There is nothing particularly insightful about saying that Canada is seeking to become less dependent on the United States. Prime Minister Mark Carney has made this abundantly clear himself, and, since Donald Trump returned to office, Ottawa has sought new markets, expanded ties with Europe and Asia, rebuilt economic relationships with China and India, and looked for ways to strengthen Canadian industrial and military capacity. Much of the commentary surrounding those moves has understandably described them as diversification, strategic autonomy, or simply Canada’s response to Trump.

That description isn’t wrong. It’s just incomplete. It treats a set of connected moves as a mood rather than a strategy, and it misses the specific logic that a number of states have used to solve versions of Canada’s problem before. China is the most fully developed practitioner of that logic among major powers, which makes it useful to think with, even though — and this is the part worth sitting with rather than rushing past — Canada is trying to do something China never had to: keep the logic and keep the alliance.

Where Canada started

For decades, Ottawa deliberately deepened its dependence on the United States because the economy and security benefits far outweighed the strategic risk. In 2025, close to 72 percent of Canadian merchandise exports went to the United States; that concentration is where Washington’s leverage comes from — a tariff threat doesn’t need to be carried out to work. The North American Free Trade Agreement (NAFTA) and its successor, the Agreement between the United States of America, the United Mexican States, and Canada (USMCA), welded North American industrial supply chains together. Geography reinforced the integration.

The security relationship developed in parallel. Canada and the United States jointly defend North American airspace through the North American Aerospace Defense Command (NORAD), cooperate through the Five Eyes intelligence partnership, and are members of the NATO alliance. When the United States invoked NATO’s Article 5 after the September 11 attacks — the only time the collective-defense provision has been used — Canada went to war alongside it, ultimately losing 158 military personnel in Afghanistan.

For successive governments in Ottawa, Washington was sufficiently predictable. Dependence rarely looked like a liability.

What changed?

President Donald Trump in his second term has treated that interdependence as leverage rather than partnership. His administration has imposed tariffs on Canadian products and threatened critical Canadian industries.

This tectonic shift in how their neighbor to the south viewed this longstanding partnership was clear to most Canadians in the first months of the administration. In July 2025, 55 percent of Canadians still described the United States as their country’s most important ally, but 59 percent identified it as Canada’s greatest threat. By July 2026, 77 percent named the United States as a threat to Canada’s economic prosperity.

Carney’s response has not been a conventional pivot away from Washington — the U.S. is too large and too close for that to be realistic. Instead, he has sought to address the considerably narrower and harder challenge of maintaining the relationship without letting it become the mechanism by which Washington wins every disagreement.

Carney signaled the doctrinal shift in March 2025, after Trump imposed automobile tariffs. He said that the United States was “no longer a reliable partner,” requiring Ottawa to “dramatically reduce” its reliance on the United States.

At Davos in January 2026, Carney made a broader argument about how “middle powers” should operate. He described an international system undergoing a “rupture, not a transition,” in which tariffs, supply chains, financial systems, and access to markets could become instruments of coercion. He called for a “dense web” of relationships built on “variable geometry” — different coalitions for different problems, rather than one bloc confronting another.

The China comparison — and where it breaks

Beijing has spent decades pursuing such optionality. It has built economic and diplomatic ties across rival camps and maintained working relationships with governments that are themselves adversaries of one another. That is close to what Carney is describing: minimize any outside power’s ability to constrain choices.

But the comparison earns its keep by showing where it fails, not just where it fits. China’s freedom of maneuver rests on staying outside the alliance system. Canada’s strategy explicitly keeps it inside. Ottawa isn’t hedging away from NATO or NORAD — it has increased defense spending, pursued more sovereign military capability, deepened defense-industrial ties with Europe, and in February 2026 became the first non-European country to join the EU’s SAFE defense-procurement initiative, while continuing a $35 billion NORAD modernization with the United States.

China built optionality by refusing commitment. Canada is trying to build optionality while holding onto the deepest commitment it has. If that works, it’s not because Canada found a Canadian version of China’s answer — it’s because Canada is attempting something China’s approach was never designed to prove is possible: durable alliance membership that doesn’t collapse into strategic dependence.

A closer comparison set

China is not the only state that has faced a version of this problem. The Association of Southeast Asian Nations (ASEAN) is the quintessential definition of hedging without formally aligning. Gulf states have long maintained security ties with Washington while building economic and diplomatic relationships with Beijing and Moscow. What unites these cases with China’s is the underlying principle, not the details: dependence narrows choices, and enough relationships enlarge them.

Canada’s variation is still the odd one out in this set, and productively so. ASEAN and the Gulf states hedge because they avoid formal alliance commitments. Canada is trying to hedge without giving one up. That makes it a more demanding test case than China, not a looser one — and a more interesting one for a security-studies audience, because it asks whether hedging logic and alliance membership are actually compatible, or whether Canada is about to find out the hard way that they aren’t.

A remnant of the past or a new beginning?

Canada has tried a version of this before. After Richard Nixon imposed an import surcharge in 1971, Foreign Minister Mitchell Sharp proposed what became known as the “Third Option” — an effort to reduce Canadian vulnerability to the United States by strengthening the domestic economy and expanding relationships elsewhere. The pull of the American market won out; Canada deepened its integration, culminating in the 1988 free-trade agreement.

Three things are different now. Canada today has far stronger trade relationships with Europe and Asia. It also is not acting alone. European and Asian countries are seeking ways to reduce their own exposure to the American market, giving Canada potential partners with many of the same concerns.

Ottawa’s motivation has also changed. Earlier Canadian governments approached diversification as a path toward sustained prosperity. Carney is framing this one around coercion and sovereignty, which changes what counts as success and what Ottawa will tolerate giving up to get there.

The conventional interpretation may still prove correct. Some of this may recede if or when a less confrontational president occupies the White House. Still, the gravitational pull of geography and accumulated economic interests will remain regardless of who governs either country.

None of that makes Carney’s strategy Chinese. It does suggest that China offers a useful way to see what might otherwise appear to be a collection of disconnected Canadian responses to Trump. Beijing has long understood that relationships are most useful when they increase a country’s choices rather than give another government the means to narrow them.

If it works, Carney will have produced something other American allies are likely to study closely. If it doesn’t, the reason will probably be visible in the China comparison in hindsight — that the alliance commitment Canada chose to keep was, all along, the one relationship “variable geometry” couldn’t route around.

Another campaign issue has gained prominence this year in a number of states: data centers. The rapid construction of these massive facilities has produced growing public opposition, particularly over electricity costs, environmental effects, and their impact on property values. That resistance has become increasingly visible at town halls across the country, with confrontations and complaints visible across local news broadcasts and on TikTok and YouTube.

I saw the beginnings of this pushback before moving away from Loudoun County in Northern Virginia two years ago. Data centers — often massive, featureless concrete boxes the size of a college football stadium — had spread rapidly across the county, which hosts more than 250 facilities, many near neighborhoods and shopping areas. Local politicians had once celebrated the industry as an economic success story. Since then, the public opposition has grown harder to dismiss as partisan.

There is little reason to expect this widening animosity to subside in the United States. The demand for computing capacity is accelerating with the expansion of artificial intelligence, which means more data centers, more pressure on electrical grids, and more competition for land and water. Public officials are increasingly cautious about approving new facilities, with some consideration tighter regulations and zoning restriction.

Americans are not alone in confronting the consequences of the data-center boom. These concerns are shared in other nations, which makes this a good moment to look outside the United States. Some of what is happening in other nations will sound familiar. Other responses will make the U.S. political and technology-industry debate look to be behind the curve.

The bottom line is that technology companies are encountering growing resistance and tighter government scrutiny around the world just as the AI buildout is gathering force. The contest is no longer simply over how many data centers can be built. Increasingly, it is over who bears their costs — and what local communities and economies receive in return.

1. Scotland: The Backlash Moves Into Politics

Scotland, like other parts of the United Kingdom, is trying to attract more investment tied to artificial intelligence and cloud computing. But proposed expansion of data centers has triggered a public backlash.

The sharpest fight is in Fife, north of Edinburgh, where developers want to build a £5 billion hyperscale data center near the small village of Auchtertool. The proposed complex would be larger than the village itself and could require as much as 600 megawatts of electricity — roughly the amount used by more than a million homes.

Residents have filed more than 1,600 objections against the project. Opponents have organized public meetings, demonstrations, and protests outside the Scottish Parliament. Similar protests are emerging around proposed facilities elsewhere in the country — local petitions opposing eight facilities have gathered, as of Monday (24 August), more than 52,000 signatures.

The complaints will sound familiar to communities in Virginia and other parts of the United States: the size of the buildings, pressure on electricity and water supplies, environmental effects, and questions as to the direct economic benefit for the communities that host such facilities.

Scottish ministers are resisting a moratorium, but they are tightening scrutiny on developers and local governments. This month, the government required the Auchtertool project to undergo a full environmental review and told local authorities to notify Edinburgh of all large data-center proposals.

Signals to watch: In October, local authorities are expected to rule on the project in Auchtertool but also in Larbert. The Auchtertool project has been delayed until the environmental review has been completed, which will require another round of public consultation. If Edinburgh intervenes further — either through a moratorium or new national planning rules — local protests will have succeeded in turning individual development disputes into a national constraint on the industry.

2. Ireland: Success Has Become the Problem

The Financial Times has been covering this issue for some time. On Sunday, it examined how data-center demand has grown faster than Ireland has expanded its grid and renewable generating capacity. That has left the government confronting energy needs it had not adequately prepared for and reconsidering options, such as nuclear power — technology for domestic generation of electricity that was banned in 1999.

Data centers consumed 23 percent of Ireland’s electricity last year, up from just 5 percent a decade earlier, and their share is expected to continue rising. Ireland might seem an unlikely place for this problem, but the country has spent decades courting foreign investment through favorable tax policies and other incentives. This campaign has brought in many of the largest U.S. technology companies and has made Ireland one of Europe’s largest data-center hubs.

The government’s support for the industry remains firm. Officials argue that data centers are important to Ireland’s broader technology economy. Still, Dublin has already had to change the terms under which new facilities can gain access to the grid.

New facilities will be expected to provide additional generating capacity to match their demand, with at least 80 percent coming from new renewable sources. The government also wants some future projects concentrated in planned “green energy parks” closer to new power generation.

In 2021, regulators imposed what amounted to a moratorium on new projects around Dublin because the electricity system could not keep pace. That restriction is now being eased. Irish contractors told The Irish Times last month that data-center activity had picked up during the previous six months, suggesting that the earlier restrictions slowed development without ending the underlying demand.

Signals to watch: Environmental groups have challenged new connection rules in the High Court, arguing that they would allow too much additional fossil-fuel generation. At the same time, EirGrid — the the state-owned electric power transmission operator in Ireland — is awaiting regulatory guidance on how future data-center applications should be processed. A court ruling against the rules, or tighter guidance from regulators, could again slow new construction; decisions favoring the government would clear more of the path for the industry’s second wave.

3. Malaysia: Growth Meets Its Limits

Malaysia’s data-center boom continues to accelerate. In July, the government reported that 36 operating facilities were drawing power from the national utility, and 23 projects were under construction. Data centers could account for 31 percent of the country’s electricity demand by 2035, up from about 7 percent this year.

The prospect that communities will be competing with the industry for basic resources is raising political concerns. Last month, the government told parliament that new data centers will be approved only when officials are satisfied that sufficient electricity and water remain available for households and local industries.

The concern is particularly acute in the state of Johor, the center of Malaysia’s data-center capacity and expansion. Local residents already deal with weak water pressure and are raising concerns about the impact of continued growth of facilities. The state and federal governments are spending hundreds of millions of dollars on new dams and water-treatment plants, while operators are being pushed toward reclaimed water and less water-intensive cooling systems. Some data centers have begun building their own treatment facilities.

The government concern for the industry’s rapid growth and the impact on community basic resources is enough that Kuala Lumpur is examining seawater desalination. However, the government’s recognition of these issues is not slowing investment. New power infrastructure contracts tied to foreign-owned data-center projects were announced in Johor in July and August.

Signals to watch: Malaysia’s government and data-center industry will meet in Kuala Lumpur on September 14 to discuss the next phase of the country’s expansion. Watch for whether the government translates its promises of protecting local communities into tougher approval standards — particularly limits tied to water and electricity use — or continues expanding infrastructure to keep pace with the industry.

4. India: Building Before the Water Arrives

India wants to become a major center for artificial intelligence and cloud computing, and the southern state of Andhra Pradesh is moving aggressively to attract the data centers needed to support that ambition. The problem is that some of the largest proposed facilities would be built around Visakhapatnam, a fast-growing coastal city that already struggles to supply enough water to its residents.

Google’s planned $15 billion data-center project in the region will require enormous amounts of electricity and water. Environmental groups and local residents have challenged the project, arguing that officials have not adequately explained how the city will support facilities of that scale without worsening existing shortages — groundwater levels have fallen and some neighborhoods have relied on tankers and newly drilled wells.

The state government has nevertheless continued to approve additional data-center projects. Officials argue that new pipelines and other infrastructure will provide enough water for both residents and industry and insist that drinking-water supplies will not be sacrificed to support the technology sector.

Signals to watch: September will test the government’s promises. Officials in Andhra Pradesh say additional water will begin reaching Visakhapatnam in the first week of the month, while a separate project is supposed to expand round-the-clock water service to thousands more households by the end of September. Opponents are already challenging the projects in court, and the state’s High Court has ordered officials to ensure that construction proceeds only after all required legal and environmental approvals are in place. Continued water shortages would give those challenges additional political weight and could increase pressure for tighter scrutiny or delays to future projects.

Watching from home, what struck me was how quickly we were learning about a war taking place thousands of miles away. Vietnam had brought war into American living rooms, but CNN’s round-the-clock coverage compressed the distance and time between the battlefield and the audience. Yet most of what we could know still depended on what reporters could witness, what their sources could tell them, and what the military was willing to disclose.

That last constraint has weakened considerably. The growth of open-source information and data (OSINT) has given journalists, academics, and independent researchers access to evidence they can collect and examine without waiting for — or relying — on a government spokesman.

Journalists and independent researchers are able to draw from such sources not just to track troop movements, identify weapons, and document destruction but also to challenge official accounts. Commercial satellite imagery, ship- and aircraft-tracking data, cellphone videos, social-media posts, and other sources increasingly allow outsiders not simply to report what governments say happened, but to test those accounts against evidence of their own.

Russia’s 2022 invasion of Ukraine accelerated this transformation already visible during Moscow’s 2014 intervention in eastern Ukraine. What became commonplace in Ukraine has since become part of the information environment surrounding the conflict in Gaza, the June 2025 attack on Iran, and the current Iran war.

I touched briefly on this evolution of OSINT in a June Signals to Watch. But reporting since outset of the Iran war — including increasingly detailed assessments of damage to U.S. bases in the Gulf — makes the media side of the transformation worth examining separately.

What Happened to the Gulf Bases?

From the outset of the Iran war, the Pentagon provided some information about attacks on U.S. forces and installations in the Gulf. But news organizations and outside researchers supplemented those accounts from publicly and commercially available information — the basis of what the U.S. intelligence community defines as OSINT.

The Financial Times asked last week whether America’s sprawling military bases in the Persian Gulf are worth rebuilding. The newspaper cited an American Enterprise Institute assessment from April estimating that repairs following Iranian attacks could cost about $5 billion. But the interesting part for me was how much AEI was able to draw separate from Pentagon briefings or congressional testimony.

AEI was not alone. Several news organizations and outside researchers were also offering regular battle-damage assessments from OSINT other than the Department of Defense. The Associated Press in late March used satellite imagery to document damage to U.S. and allied installations. In early May CNN identified damage at numerous American facilities. The Washington Post went considerably further, examining high-resolution imagery released through Iranian sources, comparing it with European and commercial satellite imagery, rejecting pictures it could not authenticate, and identifying damage that Iran itself had not publicized.

This is not simply journalism with better photographs. Reporters are employing some of the basic techniques imagery analysts have long used: comparing imagery taken at different times, and multiple perspectives, and combining with evidence collected on the ground — including photographs and videos posted to social media. Those methods can establish with considerable confidence that a building was damaged, equipment disappeared, or a previously intact facility was struck.

That does not mean the press have the means to produce a comprehensive battle-damage assessment.

I addressed that distinction last year. After President Trump declared that U.S. strikes in June 2025 had “obliterated” Iran’s nuclear program, Just Security asked me to assess what could actually be concluded from the available information. I did not share the certainty coming from Washington, but neither could the publicly available information definitively refute the administration’s claim the day after the strikes. Battle-damage analysis is iterative and fragmentary. Imagery can show that a target was struck, but it does not reveal what happened underground, whether critical equipment had been removed beforehand, whether a facility can still perform its mission, or how quickly it can recover.

The same distinction applies to the Gulf bases. Outside observers can establish physical damage with increasing confidence. Determining whether that damage prevents a base from performing its mission is considerably harder.

Tradecraft Remains Critical

That limitation points to a problem that will become more important as these sources proliferate. More evidence will not necessarily produce more accurate analysis.

Analysts inside intelligence agencies spend considerable time worrying about confirmation bias: the tendency to give greater weight to evidence supporting an existing judgment and discount information pointing elsewhere. Journalists and academics confronting thousands of satellite images, videos, social-media posts, and outside assessments face the same problem. An abundance of information can actually make confirmation bias easier because there is almost always something available to support an initial hypothesis.

Corroboration therefore matters as much as collection. Ten social-media accounts displaying the same photograph are not ten independent sources if all obtained it from a single post on the British newspaper The Telegraph. Three outside analysts agreeing about what a satellite image shows provide little additional confidence if each is relying on the first analyst’s identification.

The better journalistic investigations appear to understand it. The noted May 2026 Post assessment did not simply reproduce Iranian satellite photographs of American bases. It attempted to authenticate them against imagery from other providers and discarded images it could not independently verify. Reporters can compare overhead imagery with videos and photographs taken at ground level, examine historical imagery to establish what existed before an attack, consult specialists about weapons and equipment, and seek explanations that do not fit the initial theory.

That is essential tradecraft, whether practiced by a CIA analysts, a university professor, or a news reporter.

Peer review is equally important. Imagery that seems unambiguous to one researcher may look different to someone who understands the equipment being examined or the normal activity at a particular facility.

The intelligence community can combine overhead imagery with signals intelligence, human reporting, operational information, and other classified sources unavailable to journalists. Even then, analysts can get it wrong. During the 1991 Gulf War, CIA and CENTCOM reached sharply different conclusions about how much Iraqi armor had been destroyed — a reminder that even well-resourced, aggregate assessments can be wrong at the level of individual targets, even when the broader picture later proves roughly accurate.

More sophisticated sensors do not eliminate that problem. They simply provide analysts with more evidence to interpret.

More Eyes Are Coming

The United States can pressure an American satellite company to restrict imagery during a conflict, as it has during the Iran war. It can also tighten controls on what its own military personnel put online. Such measures can close individual windows.

They cannot close the sky from everyone.

Commercial satellite capacity is expanding outside the United States as well as within it. Chinese companies demonstrated during the Iran conflict that imagery of American military facilities could remain available even when U.S. providers restricted access. Some future providers will be owned by governments friendly to Washington. Others will operate primarily according to commercial interests. Still others may belong to countries whose interests are hostile to those of the United States.

That returns us to those damaged bases in the Gulf.

The Pentagon possesses information about their condition that the Financial Times, Washington Post, academics, and outside researchers do not. It knows more about what equipment survived, which capabilities were restored, what systems were moved, and how effectively individual installations can still perform their missions.

But reporters did not need to answer all those questions to establish that the bases had suffered substantial damage. They had a lower evidentiary threshold to negotiate: compare before-and-after satellite imagery, corroborate it against pictures and videos from the ground, consult people who understand what they are looking at, and test the resulting conclusions against other sources.

The contrast with 1991 is not that journalists today know everything the military knows, but that they no longer have to wait to accompany troops and reach the target area before building a credible picture of what happened. Their access to the information and tools needed to assess a “hot war” will only continue to grow — and be published faster.

Pyongyang was unimpressed. Yesterday (August 20), Kim Yo Jong, sister of Kim Jong Un, denigrated South Korean President Lee Jae Myung and dismissed the reduced exercises as inconsequential, arguing that Washington’s underlying “hostile policy” remained unchanged. North Korea also launched about 10 short-range ballistic missiles toward the East Sea, according to South Korea’s Joint Chiefs of Staff.

The episode comes as Seoul already faces an uncomfortable nuclear imbalance. North Korea continues producing fissile material and expanding an arsenal that South Korean Defense Minister Ahn Gyu-back said Thursday (August 19) estimates to be between 80 and 120 nuclear warheads — a range the ROK military has not officially confirmed, and outside estimates vary considerably; Trump on Wednesday put the number at 57.

Trump, meanwhile, continues to praise Kim Jong Un and seek another meeting with him without having produced an agreement that slowed that buildup. By June 2020, two years after Trump’s celebrated Singapore summit with Kim, Heritage Foundation analyst Bruce Klingner estimated that North Korea had probably built “eight or more additional nuclear weapons” — an expansion that would go unimpeded in subsequent years.

The exercise decision adds another element to broader uncertainty about Washington’s reliability as an ally. That does not suggest that South Korea — or any other American ally — is preparing to abandon U.S. extended deterrence. It does make the durability of that deterrence, and how other governments perceive it, worth watching.

Given that uncertainty, North Korea’s continued nuclear expansion, China’s military buildup, and Iran’s insistence on retaining an enrichment capability, this is a useful moment to ask a different question about nuclear proliferation. Rather than predicting which country will become the tenth member of the nuclear club, it is worth examining how quickly some of the most frequently discussed candidates could acquire a nuclear weapon if their governments decided to do so.

Japan and South Korea are obvious cases. Both possess sophisticated nuclear industries, advanced scientific and engineering establishments, and capable missile or space-launch programs. Both also face a deteriorating regional security environment dominated by China’s military expansion and North Korea’s growing nuclear arsenal. The relevant issue is not whether either government has decided to build nuclear weapons — there is no evidence that they have — but how much technical distance separates them from such a capability and what political developments might cause that distance to matter.

Saudi Arabia presents a different case. Riyadh has long figured in proliferation discussions because Crown Prince Mohammed bin Salman has said that Saudi Arabia would seek nuclear weapons if Iran acquired them. That possibility has become more salient following the Trump administration’s agreement on civilian nuclear cooperation with Riyadh, which, unlike the U.S.-UAE agreement, does not prohibit Saudi enrichment of uranium.

In the current geopolitical environment, it is worth examining what signals would indicate that this was changing — and, just as important, how quickly each country could acquire a weapon if its government made the decision to do so.

1. South Korea: A Two-Year Clock Would Start With Fissile Material

South Korea has most of the industrial and scientific capabilities needed to develop nuclear weapons, but it lacks the one thing without which it cannot build them: weapons-usable nuclear material.

Its civilian nuclear industry is among the world’s most advanced, including domestic reactor design, fuel fabrication, heavy manufacturing, and nuclear research. Its defense industry also produces ballistic and cruise missiles that could eventually serve as delivery systems. But South Korea has no uranium-enrichment plant, no reprocessing plant, and consequently no stockpile of weapons-grade uranium or separated plutonium.

Siegfried Hecker, former director of Los Alamos National Laboratory, estimated in 2023 that Seoul would need at least two years either to construct a centrifuge facility producing highly enriched uranium or use its civilian reactors and reprocessing capabilities to produce enough plutonium for a scant number of weapons. Weapon design, testing, integration with delivery systems, and creation of nuclear command-and-control arrangements would require additional work.

More recent academic assessments are broadly consistent with Hecker’s judgment. Two more recent assessments cited in a 2025 Journal of East Asian Studies study similarly put even a minimal South Korean nuclear capability at roughly two to three years.

There is already substantial domestic support for keeping that option open. The Asan Institute’s March 2025 survey found 76.2 percent of South Koreans supported acquiring indigenous nuclear weapons — the highest level recorded since it began polling the issue in 2010. That should not be confused with a government decision: South Korea’s defense minister reiterated yesterday that Seoul remains committed to relying on the U.S. nuclear umbrella and opposes acquiring its own nuclear weapons.

The signal to watch is consequently not another poll favoring the bomb. It is movement toward nuclear latency: serious proposals to develop enrichment or reprocessing capacity, changes to Seoul’s commitments under nuclear-cooperation agreements, or sustained support from governing rather than opposition politicians for creating an indigenous weapons option.

2. Japan: The Technical Barriers Are Lower

Japan begins considerably farther along the technical path than South Korea. It has an advanced nuclear industry, experience with uranium enrichment and reprocessing, a sophisticated space-launch program, and the scientific and manufacturing base required for a weapons program. Most important, Japan already possesses separated plutonium.

According to the Japanese Atomic Energy Commission, Japan controlled approximately 44 metric tons at the end of 2024 — over 9 tons stored domestically and another 36 tons held in Britain and France; the domestic cache would be enough for 1,000 Nagasaki-sized weapons. Possessing the material does not mean Japan has a bomb-making stockpile ready for immediate use, but Tokyo would not face the same initial requirement as Seoul to build a new industrial capability before it could obtain fissile material.

The time that this saves Japan is less certain. Some estimates assess that the technologically-advanced nation could acquire a nuclear weapon in several months to three years. Japan has the relevant nuclear infrastructure and materials, but a deployable arsenal would still require weapon design and manufacture, decisions about testing, adaptation of delivery systems, and creation of nuclear command-and-control arrangements.

The political barrier remains formidable, but it is no longer quite as immovable as it once appeared. Prime Minister Sanae Takaichi has opened discussion of Japan’s three non-nuclear principles — not possessing, producing, or allowing nuclear weapons into Japanese territory — as her government revises its national security strategy. That said, the immediate debate centers primarily on whether Japan should continue prohibiting the introduction of U.S. nuclear weapons, rather than whether Tokyo should build its own.

Public opposition also remains substantial. A July poll found 43 percent of Japanese seeing no need to review the principles, while 27 percent favored a review and 29 percent were unsure.

The signals to watch are therefore whether debate over the non-introduction principle broadens into reconsideration of the prohibitions on possessing or producing nuclear weapons, or whether doubts about U.S. extended deterrence begin shifting what is still a largely political barrier.

3. Saudi Arabia: Building the Infrastructure That Could Shorten the Clock

Saudi Arabia is considerably farther away from an indigenous nuclear weapon. It has substantial financial resources and increasingly capable ballistic missiles, but it lacks the nuclear infrastructure that gives Japan and South Korea their latency. The kingdom has no operating commercial nuclear power reactors, no industrial-scale uranium-enrichment capability, no reprocessing plant, and no known stockpile of weapons-usable nuclear material. There is no recent technical assessment sufficiently robust to justify assigning Saudi Arabia a specific three-, five-, or ten-year breakout time.

That makes last month’s U.S.-Saudi civil nuclear agreement particularly important. Unlike the 2009 U.S.-UAE agreement, under which Abu Dhabi renounced enrichment and reprocessing, the Saudi agreement does not prohibit either activity. It also does not require Riyadh to accept the IAEA Additional Protocol, which provides inspectors greater access and information to detect undeclared nuclear activities. The agreement does not provide Saudi Arabia with enrichment technology, and Trump subsequently said implementation would depend on Riyadh normalizing relations with Israel.

Low-enriched uranium used for reactor fuel is different from highly enriched uranium suitable for a weapon. The significance of this prospective agreement is that operating an enrichment program would give Riyadh facilities, trained personnel, supply chains, and practical experience with the technology needed to produce highly enriched uranium if a future government chose to do so.

The signals worth watching are therefore tangible ones—construction and scale of Saudi enrichment facilities, the enrichment levels permitted, accumulation of enriched uranium, the safeguards ultimately imposed on the program, and the degree of foreign nuclear assistance Riyadh receives.

A lower-probability, high-impact variable is Pakistani assistance. Saudi Arabia’s new defense pact with nuclear-armed Pakistan and Turkey gives Riyadh a closer formal security relationship with the one state in the group that already possesses nuclear weapons. None of the signatories have addressed the question of nuclear deterrence, and there is no evidence Islamabad has agreed to sensitive technology; however, if that changed, Saudi Arabia’s path to a bomb would shorten sharply.

Speaking in Manila before traveling to Indonesia, Thailand, and Cambodia, Colby said Washington was “digging in” to preserve a favorable balance of power in the Indo-Pacific. But that reassurance came with a familiar qualification. The United States wants “partners, not protectorates” — countries willing to spend more on their own defense and take greater responsibility for regional security. Colby tied that effort to “deterrence by denial” along the First Island Chain, stretching from Japan through Taiwan and the Philippines toward Borneo.

The message is consistent with the administration’s national defense strategy, which identifies China as the principal long-term military challenge and the Indo-Pacific as the priority theater. Southeast Asia has heard similar assurances from previous presidencies, however. Since the Obama administration announced its “rebalance” toward Asia, successive administrations have emphasized that American strategic attention had shifted to the region even as events elsewhere repeatedly competed for money, forces, and presidential attention.

The economic record has added to those doubts. Congress never approved the Trans-Pacific Partnership that was supposed to provide the economic foundation for Obama’s rebalance, and Trump withdrew from it in 2017. Biden’s Indo-Pacific Economic Framework returned Washington to regional economic diplomacy but offered neither a conventional trade agreement nor broader access to the U.S. market.

China has spent the same period deepening its economic position in Southeast Asia. It has been ASEAN’s largest trading partner since 2009, and two-way merchandise trade reached roughly $772 billion in 2024, about one-fifth of ASEAN’s total trade. Chinese investment has also become embedded in regional infrastructure and supply chains through the Belt and Road Initiative, including rail, ports, energy, and industrial projects. ASEAN and China are also part of the Regional Comprehensive Economic Partnership, while Beijing and ASEAN have continued to expand their own free-trade framework.

That history helps explain an ASEAN position often frustrating to Washington. Southeast Asian governments have spent years deflecting pressure, explicit or implied, that they choose between the United States and China. They want Chinese trade and investment, an American counterweight to Chinese power, and sufficient distance from both to keep either from acquiring too much leverage.

Regional analysts describe this in various ways—hedging, strategic autonomy, flexible alignment—but the practical objective is straightforward: preserve options.

Trump’s second term may be making that strategy more attractive rather than less. His administration has imposed tariffs across Southeast Asia while asking many of the same governments to increase defense spending. The war with Iran has simultaneously pushed up energy costs across a region heavily exposed to Gulf supplies and diverted American military resources from the theater Washington calls its priority. The USS George Washington, based in Japan, is now heading toward the Middle East to relieve the USS Abraham Lincoln, which has been supporting the Iran campaign since early this year.

Another reminder — shocking to most observers — came this week. Trump ordered the Pentagon to substantially reduce major exercises with South Korea, citing not just their cost but also Seoul’s refusal to participate in U.S. actions against Iran. The decision is unlikely to reassure governments already questioning how durable U.S. security commitments in the Indo-Pacific will prove when Washington’s priorities shift elsewhere.

None of this means Southeast Asia is about to choose China. That is the mistake in much of the commentary surrounding declining confidence in Washington. Beijing remains too powerful, too close, and in several cases too threatening for most ASEAN governments to regard it as a comfortable alternative.

The more plausible regional response — as it has been in the past — is to wait.

Buying Time

The 2026 State of Southeast Asia survey provides a useful starting point. When respondents were asked with which power ASEAN should align if forced to choose between the United States and China, China narrowly edged the United States. Indonesia favored China heavily, while the Philippines overwhelmingly chose the United States; Malaysia and Singapore leaned toward Beijing, Vietnam toward Washington, and Laos was evenly divided. Still, respondents continued to hold firm in their support for “flexible multiple choices,” resisting efforts to turn cooperation on one issue into alignment across the board.

Trump the individual gives them another reason to continue doing so. American presidents change; American policies can change with them. A trade arrangement disliked today may be renegotiated later. A future administration may put greater value on alliances, restore parts of the economic relationship, or again make Southeast Asia a larger diplomatic priority. Governments that avoid irrevocable choices now preserve the ability to deal with that Washington later.

China presents a different problem. ASEAN governments cannot assume that a future Chinese leader will pursue fundamentally different regional objectives. That gives Southeast Asian governments even greater reason to preserve room to maneuver: keep China engaged, limit dependence where possible, and build relationships with other powers, including Australia, Japan, South Korea, and the European Union, strong enough to prevent Beijing from becoming the region’s only option.

The Iran war has made that logic more concrete. Indonesia has had to find billions of dollars in additional savings and potentially billions more for energy subsidies as oil prices rose. Indonesian President Prabowo Subianto is already under pressure to reconcile expensive domestic programs with concerns about the rupiah, capital outflows, and fiscal discipline. His proposed 2027 budget seeks to narrow the deficit even after higher fuel subsidies this year added to the strain.

For Jakarta, the obvious answer is not to choose Beijing because Washington created an economic problem. China carries risks of its own. Nor is it to make the larger defense commitments Washington would prefer while domestic resources are under pressure. The safer choice is to protect Indonesian freedom of action, cooperate where interests overlap, and preserve enough room to maneuver to avoid commitments that narrow future options.

The same logic will not produce identical policies in Thailand, Malaysia, Singapore, or Vietnam. It does not have to. The common element, however, is a reluctance and caution.

The Philippines is the clearest exception to this pattern, because it doesn’t have the option to wait. Philippine officials say Chinese Coast Guard personnel attacked last month a Filipino sailor during a confrontation at Second Thomas Shoal — a reef in the South China Sea where Manila keeps a deliberately grounded World War II-era ship as a way of maintaining a physical presence. Further confrontations followed near Scarborough Shoal. For President Ferdinand Marcos Jr., China isn’t a long-term risk to plan around; it’s a government contesting Philippine positions in real time. Manila is the one Southeast Asian capital with a mutual defense treaty with Washington, and it’s the only regional power capable of backing that treaty with force at the scale China presents.

Even so, Marcos has kept a second track open. He said last week that joint oil and gas exploration with China remains “a distinct possibility,” and has said the Philippines isn’t trying to distance itself economically from Beijing even as the maritime standoffs continue. The country under the most direct Chinese pressure in the region is still hedging wherever it can — it has simply run out of room to hedge on security.

What Colby Should Take Home

If there is a flexible realism at work in Southeast Asia, it belongs to ASEAN, not Washington. Regional governments have spent decades taking security cooperation from the United States where it serves their interests, economic benefits from China where they can, and avoiding commitments that allow either power to define the terms of the relationship.

Undersecretary Colby may come away with indications that ASEAN governments will expand exercises, military access, training, intelligence cooperation, and defense spending. However, such promises reflect ASEAN pragmatism, not necessarily acceptance of the Trump administration’s broader strategic assumptions.

The second Trump administration has come to treat security relationships as carrying obligations that it can define. Iran exposed that gap. Japan, Australia, South Korea, NATO members, and other partners resisted requests to support a conflict they had neither initiated nor been meaningfully consulted about. Trump’s subsequent decision to reduce major exercises with South Korea will reinforce concerns that even formal alliance commitments can become entangled with this administration’s expectations outside the original bargain.

Caution toward Washington does not mean Southeast Asia is preparing to accept Chinese primacy. Beijing continues to provide governments ample reason to want an American counterweight, most visibly through coercion in the South China Sea. The Philippines shows how quickly the calculation changes when that pressure becomes immediate. Further Chinese moves against Philippine positions, particularly while U.S. forces and weapons are being drawn toward Iran, would test both Beijing’s assessment of American distraction and Washington’s ability to reconcile its declared Indo-Pacific priority with the demands of another war.

ASEAN leaders have learned to cooperate without surrendering room to maneuver. That may be the harder lesson from Colby’s trip. He may return with deliverables. He should be careful not to confuse them with alignment.

In Canada, more than 4 million hectares—about 10 million acres — have burned this year, with major fires stretching from Ontario and Quebec to British Columbia. The Bald Range fire in British Columbia alone forced more than 20,000 people to evacuate this month. In Europe, France has experienced its worst wildfire season on record, with more than 116,000 hectares burned by the end of July, while Spain had already lost more than 226,000 hectares as of Tuesday (August 11). Greece is again battling major fires as well, including one this week that forced hundreds of people to evacuate by sea.

The United States, where reports of western wildfires have become grimly familiar, is experiencing an even more severe season. More than 46,500 fires have burned 6.4 million acres — roughly 10,000 square miles — so far this year, the largest area burned by August 12 since federal records began.

Stories about these recurring disasters inevitably return to the changing climate behind them: hotter and drier conditions that make severe fires more likely, warming oceans, melting ice, and rising seas. The world is already about 1.4 degrees Celsius warmer than it was before industrialization, and scientists expect conditions conducive to extreme fires to become more common as temperatures rise.

This summer’s fire-prone weather, for example, was estimated to have been made twice as likely by climate change in southwestern France and twenty times as likely in central Spain. Efforts to slow that warming depend on substantial reductions in greenhouse-gas emissions, something made harder by the continued emissions of the world’s largest polluters, led by China and the United States.

Wildfires present a somewhat different problem. We cannot prevent every lightning strike, discarded cigarette, or act of arson, but neither must we simply watch a small fire become a large one. There is no machine waiting to vacuum greenhouse gases from the atmosphere like something out of a Ridley Scott movie. There are, however, increasingly sophisticated ways to detect a fire almost as soon as it starts, determine where it is heading, and get firefighters or aircraft to it before it spreads beyond their ability to contain it.

I addressed global warming and rising temperatures last August in this newsletter. Today, I look instead at the technological race in wildfire management, which is becoming less about how to extinguish an inferno than how to keep one from developing in the first place.

1. Finding the Fire Before It Becomes One

For generations, early wildfire detection often depended on a person seeing smoke — from a lookout tower, an aircraft, or a passing car — and reporting it. Wildlife management efforts in recent years have looked toward technological solutions to eliminating much of that delay.

California has gone furthest with cameras. The state’s ALERTCalifornia network of static and mobile tracking systems — deployed in 2023 — includes more than 1,200 cameras positioned across fire-prone areas monitoring for smoke. According to the state, the AI monitoring system has detected more than 900 fires on state lands before the first emergency call was received. Elsewhere in the West, similar systems have sometimes provided firefighters critical minutes of additional warning.

Now the same race is moving into space.

In July, the nonprofit Earth Fire Alliance, partnering with Sacremento, placed the first three operational satellites of its FireSat system into orbit. Eventually, a constellation of more than 50 satellites are scheduled to be in orbit by 2030 — a configuration designed to identify fires roughly five meters across and revisit fire-prone areas every twenty minutes or less.

Greece is trying an even more direct approach. Four thermal satellites built by Germany’s OroraTech were launched in May as part of the new Hellenic Fire System. Greece is the first country to integrate its own dedicated wildfire satellite constellation into its national firefighting system. The satellites are designed to identify heat sources about four meters across, with firefighters now being trained to incorporate the information into operations. The system is expected to become fully operational in September.

Wildfire technology has been improving for years, but 2026 may represent something closer to a threshold. Falling barriers to putting small satellites into orbit are allowing governments and private organizations to build constellations devoted specifically to fire rather than borrow imagery from satellites designed for other purposes. Thermal and multispectral sensors can now identify much smaller heat signatures; artificial intelligence can process enormous streams of camera and satellite data quickly enough to flag possible fires without waiting for an analyst to find them; and some processing is beginning to move onto the satellites themselves.

Signal to watch: Greece will provide a useful test this fall of whether firefighting systems are structured to take full advantage of earlier detection. If faster warning does not translate into faster initial attack — or into more fires being contained while still small — the technological advance will have solved only half the problem. California, now in the heart of its high-risk period, offers a parallel test of whether satellite and ground-based detection can be integrated quickly enough with dispatch, aircraft, and crews to exploit those additional minutes.

2. Fighting the Fire When Piloted Aircraft Are Not Enough

Early detection solves only part of the problem. The other half of the solution is ensuring firefighters can deploy to a newly detected fire quickly enough to matter. Numerous companies and firefighting agencies in recent months have carried out successful trials of autonomous aircraft not simply to observe fires, but to serve as the first responders and suppress them.

In California last month, CAL FIRE tested five autonomous drones working together to deliver between 500 and 1,000 gallons of foam against simulated fires. The aircraft, developed by California startup Seneca, can carry roughly 100 pounds of water or retardant. The drones are small enough to be transported by pickup truck and are designed to be refilled and returned to the air within minutes.

In June, finalists in the $11 million XPRIZE Wildfire competition were tested outside Fairbanks, Alaska, on whether autonomous systems could find and suppress fires. One finalist, Germany’s Dryad Networks, linked sensors attached to trees with reconnaissance and suppression drones. When the sensors detected smoke, one drone confirmed and located the fire before another delivered 26 gallons of suppressant.

Other approaches are developing at different scales. The startup company Rain is adapting larger aircraft for autonomous wildfire suppression, including work with Sikorsky on an unmanned Black Hawk, while other XPRIZE teams are experimenting with heavy-lift aircraft and drone swarms. Israel this month completed an initial trial of an autonomous aircraft equipped with infrared sensors that can patrol for fires for several hours, including in darkness and smoke.

None of these aircraft can replace specialized fixed-wing aircraft and helicopters against a major wildfire — for now. However, such autonomous aircraft could be dispatched against a detected ignition, operate after dark, or continue through smoke that grounds piloted aircraft. NASA’s Advanced Capabilities for Emergency Response Operations (ACERO) program is preparing for increasingly crowded wildfire airspace and the need for all firefighting elements to have a common operating picture. ACERO is intended to provide the architecture through which remotely piloted aircraft can eventually operate alongside conventional firefighting aviation, including at night and in poor visibility.

Signal to watch: The XPRIZE Wildfire competition is expected to announce its winners next month, but there are no specific near-term thresholds that would by themselves mark autonomous firefighting as operationally mature. Key signals will come later: whether any of these systems are actually used during a wildfire to detect, verify, and attack an ignition before conventional aircraft or crews arrive. Until then, 2026 is best understood as a year of field trials and early deployment rather than the year autonomous aircraft fundamentally changed wildfire suppression.

3. When Technology Runs Into Capacity

Better detection and autonomous aircraft may eventually help close one of wildfire management’s biggest gaps: the shortage of aircraft and crews available to respond quickly. But those technologies are not yet deployed at anything close to the scale required. For now, once a wildfire grows beyond the point where a small initial response can contain it, firefighting agencies still depend on a limited pool of specialized aircraft, pilots, and maintenance crews.

Europe is confronting that problem today.

The continent’s largest aerial emergency-services company, Avincis, warned in July that longer fire seasons are disrupting a system that traditionally shifted aircraft between the northern and southern hemispheres as seasons changed. Aircraft are increasingly needed somewhere almost year-round, while experienced pilots are becoming harder to find. In Spain and Portugal alone, Avincis aircraft had already flown more than 5,000 firefighting hours by late last month — more than twice the level for the same period in 2025.

France offers a sharper example. During the worst days of its July fires, only five to eight of the country’s twelve Canadair water bombers were reportedly available. The aircraft average roughly thirty years old, while heavy use, maintenance demands, and shortages of spare parts have reduced availability. France has ordered replacements, but the first new aircraft are not expected before 2028.

The European Union leases 22 firefighting aircraft from member states and has tried to create additional depth through its strategic reserve of European disaster response capabilities and stockpiles (rescEU). Europe had expected delivery of the first of 22 new De Havilland Canada DHC-515 water bombers before the 2027 fire season, but the delivery has slipped to 2028. Even as France and Spain battled major fires in July, officials made the pragmatic decision to hold some aircraft back, recognizing that another emergency could erupt elsewhere.

For now, Europe is managing scarcity — precisely the problem the next generation of autonomous systems is supposed to ease.

Signals to watch: With the delivery of DHC-515s delayed, the continent faces at least two more — and likely longer — fire seasons in which aging fleets must carry much of the burden. The more consequential development will be whether governments decide they can no longer wait. Several European companies are developing converted or new firefighting aircraft. France’s Kepplair says it hopes to begin deliveries in 2027, while Positive Aviation is targeting 2028 with a converted ATR-72.

That helps explain why the Chinese shipping company Sea Legend Line is launching the first regular container service between China and Europe through Russia’s Northern Sea Route. The Arctic route will allow the company to bypass the strait of Malacca and Bab el-Mandeb and the Suez Canal, and in some cases cut sailing to some destinations in half. The warmer climate and melting ice in the Arctic region makes that possible.

Physical closure of the major maritime arteries has historically been unusual. Canals have been blocked by war, wrecks, and geography: Suez remained closed from 1967 until 1975, while the grounding of the Panama-registered Ever Given stopped traffic for nearly a week in 2021. Natural straits are harder to plug. Even during the Tanker War of the 1980s, when Iran and Iraq attacked neutral shipping and Iran laid mines in the Gulf, Hormuz remained open.

A declassified CIA assessment from 1987 judged that Iran had plans to close Hormuz but probably could not sustain a closure for more than one or two weeks. Washington nonetheless treated the threat seriously. Operation Earnest Will, which escorted reflagged Kuwaiti tankers through the Gulf, became the largest U.S. naval convoy operation since World War II. Its first convoy also demonstrated the limitations of protection when the tanker Bridgeton struck an Iranian mine in July 1987; a missile damaged the Sea Isle City months later, and an Iranian mine nearly sank the USS Samuel B. Roberts the next year.

That episode harkens back to the previous century. The United States sent naval forces into the Mediterranean in the early 1800s in part to protect American merchant ships from Barbary corsairs. European merchants had been writing piracy and corsair losses into marine-insurance contracts centuries earlier. The problem was dangerous but predictable enough for commerce to continue.

Modern piracy reproduced much the same arrangement. Somali piracy became a major threat after 2005; piracy and armed robbery had already become a serious concern around the Strait of Malacca in Southeast Asia. Ships were hijacked, crews held for ransom, and premiums increased, but neither route ceased functioning. A 2013 World Bank study estimated that Somali piracy imposed about $18 billion annually in additional trade costs at its peak—roughly equivalent to a 1 percent tax on commerce through the affected region.

Shipping companies adapted. They increased speed through dangerous areas, posted additional lookouts, installed barriers and citadels, hired armed security, coordinated with naval forces, and paid higher insurance premiums. Governments and industry developed reporting networks and common practices.

Yet, missiles and drones have changed what a merchant vessel can reasonably be expected to defend against.

The Yemeni-based Houthis did not need to occupy Bab el-Mandeb Strait or sink enough ships to form a physical barrier. Their arsenal has given them the ability to combine cruise and ballistic missiles, one-way attack drones, unmanned surface vessels, and small boats in repeated attacks launched from shore. Major carriers responded by diverting around the Cape of Good Hope. The route itself remained navigable; commercial decisions produced much of the effective denial. More than 100 Houthi attacks on commercial vessels have affected ships connected to more than 60 countries since late 2023, according to the U.S. Maritime Administration.

This is where the comparison with piracy begins to break down. Razor wire can frustrate men climbing aboard from a skiff. Armed guards and maneuvering can deter some small-boat or unmanned-surface attacks. They offer little protection against an anti-ship missile or an airborne or maritime drone. Current industry guidance concentrates increasingly on avoiding targeting rather than defeating the weapon: voyage-specific threat assessments, variations in routes and timing, scrutiny of electronic emissions, and coordination with maritime-security centers.

The Iran war has revealed the same principle on a larger scale. A proxy supplied with sufficient weapons and targeting assistance does not need to destroy every vessel to alter global shipping, but to create a persistent probability of loss.

A terrorist organization could conceivably seek a maritime equivalent of a 9/11 attack: disable or sink a large vessel in exactly the right section of a canal and produce a catastrophic, long-term obstruction. However, the success of such an operation would require an unusual combination of access, target, and location. Precision weapons offer a capable state or proxy something strategically different. Repeated attacks can impose uncertainty without requiring the spectacular success of physically plugging the waterway.

The private sector is already adapting, but mostly at the tactical level. U.S. guidance warns that Automatic Identification System transmissions and other commercial electronic emissions can assist adversary targeting and recommends that ships avoid predictable routes and schedules where circumstances permit. The Joint Maritime Information Center, UK Maritime Trade Operations, U.S. Naval Forces Central Command, and partner organizations fuse incident reports, threat assessments, and vessel information for operators. These measures can reduce exposure; none provides a merchant ship with an air-defense system.

That leaves governments with an awkward problem. Permanently placing warships around thousands of commercial transits at several geographically separated chokepoints is impractical. The more important safeguard, therefore, is redundancy, and preparedness that is much less even.

Energy has gone furthest because governments have spent decades planning for supply disruptions. Saudi Arabia’s East-West pipeline and the UAE’s line to Fujairah together can provide about 4.7 million barrels a day of capacity that bypasses Hormuz. The UAE is constructing another 1.5-million-barrel-a-day pipeline expected to become operational in 2027. Strategic petroleum reserves provide another buffer, and this year the International Energy Agency conducted the largest coordinated emergency oil-stock release in its history.

Those mechanisms cushion a disruption but cannot replicate the volume normally moving through a major chokepoint indefinitely. Nor does it eliminate the vulnerability: these pipelines are long, fixed linear targets.

The distinction becomes more troubling outside oil. LNG has no comparable international strategic-reserve system. Fertilizer, food inputs, industrial components, and containerized goods depend on different combinations of inventories, suppliers, and transport routes, some considerably easier to substitute than others.

The Arctic is another form of redundancy, not an escape from vulnerability.

China new service through the Arctic region reduces dependence on Malacca, Bab el-Mandeb, and Suez, and rising orders for ice-class ships suggest that others see commercial possibilities. However, the Northern Sea Route introduces a different concentration of risk. Russia controls access and much of the supporting infrastructure. Allianz has recorded more than 500 reported shipping incidents in Arctic Circle waters during the past decade, with machinery failure or damage accounting for almost half. Search-and-rescue, salvage, and pollution-response capacity remain thin compared with established routes, while a serious casualty would occur in an environment far less forgiving than the Red Sea or the Mediterranean.

Nor would moving north make military vulnerability disappear. A Houthi-like organization may struggle to maintain surveillance, communications, and targeting across Arctic distances, but a capable state controlling the coastline would not.

The next stage should be visible before another chokepoint closes.

The first signals will probably come from the private sector rather than intelligence warnings or government declarations. War-risk premiums, exclusions, and the expansion of designated high-risk areas can change before traffic collapses. So can carrier schedules. The Center on Foreign Relation’s recent study identifies rising insurance premiums, commercial rerouting, and advisories from flag states or classification societies as early indicators precisely because commercial actors make their own judgment about danger.

Military planning offers another set of indicators. The capabilities to watch are not simply additional destroyers, but cheaper counter-drone defenses, stronger mine-countermeasure forces, persistent surveillance of approaches to chokepoints, and strengthened arrangements that allow navies to share targeting and threat information rapidly with commercial operators. Equally important will be evidence that contingency exercises are moving beyond oil and testing what happens when attacks disrupt LNG, fertilizer, food, or industrial supply chains simultaneously.

The final indicator is proliferation. The Houthis have demonstrated something that maritime leverage no longer requires a blue-water navy. The actors worth watching are those near heavily traveled waters acquiring shore-based anti-ship missiles, longer-range drones, and unmanned boats—and, just as important, the surveillance and external intelligence needed to identify the right ships at the right time.

The Chinese shipping company is responding to a world in which the reliability of geography can no longer be assumed. If such adjustments grow — new routes, bypass infrastructure, inventories, and permanent changes in shipping patterns — the measure of a chokepoint will no longer be whether someone can physically close it. It will be how much investment the rest of the world is willing to make to avoid having to pass through it.

Finland has spent decades preparing young people to navigate an information environment in which fabricated content is increasingly difficult to distinguish from the real thing. That emphasis fits a broader Finnish tradition of societal preparedness, shaped in no small part by the country’s history and proximity to Russia: build resilience by giving citizens the tools to recognize and withstand threats rather than simply trying to shield them from them.

Helsinki is now considering adding restrictions on children’s social-media access to that longstanding emphasis on education and resilience. Finland may therefore be moving toward a hybrid approach: limiting some access while also preparing young people for the digital environment they will eventually have to navigate. Elsewhere, however, the policy momentum has shifted more heavily toward restrictions.

Last November, I alerted readers that Australia was preparing to bar children under sixteen from having accounts on the major social-media platforms. I described it as the beginning of a broader period of regulatory experimentation across the globe and argued that the important question would not be which model “won,” but which combination of tools actually reduced harm without sacrificing privacy or simply pushing young people elsewhere.

Eight months later, Australia has some early results, European governments are moving from debate toward legislation and enforcement, and similar measures are appearing in Asia. That makes this a useful moment to see what those experiments are beginning to tell us.

1. Australia Social Media Restrictions Have Reduced Accounts, Not Use

In 2025, Australia entered the debate over children’s access to social media with a straight-forward proposition: restrict anyone under 16 years old from holding an account on such a platform. After the law took effect last December, the first returns appeared to justify the approach. Companies reported removing or restricting 4.7 million accounts — an extraordinary figure in a country of 28 million people.

The first government-commissioned assessment, released last week, tells a more complicated story. The proportion of children aged 10 to 15 who owned social-media accounts fell from 52 percent to 42 percent. But more than 81 percent continued to use at least one restricted platform, compared with 86 percent before the ban, and daily use has hardly changed. About half of those who retained accounts said they had never been asked to verify their age. Others entered false birth dates or were incorrectly classified as 16 or older.

Nor did the children simply put down their phones. Some activity shifted to messaging and gaming services outside the restrictions. Parents also reported knowing less about what their children were doing online. The Australian government has responded by doubling the maximum penalty for noncompliant companies and expanding the eSafety commissioner’s power to demand evidence of what platforms are actually doing.

It is too early to know whether tougher penalties for the platforms will provide the simple fix, although I have my doubts. Even if enforcement improves, the true measure will be whether the law reduces the harms that prompted it: bullying, exposure to damaging content, and compulsive use.

The early results do not settle whether Australia’s experiment will work. They do expose the distinction other governments will have to confront: preventing children from owning accounts is not the same as preventing them from using social media.

Signals to Watch

• Whether stronger enforcement produces measurable declines in daily use, rather than further reductions in formally registered accounts.

• More intrusive age-verification requirements as Australia presses companies to close the gap between claimed compliance and actual use.

2. Europe Moves From Platform Safeguards to Age Limits

Europe’s response to children’s use of social media has hardly followed a straight line. The European Union’s first major step was the Digital Services Act, adopted in 2022. It required online platforms to address risks to minors and barring targeted advertising to children, but it did not establish an EU-wide age limit for social-media use.

More recently, Brussels has been using the DSA to press the platforms directly. In April, the European Commission announced preliminary findings that Meta was failing to keep children under 13 off Facebook and Instagram; in July, it reached a similar preliminary conclusion that features such as infinite scroll and autoplay posed risks to minors. If those findings are upheld, the Commission can require changes and impose penalties.

A shift toward age restrictions for social media appears to be underway. France last month became the first European country to approve a national ban, barring children under 15 from social-media platforms beginning September 1. Greece plans to prohibit social-media access for children under 15 beginning in January. Denmark has announced the same minimum age with limited parental exceptions beginning at 13. Spain, Poland and Slovenia are considering variations of these laws.

Brussels itself is also moving toward a common approach. In July, an expert panel convened by European Commission President Ursula von der Leyen recommended that children under 13 not have access to what it calls “social media plus”—a category that extends beyond traditional social media to services such as some video games and AI chatbots. Von der Leyen endorsed the need for age-based restrictions and promised draft EU legislation this autumn.

Europe is also trying to solve the enforcement problem that Australia is confronting. Last year, the European Commission rolled out an open-source blueprint for a system that would allow users to prove they meet an age threshold without revealing their identity or exact age. However, the system has yet to be tested at anything approaching EU-wide scale, and the Commission itself acknowledges that the checks can be circumvented.

Europe has not settled on a common age, much less demonstrated that its technology will prevent the evasions already evident in Australia. But the direction is considerably clearer than it was a year ago, with governments treating the issue as a regulatory problem requiring enforceable rules, rather than another matter to be left to parents and the platforms themselves.

Signals to Watch

• Von der Leyen’s promised autumn legislation will show whether Brussels settles on the expert panel’s under-13 threshold or moves closer to the higher age limits favored by several national governments.

• France’s law still faces constitutional review and scrutiny over its compatibility with EU rules, which will provide an early legal test of how far member states can go on their own.

• If Europe is unable agree on a common minimum age, will it revert back to producing a patchwork of national thresholds, exceptions, and enforcement rules.

• Europe’s age-verification plans will face the same basic test already emerging in Australia—and the same simple question: What happens when children circumvent the system?

3. Asia Is Emerging as the Next Testing Ground

Indonesia began restricting social-media access for children under 16 in March, requiring platforms it considers high-risk — including TikTok, YouTube, Instagram, Facebook, X and Roblox — to deactivate accounts held by younger users. Malaysia followed in June, prohibiting anyone under 16 from registering a new social-media account. Together, the two countries have turned Southeast Asia into the next large-scale test of whether age restrictions can work across populations far larger and more digitally uneven than Australia’s.

The approaches are similar in ambition but remain unsettled in practice. Indonesia introduced its restrictions before platforms, parents, or even some officials appeared to understand precisely how they would be enforced. Deactivation was to proceed gradually, with different companies applying different rules; YouTube subsequently said it had complied, while questions remained about other services and the reliability of age checks. Malaysia has so far focused on preventing new registrations rather than removing every existing account.

Vietnam is considering a different boundary. A draft decree would allow children under 16 to view social-media content through accounts registered in a parent’s or guardian’s name, but would prevent them from posting, commenting, or reacting. It would also require platforms to verify age, filter inappropriate material, and impose a daily one-hour limit on online gaming.

Vietnam’s proposal reflects a wider Asian preference for controlling particular forms of participation rather than simply drawing one line between access and exclusion. It also raises a separate concern: child protection rules can provide authoritarian governments with another justification for expanding control over online expression.

Asia, therefore, may provide a more revealing test than Europe. Indonesia and Malaysia must enforce restrictions across enormous and diverse digital markets. Vietnam is combining parental supervision with state control over participation. And India — one of the world’s largest social-media markets — is beginning to confront the same choices, with the state of Karnataka moving toward an under-16 ban even as some experts argue that restrictions should be paired with digital literacy rather than treated as a solution on their own.

The emerging question is not whether Asia will act, but which model its governments will choose — and whether protecting children becomes indistinguishable from giving the state greater authority over everyone online.

Signals to Watch

• Early compliance data from Indonesia and Malaysia showing how many underage accounts platforms are actually identifying and removing.

• Evidence that younger users are migrating to messaging apps, gaming platforms, or smaller services that fall outside the restrictions.

• Changes to Vietnam’s draft rules that broaden enforcement beyond child safety into tighter controls on online expression.

• Any move by New Delhi to turn India’s state-level experiments into a national policy.

The strike drew attention because of the deaths, but an equally consequential fact was procedural: American and partner air defenses did not stop everything. In a war fought substantially over the question of how much a modern, layered defense can intercept, something got through at the moment it mattered.

A single number has dominated public discussion of the air war between the United States, Israel, and Iran: 95 percent. It has been used, imprecisely, to describe two different things.

In March, Defense Secretary Pete Hegseth told reporters that Iran’s one-way drone attacks had fallen “95 percent” since the war began on February 28, with missile launches down roughly 90 percent. Gen. Dan Caine, chairman of the Joint Chiefs, gave similar figures in separate briefings that week, attributing the decline to strikes on Iranian launchers, storage sites, and production lines.

Those numbers describe how much less Iran was firing — a measure of degraded Iranian output, not of how many incoming weapons were being stopped. A War on the Rocks analysis published in April made the same point: the metric answers what Iran is currently doing, not what it retains the capacity to do, and a lull driven by conservation looks identical, from the outside, to a lull driven by destruction.

The interception question is a separate one, and it has been answered less precisely. The Center for Strategic and International Studies (CSIS) researchers estimated in late March that reported interception rates across Gulf states ran from 80 to 90 percent — comparable, they noted, to the best rates Ukraine has achieved against Russian drone and missile salvos. A Jewish Institute for National Security of America report put the figure above 90 percent theater-wide.

Washington has never published a single cumulative interception rate for the war; CENTCOM has instead released after-action figures for individual engagements, which establish that defenses performed well in those specific encounters but cannot be summed into a reliable theater-wide percentage.

Gulf governments have sometimes been more precise about individual weapon types, and the detail is instructive. Early in the war, Qatar reported intercepting the overwhelming majority of ballistic missiles fired at it but a much lower share of drones — illustrating why combining different weapons into one rate obscures more than it reveals. By late July, as Iranian strikes resumed against Jordan, Kuwait, and Bahrain, Jordan’s military said only that it had intercepted ten missiles in one attack, without stating how many were fired in total or how many reached the ground.

“Five percent,” in other words, is shorthand, not an audited total. Whether the share reaching a target is five percent or fifteen, the operational problem for the defender is the same: identifying which interceptors were expended, which sensors and radars were disabled or degraded, how long batteries took to reload under sustained fire, and whether the same performance could be sustained across weeks rather than days.

What Beijing Can Now Measure

None of this is news to the PLA in the abstract. Chinese planning for a Taiwan contingency has assumed for years that an opening phase would combine missiles, drones, cyberattacks, and strikes on the bases and logistics nodes that would let the United States intervene. What the Iran war supplies is something a simulation cannot: evidence of what actually happens when a real, integrated American-led air defense network is pressed for months rather than modeled for an afternoon.

Two data points from the war are likely of particular interest in Beijing. First, the CSIS estimate — since revised downward further — that the United States had by late July expended 65 percent of its most modernized Patriot interceptors and a similar share of its THAAD inventory, dropping from roughly 2,200 top-tier Patriots and 452 THAAD missiles before the war to fewer than 830 and 280, respectively. Second, the recurring observation, made by multiple outlets covering the war, that Iran’s cheapest weapons — a Shahed-type drone costs a small fraction of the interceptor fired at it — have been the most disruptive precisely because they force defenders to keep expending expensive stock against inexpensive threats.

The lesson available to the PLA is not the simple one — that mass produces leakage. Any planner already assumes that. The more useful lesson is that directed leakage compounds. Rather than spreading a strike package evenly across many targets, a campaign that concentrates opening attacks on the small number of radars, command links, and interceptor batteries that let a wider network function can turn a high interception rate into a declining one — with each successful strike making the next one cheaper.

The interceptor-depletion numbers coming out of the Iran war are, in effect, a real-world demonstration of how fast that arithmetic moves once a defender is forced to sustain a high intercept rate for months rather than days.That gives the “five percent” pattern its strategic weight for Beijing.

But the traffic runs both directions.

Why the Analogy Doesn’t Prove What It Looks Like It Proves

American and Japanese forces, in a Taiwan scenario, would be attacking the same category of Chinese systems — reconnaissance satellites, over-the-horizon radar, command links, and launch units — that the PLA would need to sustain its own salvos. A strategy built around exhausting an opponent’s defense only works if the attacker can also protect its own targeting and launch chain while under reciprocal pressure.

The Iran war offers no evidence either way on that question, because Iran has not had to fight for its own ability to see and strike while simultaneously being struck. The United States and Israel have had largely uncontested air superiority over Iranian territory for most of the campaign — a condition no one expects to hold for China over the Taiwan Strait.

The two conflicts also differ in what they are trying to accomplish, which matters for how much force is “enough.” Iran is attempting to impose cost and preserve leverage in ceasefire negotiations, not seize and hold territory. China’s requirements in a Taiwan scenario — isolating the island, defeating its armed forces, sustaining operations across a contested strait, and preventing the United States and Japan from reversing early gains — begin roughly where Iran’s campaign stops. A missile and drone barrage that damages bases and depletes interceptor stocks says little about whether an attacker can convert that damage into the sustained, multi-domain campaign a Taiwan contingency would require.

Finally, the political effect of the Iran war’s strikes on U.S. facilities has been to complicate American operations, not to change American objectives. The U.S. servicemen killed and wounded in the region have not visibly moved Washington toward ending the campaign; if anything, U.S. officials have continued to describe an intensifying air campaign against Iranian targets in the weeks since. A Chinese campaign against Taiwan and U.S. regional bases could just as easily harden American and allied resolve as weaken it — and a strategy premised on inflicting just enough damage to produce a quick political capitulation becomes considerably less attractive if it instead produces a longer war that Beijing cannot contain on favorable terms.

The Measure That Will Actually Matter

The Iran war is giving Chinese planners something genuinely valuable: real data on how a modern, American-led air defense network performs, degrades, and reconstitutes under sustained pressure, and evidence that concentrated attacks on the right nodes can turn a strong interception rate into a weaker one over time. It is not giving them proof that such a window, once opened, can be kept open long enough to force a political outcome — or that China’s own targeting and launch systems would survive the reciprocal pressure a Taiwan conflict would put on them.

The more telling numbers from Jordan, Kuwait, and Bahrain are not really about the five percent that gets through in the first weeks. They are about what happens afterward: how quickly a damaged runway reopens, how fast a destroyed radar is replaced, how quickly interceptor stocks are resupplied, and whether the attacker can strike again before the network recovers.

That is the test a five-percent solution would eventually have to pass — and the Iran war, so far, has only shown the first half of it.

Many of the movements across Asia have emerged from a similar set of grievances: scarce employment, unequal access to education, corruption, and blocked social mobility. Yet these frustrations are not generating a single regional political model.

The question is therefore no longer whether younger Asians are becoming more willing to protest. It is what kind of power they are acquiring, whether that power survives its first victory, and whether governments can sustain its hold without addressing the conditions that produced it. These are the signals worth watching in the weeks ahead.

1. India’s Cockroach Movement Wins a Resignation—but Not Yet Reform

On Saturday, July 25, Indian Education Minister Dharmendra Pradhan resigned following weeks of protests over examination failures, alleged corruption, and the shortage of employment opportunities for educated young people. It was the most visible victory yet for the loosely organized Cockroach Janta Party — a movement that began in satire but has become a vehicle for broader anger about whether India’s education and employment systems still offer a credible path to advancement.

The immediate trigger was the cancellation of a medical-school entrance examination after questions were leaked. Students had spent months or years preparing for the test, often supported by families paying substantial fees to coaching centers and private tutors. Reports that several despondent students committed suicide deepened the outrage and turned what might have remained another examination scandal into a national protest movement.

The cockroach became its symbol after India’s chief justice compared unemployed young people to the insects. Protesters reclaimed the image, appearing in masks and circulating jokes, cartoons, and videos that portrayed young Indians as numerous, unwanted, and impossible to eliminate.

Humor and satire have been central to the campaign, but the grievances behind them are not new. Examination leaks, teacher shortages, poor educational standards, and intense competition for a limited number of government jobs have shaped political discontent in India for decades. By the late 1990s, thousands of graduates could be found competing for a handful of public-sector positions, while many others continued studying largely because few alternatives existed.

Researchers have described this condition as “timepass”: young people accumulating degrees, preparing for examinations, and waiting for employment that might never materialize. Some referred to themselves as “Generation Nowhere.” Their frustration reflected not idleness but the absence of a reliable route from education into stable work.

Pradhan’s resignation demonstrates that the movement can impose political costs — strong enough even to push Prime Minister Narendra Modi, a leader who rarely submits himself to unscripted questioning, into recording a cellphone video on July 23 appealing directly to students to give the government time to act. However, it does not yet show that the government is prepared to address the deeper problem. Replacing a minister may relieve immediate pressure, but it will not create jobs, improve schools, regulate coaching centers, or ensure that examinations are conducted fairly.

The Cockroach movement has succeeded because it created a political identity to the painful reality that advancement depends less on merit than on money, connections, or luck — a recognizable identity that did not require protesters to join an established opposition party. The movement emerged outside that machinery and forced a cabinet resignation without contesting an election, controlling a formal organization, or presenting a comprehensive political platform.

Signals to watch:

• The first is obvious: whether Pradhan’s resignation leads to visible institutional changes. A clear timetable for protecting examinations, prosecuting those responsible for leaks, and compensating students whose time and money were lost would indicate that the government regards the protests as more than a temporary political crisis.

• The second will be whether the movement remains focused on examination integrity or expands into the larger questions of employment, education spending, and the regulation of coaching centers.

• Equally important will be the relationship between protesters and established political parties. If opposition groups succeed in absorbing the movement, the Cockroach campaign may become another instrument of conventional partisan competition. If organizers maintain their independence, India could see the emergence of a more durable form of issue-based mobilization capable.

2. Nepal’s Protest Generation Becomes the Opposition

On July 12, hundreds of protesters gathered outside Nepal’s main government complex carrying signs demanding an end to “atrocity against the poor” and greater respect for human rights. The demonstrations followed government-backed evictions of landless residents and the death of Ganesh Nepali, a 25-year-old ride-share driver who set himself on fire after municipal police wheel-locked his motorcycle during a dispute outside the passport office.

The protests would have been politically uncomfortable for any government. For Prime Minister Balendra Shah, they carry a sharper message. Less than a year ago, Nepal’s youth-led uprising helped remove the previous government. In March, voters gave Shah’s Rastriya Swatantra Party 182 of the 275 seats in parliament, turning the anger of the streets into one of the strongest electoral mandates in Nepal’s recent history. Shah took office promising clean government, faster public services, and a break with the political practices of the parties he displaced.

The young people now returning to the streets are not confronting the old establishment. They are challenging a government brought to power in their name.

That development complicates the familiar story of generational political change. Taking power did not end the movement that produced it. It divided that movement between those working within the new government and those who believe their role is to hold it accountable.

Some youth activists acknowledge that Shah has begun changing public services and pursuing corruption cases. However, the complaints are not simply that reform is moving too slowly, but whether the methods used to impose change resemble those of the political order the movement rejected in March.

Independent assessments have found that many commitments remain incomplete or have produced little visible effect on daily life. Even lawmakers from Shah’s own party are warning that the government cannot depend indefinitely on the prime minister’s popularity.

The return of youth protesters may not just reflect disappointment but something more durable. The movement that helped elect Shah has not surrendered its political identity to him. Some of its members appear willing to oppose the government they helped create rather than treat generational solidarity as a permanent claim on their loyalty.

Signals to watch:

• As with the India piece, the first signal is obvious: whether the protests broaden into a sustained campaign over transparency, jobs, and the government’s use of police power — a development that would show that the unrest is becoming a judgment on Shah’s governing model rather than a response to one incident.

• The government’s treatment of youth activists will be equally important. An independent investigation, compensation, and changes to municipal enforcement would suggest that Shah recognizes the protesters as part of the constituency that placed him in office. Arrests, restrictions, or efforts to dismiss them as agents of the old parties would indicate that the administration increasingly regards independent youth mobilization as a threat.

• Nepal’s generational transfer has already occurred. The question now is whether it produces a new political system or merely a younger governing class. The willingness of Nepal’s young people to return to the streets against their own government may provide the first indication of which direction it is taking.

3. Bangladesh’s Protest Generation Fights Over Its Inheritance

Two years after the student-led uprising that drove Sheikh Hasina from power, the movement’s political heirs are again marching through Bangladesh. The National Citizen Party (NCP) has organized thirty-six days of rallies, memorial events, and public meetings extending from July 1 to August 5, the anniversary of Hasina’s fall. Its leaders say the uprising removed an authoritarian government but failed to deliver the institutional reforms for which protesters fought.

That claim reflects the peculiar position in which Bangladesh’s protest generation now finds itself. Young activists demonstrated that they could bring down a government that had ruled for fifteen years. They helped establish an interim administration, created a political party, and placed constitutional reform before the electorate. Yet when national elections were held in February, the NCP won only six of 300 directly-elected parliamentary seats. The Bangladesh Nationalist Party (BNP), one of the established forces the uprising was supposed to weaken, returned to power with a commanding majority.

In other words, the students won the revolution; the old political system won the election.

The reasons offer a warning for youth-led movements elsewhere in Asia. Protest organization and electoral organization require different capabilities. A movement can mobilize students through social media, shared outrage, and a clear demand for the removal of a government. Winning constituencies requires local candidates, money, party workers, policy positions, and relationships extending well beyond university campuses and urban centers.

The BNP faced no comparable burden.

The dispute has now shifted to the July National Charter, the package of constitutional and institutional reforms developed during the interim period and endorsed in the February referendum. NCP leaders accuse the BNP government of delaying or diluting those commitments. On July 18, party convener Nahid Islam warned that failure to implement the charter could produce another mass uprising.

The warning exposes the movement’s central dilemma. The NCP is now a parliamentary party, however small, operating within the institutions it once challenged from the streets. Yet, its principal source of leverage remains the possibility of returning to mass mobilization. It is attempting to function simultaneously as a conventional opposition party and as the guardian of an unfinished revolution — a strategy that carries risks.

Repeatedly invoking another uprising may preserve pressure on the government, but it can also suggest that electoral outcomes remain conditional on the approval of those who led the protests. The BNP won the election and possesses its own democratic mandate. The NCP cannot claim exclusive ownership of the political meaning of July simply because its leaders played a central role in the movement.

Bangladesh thus represents a different generational outcome from both India and Nepal. India’s protesters have demonstrated that they can extract concessions without forming a party. Nepal’s younger political forces converted protest into a governing majority. Bangladesh’s student leaders entered formal politics but discovered that the electorate was willing to honor their uprising without entrusting them with the state.

Signals to watch:

• The first will be whether the dispute over the July Charter produces a negotiated implementation timetable. Agreement on specific reforms, deadlines, and parliamentary procedures would indicate that the uprising’s legacy is becoming embedded in institutions rather than preserved mainly through protest rhetoric.

• The second will be whether the NCP uses the coming local elections to build the organization it lacked in February. Continued weakness would leave it dependent on alliances with larger parties and periodic street mobilization.

• The party’s relationship with Jamaat-e-Islami — Bangladesh’s long-established Islamist party — will be equally important. A deeper alliance could provide organizational reach while blurring the generational break the NCP once claimed to represent. Moving away would test whether it can survive without Jamaat’s political machinery.

• The final signal will be whether the July commemorations draw participants beyond the NCP’s existing base. Large, diverse crowds would suggest that the reform agenda retains national force despite the party’s poor electoral showing.

Takaichi is not unique. Donald Trump has repeatedly boasted that he needs little sleep, though recent images of what appears to be the president nodding off in meetings have called that claim into question. Business profiles still celebrate chief executives who begin answering emails before dawn and work past midnight. Few phrases better captured this strain of the late-twentieth-century work ethic than “I’ll sleep when I’m dead.” It appeared in songs, business books, and corporate offices alike.

Sleep deprivation among the people responsible for consequential decisions is not a wellness problem. It is a readiness problem — the same category of risk that aviation regulators, hospital accreditors, and military planners have already spent decades managing. A society that dismisses that risk, or lauds those who conspicuously court it, is degrading a foundational element of its own resilience. And the leaders who wear sleeplessness as a badge of honor are not just modeling something for themselves — they are normalizing exhaustion down through every institution that watches how the people at the top behave.

The Science

The belief that judgment and rest are unrelated collides with the research. Over the past two decades, sleep has moved from a subject confined to medicine into one that intersects with economics, education, transportation, military readiness, and public policy. Matthew Walker’s bestselling Why We Sleep brought much of that research to a broad audience, but it drew on decades of work showing that chronic sleep restriction degrades vigilance, learning, memory, mood, and complex judgment.

Sleep performs vital housekeeping in the brain — clearing metabolic waste, supporting cellular repair, consolidating memory — and does real work in the regions associated with abstract thinking and creativity. Deprivation, by contrast, is linked to impaired judgment and higher long-term risk of heart disease, stroke, and Alzheimer’s.

Catching up is also something of a myth: extra sleep can blunt some short-term effects, but chronic sleep debt does not simply erase over a weekend.

The Institutional Precedent

Several organizations built around sustained judgment have already responded to this evidence, formally. Federal Aviation Administration rules and guidance combine limits on flight and duty periods with formal fatigue-risk management systems tailored to operational conditions. The Accreditation Council for Graduate Medical Education limits clinical and educational work to 80 hours a week averaged over four weeks and generally caps continuous scheduled clinical assignments at 24 hours, with limited additional time for transitions and education. Professional sports organizations have invested in sleep and recovery because marginal improvements can translate into measurable competitive advantage.

What these institutions share is a recognition that judgment is a perishable resource, not a fixed trait — and that protecting it requires rules, not just individual willpower.

The Intelligence Lesson

The U.S. intelligence community rarely comes up in discussions about sleep, but it offers a useful case study, because intelligence work alternates between long stretches of routine monitoring and sudden crises that demand extraordinary effort. During my government career, I encouraged employees to leave after completing a full workday — not because the mission lacked urgency, but because I knew urgency would eventually arrive, and I wanted people to have something left when it did.

When those moments came, stamina mattered as much as discipline or sacrifice. There was little value in letting people exhaust themselves during quiet periods merely to signal commitment to the mission. I was not always the best role model on this myself — I would often delay my own departure even after the day’s work was done, uneasy about leaving after only nine hours.

Organizational culture exerts a powerful pull. The last car in the parking lot quietly becomes a symbol of dedication, and leaving at a reasonable hour can feel like an admission — however inaccurate — of insufficient commitment.

That dynamic is the real danger in stories like Takaichi’s. It is not that one exhausted leader will necessarily make one bad call. It is that valorized exhaustion at the top gives every subordinate institution permission to treat burnout as the price of seriousness — in exactly the fields, like intelligence, cybersecurity, and finance, where degraded judgment is most expensive.

The Military Distinction

The military offers an instructive contrast, because it holds two seemingly opposite truths at once. Army Ranger School, Navy SEAL training, and other elite programs deliberately expose candidates to severe fatigue. The purpose is not to celebrate sleeplessness — it is to prepare people for a battlefield where sleep deprivation will happen anyway, and to teach them its effects on their own judgment under controlled conditions.

Operational practice reflects the other half of the lesson. Military aviation, naval operations, and extended command posts manage crew rest, staffing, and rotation with real discipline, because fatigue creates risk. The military distinguishes sharply between training people to endure degraded conditions and choosing degraded conditions as a routine operating model. Civilian institutions routinely blur that distinction. A corporation is not a battlefield, and an ordinary workweek is not a crisis. If exhaustion becomes the normal condition of routine operations, an institution has less reserve left for the moment it’s genuinely needed.

The pandemic offered a smaller-scale version of the same lesson. Remote work weakened collaboration and mentoring in some settings, but controlled studies found that well-designed hybrid arrangements could preserve performance while improving retention, and earlier randomized research found a 13 percent productivity gain among remote call-center employees. The consistent finding across that research: output tracks job design, concentration, and recovery — not hours logged in a particular building.

The AI Wildcard

The stakes here extend beyond individual workplaces. Advanced economies compete through scientific research, software development, financial analysis, intelligence, and healthcare — sectors defined less by physical output than by the quality of human judgment. A software engineer who introduces a critical security flaw, or a financial manager who misjudges risk, can impose costs that dwarf the value of the extra hours worked to get there.

AI is likely to sharpen this problem rather than solve it. As AI systems absorb more routine analysis, pattern recognition, and summarization, human beings will increasingly be left holding the ambiguous decisions machines cannot resolve — precisely the judgment calls most vulnerable to frayed attention and weakened impulse control. AI may also make overwork less visible, extending the workday quietly through constant alerts, summaries, and automated tasking.

Organizations that expect human decision-makers to match AI’s speed and consistency are confusing technological capacity with biological reality. The comparative advantage of people will lie not in matching machine endurance, but in knowing when to slow down, challenge an answer, and notice what the data don’t show — none of which is available to a mind running on two hours of sleep.

The Bottom Line

None of this argues that long hours will, or should, disappear. Every profession has moments when sleep must yield to necessity — that surge capacity is real and worth preserving. The institutional task is to protect it for the moments that actually demand it, rather than spend it as a routine operating model.

Much of the modern workplace still treats time at the desk as a proxy for value, rewarding those who arrive earliest, leave latest, and stay reachable around the clock — even as the added hours quietly degrade the judgment those hours were meant to demonstrate.

For institutions that depend on sustained judgment — national security agencies, financial firms, technology companies, hospital systems — the practical takeaway is not a wellness memo. It is to borrow, deliberately, from the playbook aviation and medicine already wrote: formal fatigue-risk management, real limits on continuous duty, and leadership that treats rest as part of readiness rather than its opposite.

The strategic challenge for the coming decade will not be persuading people to work harder. It will be enabling them to think clearly when it counts — and protecting the cognitive capacity of the people making the hardest calls is not an employee benefit. It is an investment in national resilience.

That growing reliance on this source of information is producing a new set of problems for intelligence agencies. The challenge is no longer simply finding useful information in public view. It is developing data management systems to ingest and organize the information, avoiding parallel collection among various agencies within the national security community, and implementing timely and applicable tradecraft guidelines for its use.

OSINT’s appeal has always rested on its openness, but its emergence in many cases as the prime source of reporting in analytic assessments — CIA Director John Ratcliffe told the House in 2025 that it was the most cited material in the President’s Daily Brief — has come from its unmatched scale and speed. Commercial satellites, artificial intelligence, social media, shipping records, corporate databases and other public or purchasable sources can now be fused into finished intelligence that once relied solely on government collection systems. The result is a discipline that is moving closer to the center of intelligence.

This week’s Signals to Watch focuses on this transition not as a retrospective, but to provide an update on how — and if — the institutions, rules, and habits built for a classified age are keeping pace with an intelligence environment increasingly centered and reliant on sourcing that is available to all of us.

1. Congress: OSINT Reform Without an OSINT Institution

The House Permanent Select Committee on Intelligence (HPSCI) this week approved an intelligence authorization bill that treats open-source intelligence less as a supporting activity and more as a community-wide capability requiring its own budgets, standards, and workforce, and broader oversight.

In 2024, the Biden administration’s intelligence-community OSINT strategy issued a comprehensive OSINT strategy for coordinated data purchases, integrated collection management, common platforms, and improved tradecraft and training standards. It also assigned implementation to a familiar collection of officials: the CIA director as OSINT functional manager, CIA’s Open Source Enterprise, a Defense Intelligence Agency manager, and an OSINT executive at the Office of the Director of National Intelligence.

The new House bill suggests that Congress is not convinced those arrangements have produced an integrated enterprise. It would require agencies to identify OSINT activities, expenditures, and personnel in their budget submissions. It calls for a community-wide framework covering collection, analysis, dissemination, training, and technology; an assessment of gaps in workforce, tools, data access, and analytic methods; stronger integration of OSINT into Defense Department collection management; and consideration of centralized services for sharing unclassified intelligence and commercially available information. It would also establish a chief artificial intelligence officer at ODNI to coordinate AI strategy across the intelligence community.

The provisions amount to another congressional endorsement of OSINT; HPSCI established an OSINT subcommittee only last year. But Congress, as with the intelligence community it oversees, has largely been reacting to a transformation already well advanced. Journalists, academics, and a growing private-sector intelligence industry have exploited open and commercially available information for years while government institutions have moved more slowly to adapt its structure and guidelines for regulating its collection, distribution, and application.

OSINT remains the only major intelligence discipline without an independent agency or organization primarily responsible for its performance. Signals intelligence has the National Security Agency. Geospatial intelligence has the National Geospatial-Intelligence Agency. Human intelligence has the CIA. OSINT is distributed across all of them, while the CIA serves as functional manager for a discipline whose value depends on wide dissemination beyond CIA’s own mission.

The issue of establishing a separate agency has been a recurring topic debate, but that has come from the outside of the IC and Congress. Ben Scott made the case in a 2024 Lawfare essay, arguing that a dedicated agency would “leave the IC better postured for the information age.”

The House bill does not resolve that problem. It improves visibility, directs common standards, and creates additional coordinating mechanisms, but it leaves budgets, personnel, and operating authority largely inside the existing agencies. Budget transparency could expose duplicative commercial-data contracts. Common dissemination services could make OSINT available to agencies that cannot access classified systems. Integrating OSINT into collection management could reduce the expensive habit of tasking sensitive systems to answer questions already resolvable in the open.

Signals to watch: The first is whether these House provisions survive negotiations with the Senate and reach the final authorization bill. The second is what Congress learns once agencies must separate OSINT spending and personnel from larger intelligence accounts. Significant duplication — or unexpectedly modest investment — would strengthen the argument that the discipline requires more than better coordination.

The most consequential signal will be whether the proposed framework produces a genuine service of common concern with its own funding and authority. If it does not, Congress may eventually confront the question the bill avoids: whether OSINT can become a core intelligence discipline without an institution built to lead it.

2. Commercial Satellites: When Transparency Becomes Targeting Support

Commercial satellite imagery (IMINT) has been an equally transformative source of information outside the intelligence community. It has enabled journalists to document collateral damage, corroborate reports of civilian casualties, and compare government claims against visible evidence. But the same open-source collection can track military deployments, locate air defenses, and reveal the layout of troop housing and equipment caches.

The Netherlands-based investigative group Bellingcat provided an early demonstration of the power of commercially IMINT and other open sources in its investigation of the 2014 downing of Malaysia Airlines Flight 17 over Ukraine. Its researchers combined photographs and videos posted online with geolocation and satellite imagery to trace the missile launcher that fire at the aircraft through separatist-controlled territory and link it to a Russians air defense unit, undermining Moscow’s denials of involvement.

Its dual-use potential is no less evident in the Iran war. The Chinese geospatial company MizarVision publicly posted satellite imagery and AI-assisted analysis tracking U.S. aircraft, naval forces, and military installations in the Gulf. U.S. intelligence concluded that Iranian forces used some of that material to help identify potential targets. The Trump administration later sanctioned MizarVision and two other Chinese companies, alleging they facilitated Iran’s strikes on American and allied facilities.

The United States responded to the same danger from the opposite direction. At Washington’s request, Planet Labs indefinitely withheld recent imagery of Iran and surrounding conflict areas, allowing release only in selected, public-interest cases. Other American providers also tightened access. The restrictions followed an earlier decision to delay publication by 14 days.

The immediate security logic is understandable. But the Iran experience also exposed the limits of controlling OSINT through American companies. Restricting Planet Labs may prevent an ordinary subscriber from receiving fresh imagery. It does not prevent China, Russia, or another government from using its own satellites, purchasing imagery through foreign providers, or combining older pictures with flight tracking, social media, and other commercial data.

Signals to watch: Whether the United States replaces ad hoc requests with formal standards distinguishing near-real-time operational imagery from delayed material used for accountability. Another indicator will be whether American restrictions push customers toward foreign providers, reducing U.S. influence over the commercial imagery market without meaningfully denying information to adversaries.

The most important signal will be whether wartime restrictions expire when their operational justification does. Do such restrictions apply during a White House-described “cease fire”? If temporary controls become the normal response whenever commercial imagery complicates a government’s preferred account, OSINT will remain open in principle but increasingly selective in practice.

3. The U.S. Navy: AI Is Only as Useful as the Data It Can Reach

The Navy last month published a strategy to build what it calls an “AI-first” fleet, promising to “weaponize” data so commanders and sailors can make decisions faster than an adversary. The Navy wants artificial intelligence integrated across operations, logistics, maintenance, and planning. But the strategy also exposes a problem that has followed the intelligence community’s effort to modernize OSINT: AI cannot exploit information that agencies cannot acquire, organize, share, or trust.

Much of the data with greatest operational value no longer begins inside a classified collection system. As noted earlier, commercial IMINT can reveal changes at a naval base. Shipping databases can trace unusual vessel movements. Corporate records can identify suppliers and ownership networks. Social media, weather data, port records, and commercially collected location information can fill gaps that exquisite national systems may not cover continuously.

The Navy’s challenge is therefore not simply to build better models. It is to connect those models to information scattered across classified networks, commercial contracts, operational databases and publicly available sources — without allowing uncertain or manipulated material to move directly into military decisions.

That distinction matters because speed can magnify error as readily as insight. An automated system may identify a pattern across hundreds of sources more quickly than an analyst, but it may also count the same report several times, confuse commercial activity with military preparation, or incorporate data deliberately placed in the open to deceive it. The faster the system produces an answer, the easier it becomes for users to mistake technical fluency for analytic confidence.

The Navy strategy and the House intelligence bill are therefore addressing different sides of the same problem. Congress is asking who should organize OSINT and AI across the intelligence community. The Navy is trying to make the resulting data operational before those institutional questions have been settled.

Signals to watch: Whether the Navy’s implementation plan explains how publicly and commercially available information will enter AI-enabled operational systems. The important indicators will be common data standards, access to commercially acquired information, mechanisms for identifying duplicated or manipulated sources, and rules requiring analysts to show the evidence behind machine-generated judgments.

Also watch whether the Navy builds its own data architecture or relies on services established across the intelligence community and Defense Department. Separate service systems may deliver progress quickly, but they could reproduce the fragmentation that the House bill is trying to correct.

The implications extend well beyond Ukraine. Governments are committing hundreds of billions of dollars on aircraft, ships, armored vehicles, and missiles that are meant to last for years, if not decades — while some of the assumptions behind them are being tested on an active battlefield.

A recent Wall Street Journal examination framed the central debate as one between “revolution” and “evolution”: Whether new technologies are overturning the established way of war, or simply accelerating changes already underway? That framing is a useful starting point, but it collapses two different arguments into one axis, and leaves out a third that may matter more than either.

If you put aside the question of whether warfare has changed, a separate question emerges: Even if we knew the answer, would it decide who wins the war? Sometimes it doesn’t. That possibility — that endurance may outrank cleverness — deserves to be treated as its own argument, not folded into either side of the WSJ’s framing.

And running underneath all of it is the inconvenient truth of speed: no matter which side of the revolution-versus-evolution debate is right, and no matter whether endurance ultimately decides the war, how quickly a military can learn and adjust may determine whether it gets the chance to find out.

Each interpretation points toward different choices about military spending, industrial production, and force structure. A lesson drawn too quickly could leave a country preparing for the wrong war. One dismissed too easily could leave it dependent on weapons and institutions that no longer work as intended.

No single battlefield development will settle any of these arguments on its own. But each is producing evidence that can be watched over time.

1. The Revolution View: Are Expensive Weapons Becoming Too Vulnerable?

The boldest interpretation is that warfare is undergoing a technological revolution. Inexpensive drones, widely available sensors, and increasingly autonomous weapons are changing not only how armies fight, but which weapons remain worth buying.

The argument begins with an imbalance. A drone costing several thousand dollars can help destroy a tank worth millions. A swarm can force an opponent to expend scarce air-defense missiles that cost far more than the incoming weapons. Commercial satellites, surveillance drones, and electronic sensors can expose troop concentrations that previous generations might have concealed until an attack was underway.

Movement has consequently become more dangerous. Tanks, artillery batteries, supply convoys, and command posts cannot assume they will remain hidden long enough to prepare an assault. According to CIA Director John Ratcliffe, a Russian recruit arriving on the front lines in Ukraine is expected to be killed or wounded within 20 to 30 minutes.

The safety of rear areas is also disappearing. Ukraine has struck Russian airfields, refineries, and military installations hundreds of miles from the front. Russia has used missiles and drones against Ukrainian energy infrastructure, factories, and cities. Capabilities once confined to major powers are becoming more numerous and more accessible.

The deeper claim is that military power will increasingly reside in networks rather than individual weapons. Sensors find targets, software sorts the information, weapons carry out the attack, and factories produce replacements. A force built around a small number of sophisticated platforms may struggle against an opponent able to manufacture, lose, and replace thousands of cheaper systems. The Pentagon’s Replicator initiative represented an early bet on this logic, designed to field thousands of relatively inexpensive autonomous systems.

Ukraine has not settled the argument, however. Drones have made traditional forces more vulnerable and have helped prevent large breakthroughs. They have not consistently enabled either side to penetrate prepared defenses, advance rapidly, and hold substantial territory on their own.

Signals to watch: The clearest test will be whether drones and autonomous systems begin doing more than finding and destroying targets — if largely uncrewed formations can breach defenses and support a sustained advance, the case for a military revolution will become much stronger. A second indicator will be whether major powers begin canceling or substantially reducing purchases of tanks, crewed aircraft, or large warships, rather than simply adding autonomous systems on top of the forces they already have. A third signal will come from defensive innovation: if electronic warfare, deception, and uncrewed escorts restore the ability of conventional forces to move and concentrate safely, the revolution argument will weaken; if those defenses remain too costly or unreliable, the balance will keep shifting toward cheaper weapons produced in large numbers.

2. The Continuity View: Technology Has Not Repealed the Fundamentals of War

The second view accepts that drones and sensors have changed the battlefield, but rejects the claim they have made traditional forces obsolete. Weapons can destroy and delay an enemy, but armies still need people and equipment to seize territory, protect supply routes, and hold what they capture.

Ukraine offers substantial evidence for this view. Britain’s defense secretary said in October 2025 that drones accounted for approximately 80 percent of casualties in the war, yet neither Russia nor Ukraine has converted that lethality into a sustained breakthrough.

Ukraine’s failed 2023 counteroffensive illustrates the problem. According to a U.S. Army study, Ukrainian forces advanced approximately 20 kilometers along one principal axis while losing 518 vehicles, including 91 tanks and 24 engineering vehicles. Minefields, artillery, drones, and attack helicopters prevented Ukrainian units from opening and exploiting a path through Russian defenses.

The lesson was not that tanks had become useless. It was that breaching a defended line still required combined-arms warfare: engineers to clear mines, artillery to suppress enemy positions, air defenses to protect the attackers, infantry to secure the ground, and armored vehicles to carry forces forward, all operating together.

That helps explain why armies have not abandoned heavy weapons. The U.S. Army’s Transformation Initiative, for example, envisions combining modern drones and counter-drone defenses with long-range missiles, aircraft, and armor, on the wager that the tank will remain necessary but must become lighter, better protected, and easier to update.

The continuity view is most persuasive when it argues that the fundamentals endure but their execution must change. Its weakness is that continuity can become an institutional defense of programs already funded and weapons already familiar. Adding drones to an armored brigade will not solve the problem if the brigade remains too concentrated, slow, or visible to reach the battlefield.

Signals to watch: The strongest evidence for continuity would be a successful mechanized offensive in which infantry, armor, engineers, artillery, electronic warfare, and drones break through prepared defenses and then sustain a rapid advance. A second signal will be whether new protections allow armored forces to regain freedom of movement, preserving the value of tanks without restoring the large armored columns of earlier wars. A final indicator will be how governments divide their budgets: Continued purchases of tanks, aircraft, and warships alongside investment in autonomous systems would amount to a bet on continuity.

3. The Attrition View: Whoever Outlasts the Other Side Wins

A third interpretation argues that the first two debates, for all their disagreement, share an assumption worth questioning: The outcome of the war will be decided primarily by which weapons or tactics prove superior on the battlefield. The attrition view holds that the more basic determinant is simpler and older — who can keep replacing lost soldiers, shells and equipment for longer. War becomes a contest of endurance and supply as much as of cleverness.

There is a substantial case for this reading of Ukraine. One strategic assessment frames the war as running on four simultaneous clocks: the pace of events on the battlefield itself, the pace of industrial production on each side, the patience of the governments supporting Ukraine, and the internal stability of the Russian government.

This view also reframes certain tactics as economic rather than technological in character. Ukraine’s long-range strikes on Russian oil terminals and refineries, which analysts estimate have cut Russian export capacity by roughly 40 percent, are not best understood as evidence of a new way of fighting. They are a deliberate campaign to raise the cost of the war for Russia to sustain, carried out one barrel at a time — not a single decisive blow, but a steady erosion of the resources needed to keep fighting.

The risk in this view is fatalism: Treating industrial capacity and population size as destiny risks understating how much better tactics and technology can reduce the losses either side actually suffers per engagement, which in turn affects how long either side can hold out.

Signals to watch: The clearest indicator will be the trajectory of recruitment and munitions production on both sides — whether Russia and Ukraine are able to sustain or grow their output, or whether either is beginning to show real strain. A second signal will be the durability of outside support to Ukraine, particularly whether financial and military assistance from its partners holds steady through the coming budget cycles or begins to erode. A third will be visible cracks in either side’s capacity to continue: Labor shortages, currency pressure, conscription resistance, and falling public support for the war effort would all suggest that endurance, not technology, is becoming the deciding factor; all four cracks may be forming in Russia.

The Separate Issue: Speed of Learning

Regardless of which camp turns out to be right, there’s a second question that matters just as much: How quickly can a military learn from what’s working and what isn’t, and get that lesson into the hands of soldiers?

Think of it like a smartphone versus a car. A car is designed once and stays mostly the same for years. A phone gets a software update practically every week. Old-fashioned military purchasing treats weapons like cars — designed to a fixed spec and used for decades without much change. Ukraine has been forced to treat weapons like phones, because the enemy adapts to any weapon within weeks.

Ukraine’s Brave1 program supports roughly 2,500 defense companies and over 5,000 products. Frontline units can request new gear directly and give feedback that gets built into the next version almost immediately. That fast feedback loop — battlefield problem, quick fix, redeployment — is arguably a bigger advantage than any single weapon.

The U.S. is trying to copy this. The Government Accountability” Office reported in 2026 that major Pentagon weapons programs “take over 12 years, on average, to deliver a first working version — far too slow for a battlefield that changes month to month. Programs like the Army’s “Transformation in Contact” and the U.K.’s “Project Beehive” — a fleet of low-cost, upgradeable naval drones — are attempts to shorten that cycle.

It’s not clear peacetime militaries can fully copy this. Ukraine’s fast learning happened under the pressure of national survival, with constant, large-scale feedback from real combat. Training exercises can’t fully recreate that. Speed also has downsides: a flood of constantly-changing gear can create supply and compatibility headaches, and units may get rewarded for flashy, easily-measured kills rather than for less visible things like logistics and communications, which matter just as much.

Signals to watch: The first will be whether programs such as Transformation in Contact and Project Beehive produce equipment that is repeatedly modified after use. The second will be whether Ukraine’s innovations spread beyond individual teams, since a system that works for one experienced drone unit is not yet a military transformation. The final signal will be how long battlefield advantages endure: If each new drone, jammer or targeting method is quickly neutralized, the case for treating speed itself as the decisive capability will grow stronger.

The Bottom Line

The Wall Street Journal’s revolution-versus-evolution framing captures part of the disagreement but not all of it. Once the question of what determines victory is separated from the question of how fast any answer can be acted upon, the debate resolves into three substantive arguments — over technology, over the enduring fundamentals of combined-arms warfare, and over sheer endurance — running alongside one cross-cutting argument about institutional speed.

None of the four will be settled by any single battlefield development. But together, they define most of what is actually worth watching for as the war, and the arguments about it, continue

Late yesterday afternoon, Just Security—an online publication focused on national security, law, and policy, for which I have written a number of articles on the U.S. intelligence community—asked whether I could produce a quick-turnaround initial assessment of President Trump’s election-intelligence speech and the documents released by the White House.

That article was published around 1 a.m. this morning.

After spending much of the night working through the president’s speech and the accompanying declassification release, I discovered I had just enough brainpower left to write this note—not to give Ukraine the final edit it deserves.

So today’s post instead points you to my Just Security article, “An Initial Readout of Trump’s Election-Interference Speech.” The piece examines what the declassified intelligence appears to show, what it does not yet establish, and why understanding the intelligence process matters as much as reading the documents themselves.

I will get to finishing a final edit of the Ukraine article—after I’ve had a little more sleep—and post it on Tuesday’s edition of the newsletter. Here is a quick preview of that article:

• The war in Ukraine is becoming more than a battlefield—it is the world’s largest laboratory for testing the future of warfare. But the lessons are proving far less straightforward than many headlines suggest.

• Is Ukraine demonstrating that inexpensive drones are making tanks and other traditional weapons obsolete? Or is it showing that armies still win through the enduring principles of combined-arms warfare? Or is the real lesson that the military able to learn and adapt fastest will hold the decisive advantage?

• Three competing schools of thought have emerged from the war, each implying a different future for military doctrine, procurement, and force design.

Western media noted the visit. Proliferation specialists studied the images. There was no comparable reaction from the White House — no sustained presidential warning, no push to make North Korea the center of America’s nuclear debate.

That silence is the story, not an absence of one. It isn’t that Trump has forgotten Kim. After the G7 summit in France, South Korean President Lee Jae Myung said Trump told him it was “time to pay attention” to North Korea and seemed open to resuming dialogue. Lee said he urged a phased approach: halt production of additional nuclear material, prevent weapons transfers, slow further ICBM development, and treat denuclearization as a longer-term goal rather than a precondition.

That proposal is serious. It has also gone nowhere near the center of U.S. policy, which is still organized around Iran — even though North Korea is the country that already has the weapons and Iran, by U.S. intelligence’s own repeated account, does not.

The gap between the two files runs on three asymmetries: what Washington says, what the intelligence community has found, and what each program has actually built.

The first asymmetry is rhetorical. Trump’s language toward the two programs has moved in opposite directions since he returned to office. In March 2025, then-Director of National Intelligence (DNI) Tulsi Gabbard told Congress the intelligence community still assessed Iran was not building a weapon and that the now-deceased Iranian supreme leader Ali Khamenei hadn’t reauthorized the program suspended in 2003.

In the days leading up to the U.S. airstrikes on three Iranian nuclear facilities, Trump contradicted his own DNI directly, calling Iran “very close” to a nuclear weapon and dismissing the community’s finding. North Korea got the opposite treatment. Weeks earlier, Trump had called it a “big nuclear nation,” emphasized his “very good relationship” with Kim, and said he’d “probably do something with him at some point.”

The second asymmetry is evidentiary, and it predates Trump entirely. From DNI Dan Coats in 2019 through Gabbard in 2025, the U.S. intelligence community has repeatedly assessed that Iran was not undertaking the activities necessary to produce a weapon — even while flagging an enriched-uranium stockpile unprecedented for a non-nuclear state and rising domestic debate inside Iran about weaponization. On North Korea, the same years of assessments describe a state retaining its full WMD capability, unlikely to disarm because its leadership treats nuclear weapons as the guarantee of regime survival, and by 2026, actively expanding that arsenal to deter the U.S. and hold South Korea at risk.

Three administrations, two parties, one consistent finding on each side. Political rhetoric keeps compressing that into a simpler story than the intelligence supports.

The third asymmetry is material, and it’s the one that will outlast any administration. The Stockholm International Peace Research Institute (SIPRI) 2026 Yearbook estimates North Korea has assembled around 60 warheads, holds fissile material for at least 30 more, and is accelerating production — while unveiling new systems including the solid-fueled Hwasong-20 ICBM and maneuverable medium-range missiles built to evade defenses.

The Verification Research, Training and Information Centre’s (VRTIC) March 2026 report puts North Korea’s highly enriched uranium stockpile as high as 1.8 to 2.6 metric tons by the end of 2025, based on new open-source imagery of the Kangson and Yongbyon enrichment halls. A January 2026 assessment by 38 North — a publication of the Stimson Center — found that solid-propellant ICBMs, tactical nuclear weapons, and land-attack cruise missiles have done more for North Korea’s military position than any single diplomatic development since Kim’s 2021 weapons plan.

Solid fuel compresses warning time; tactical weapons and cruise missiles add regional coercive options that have nothing to do with reaching the U.S. mainland. These are estimates, not precise counts — North Korea remains opaque. The direction is not in dispute.

This is the fact Washington keeps working around: whether Kim is building leverage to coerce South Korea, pressure Japan, and complicate any future U.S. reinforcement plan. Increasingly, the answer looks like yes.

None of this started with Trump’s second term. In May 2018 he withdrew from the 2015 nuclear deal, known as the Joint Comprehensive Plan of Action (JCPOA), arguing it had failed to block Iran’s path to a weapon. A month later he met Kim in Singapore, where the two sides signed a declaration on “complete denuclearization of the Korean Peninsula” that meant entirely different things to each government. Arms Control Today’s later verdict was blunt: three summits, few results, an expanded arsenal afterward.

The Biden administration inherited the same split file and handled it in the opposite register — a “calibrated, practical approach” to North Korea that avoided both Trump’s summitry and Obama-era strategic patience, while on Iran continuing to treat Tehran’s post-JCPOA violations as reversible if Washington returned to its own commitments. President Joe Biden’s approach was more disciplined. It also produced nothing: North Korea never returned to serious talks, and Kim kept building on the assumption that time would eventually bring acceptance.

That’s the bipartisan part of the record. Trump personalized the North Korea file and made Iran the symbol of a broken deal. Biden lowered the temperature but didn’t revive the Iran agreement or slow North Korea down. Both administrations handed the next one a more advanced North Korea and a more volatile Iran.

The consequence shows up in what each state has room to do next. Iran, still short of a weapon and still under direct military pressure, has every reason to conclude that stopping short of a bomb bought it airstrikes rather than security. North Korea, already armed, has faced none of that pressure and has no incentive to trade away a capability that has functioned exactly as intended — it has kept the regime in power and kept Washington talking about relationship-building rather than regime change.

That is the asymmetry other proliferators are watching. A program short of a weapon can still be pressured, bombed, and sanctioned. A program that crosses the line, survives, and keeps building gets treated as a problem to manage rather than one to reverse. Pyongyang has already drawn that conclusion. Tehran may be running the same calculation now.

By week’s end, the immediate exchange of missiles and drones had largely subsided. But the larger strategic picture remains no clearer than it was a week ago. The debate has shifted from battlefield developments to political calculations: whether Iran is overplaying its hand, whether Trump has a strategy beyond additional airstrikes and intermittent negotiations with a regime he this week called “scum” and “cuckoo,” and whether diplomacy can still produce more than another temporary pause in the fighting.

Those are not the only developments that matter.

Away from the Strait of Hormuz and the headline military exchanges, two quieter developments are unfolding in the region. Neither has dominated the news cycle. Each, however, has the potential to complicate diplomacy, widen the conflict, or reshape the strategic environment in ways that are easy to overlook. These are the signals worth watching in the days ahead.

1. The Houthis: The Dog That Has Yet to Bark in the Red Sea

On Monday, July 6, shipping giants Maersk and Hapag-Lloyd announced that one service operated through their joint Gemini network would resume using the Red Sea and Suez Canal rather than sailing around the Cape of Good Hope. Maersk said its latest security assessment found that conditions had returned to a level of stability sufficient to justify the change, although the decision remained contingent on continued stability in the Red Sea and the absence of further regional escalation.

That is a notable commercial judgment because it rests partly on an assumption that has not been fully tested: that Yemen’s Iran-aligned Houthis will continue to hold their fire. The group, which has controlled Sana’a and much of northwestern Yemen since 2014, carried out repeated missile and drone attacks against commercial and military vessels in the Red Sea and Gulf of Aden after the Gaza war began in 2023. During the war with Iran, however, the Houthis largely limited their direct participation to missile and drone attacks against Israel, while threatening—but not undertaking—a sustained campaign against international shipping.

That restraint surprised many observers. As Iran came under sustained U.S. and Israeli attack, the Houthis controlled one of the few remaining means of imposing rapid economic costs far beyond the battlefield. A renewed campaign against shipping through the Bab el-Mandeb, the narrow southern gateway to the Red Sea and Suez Canal, could have disrupted global commerce, raised insurance and freight costs, and complicated military logistics across the region. Instead, the group avoided a sustained effort to close a second critical maritime chokepoint.

Why? There is no single answer, but three explanations stand out.

The first is deterrence. U.S. and Israeli strikes during 2024 and 2025 damaged Houthi radar networks, launch sites, logistics routes and senior leadership. Reopening the Red Sea front could invite another sustained air campaign while the group is still rebuilding its capabilities and replenishing missile stocks. The Houthis may retain the ability to conduct individual attacks without being confident that they could sustain a prolonged maritime campaign.

The second is domestic politics. However closely aligned with Tehran, the Houthis govern territory and pursue interests of their own. Their fragile accommodation with Saudi Arabia has reduced military pressure and brought financial benefits, including support for public-sector salaries in Houthi-controlled areas. Attacking Gulf energy infrastructure or reopening the Red Sea campaign could jeopardize those arrangements while imposing renewed military and economic costs at home. Fighting Israel can be presented to Yemenis as support for the Palestinians. Risking another war with Saudi Arabia in Iran’s defense would be a more difficult case to make.

The third possibility is that the Houthis regard the Bab el-Mandeb less as an active battlefield than as a strategic reserve. Merely retaining the ability to threaten one of the world’s busiest shipping corridors generates leverage: carriers divert vessels, insurers price in additional risk, and governments must prepare forces for a campaign that may never begin. Employing that capability would consume missiles and drones, expose launch networks and invite retaliation. In that sense, restraint does not mean the threat has disappeared. It may mean the Houthis believe the threat is more valuable while it remains largely unused.

The events of the past week are unlikely by themselves to alter that calculation. The exchange between Washington and Tehran remained limited, Gulf governments did not enter the fighting directly, and no ground operation against Iran—the sort of development that could threaten the survival of the regime and cross a Houthi red line—appeared imminent. Maersk and Hapag-Lloyd are therefore testing not whether the Houthis have been defeated, but whether the conditions that have encouraged their restraint still hold.

Signals to watch: The first will be whether other major carriers follow the Gemini service back through the Suez route. A broader return without Houthi interference would support the judgment that the group is deliberately preserving its maritime leverage. Evidence of renewed weapons shipments, reconstructed coastal launch sites or explicit coordination with Iran would strengthen the alternative explanation that the pause has been used to rebuild. Movement—or breakdown—in Saudi-Houthi negotiations would offer another clue about the importance of domestic considerations.

The most consequential signal would be a change in what the Houthis describe as their threshold for intervention. New warnings against Gulf states, U.S. military logistics or commercial traffic generally—not merely Israeli-linked vessels—would suggest that the group is preparing to move from symbolic participation to economic warfare.

2. Pakistan–Afghanistan: Allies Are Becoming Adversaries

For much of the past two decades, Pakistan regarded Afghanistan’s Taliban as an imperfect but useful partner. Since the Taliban returned to power in Kabul in 2021, that relationship has steadily deteriorated as Pakistan accuses the Taliban government of allowing the Tehrik-i-Taliban Pakistan (TTP)—a separate militant group seeking to overthrow the Pakistani state—to operate from sanctuaries inside Afghanistan. Kabul rejects the accusation, insisting it will not allow Afghan territory to be used against another country while arguing that Pakistan’s security problems cannot simply be blamed on Afghanistan.

The fighting over the past two weeks underscores how dangerous that deterioration has become. Following a series of deadly attacks inside Pakistan that Islamabad blamed on the TTP, Pakistani aircraft on Thursday struck what officials described as militant hideouts in eastern Afghanistan. Taliban authorities said the strikes killed 13 civilians—a claim subsequently confirmed by the United Nations Assistance Mission in Afghanistan—and denounced the attacks as violations of Afghan sovereignty before launching retaliatory strikes across the border.

Efforts by China to broker a lasting accommodation earlier this year reduced the immediate danger of a wider war but left the central dispute unresolved. The significance of the exchange of fire this week is not simply that violence continues along a historically unstable border. It is that both governments appear increasingly willing to rely on military force to pursue political objectives that diplomacy has failed to achieve.

Pakistan has shown little confidence that negotiations alone will persuade Kabul to dismantle TTP networks. The Taliban, for their part, cannot easily ignore repeated Pakistani strikes without undermining their claim to defend Afghanistan’s sovereignty. The result is a relationship in which military retaliation is becoming a more routine instrument of state policy rather than an exceptional response to a crisis.

The relevance to the Iran conflict is not that Pakistan or Afghanistan is likely to enter the war, but that another regional confrontation would weaken efforts to contain it. Pakistan has played a diplomatic role between Washington and Tehran while also confronting militant violence at home and renewed fighting along its Afghan border. A prolonged conflict with the Taliban would absorb military attention, narrow Islamabad’s room to mediate and increase instability along Iran’s eastern approaches.

It would also complicate China’s position. Beijing has tried to restrain Pakistan and Afghanistan while supporting diplomacy over Iran. If fighting intensifies, China would be managing two crises involving countries central to its trade, investment and energy interests. That would not decide the Iran negotiations, but it would further weaken the regional diplomacy needed to keep the conflict from widening.

Signal to watch: The next indicator is not whether another border clash occurs, but whether the remaining political guardrails continue to erode. If Pakistan begins routinely striking targets well beyond the frontier, or if the Taliban sustain retaliatory operations inside Pakistan rather than responding episodically, the conflict will increasingly resemble a prolonged interstate confrontation rather than a counterterrorism dispute.

Equally important will be China’s response. Beijing has invested heavily in Pakistan through the China-Pakistan Economic Corridor while cultivating ties with the Taliban. If Chinese mediation loses momentum or Beijing begins emphasizing security over diplomacy, it would signal diminishing confidence that the dispute can still be managed through negotiations.

That would miss the larger story.

Every NATO summit reflects change, but this one reflects discontinuity. The assumptions that shaped the alliance only a year ago have shifted with unusual speed: the United States is treating NATO less as a community of shared strategic purpose than as a ledger; Europe and Canada are spending more but also trying to reduce the consequences of American unreliability; Ukraine’s war has entered a more fluid phase; Iran has reintroduced the Middle East into alliance calculations; and Hungary’s post-Orbán government may remove one of the most persistent internal obstacles to collective action against Russia.

That is why the most important question in Ankara is not whether NATO will produce another statement of unity. It almost certainly will. The more important question is whether the alliance that leaves Ankara is still the alliance that arrived at last year’s summit.

The first rupture is American. President Donald Trump pressed allies last year into a much higher spending framework. This year, he appears to be moving the goalposts again. NATO Secretary-General Mark Rutte is demanding “clear, concrete and credible plans” to reach the new 5 percent defense investment target, divided between 3.5 percent for core defense and 1.5 percent for broader security infrastructure. European allies and Canada increased defense spending by about 20 percent in 2025, but the issue in Ankara is no longer just whether allies spend more. It is whether they spend in ways Washington approves, including purchases from American defense firms.

That matters because the alliance is moving from burden-sharing to burden-enforcement. In earlier NATO debates, the United States pressed Europe to do more so that the alliance could remain strategically coherent. Trump’s version is narrower and more transactional. He wants visible numbers, visible concessions, and visible credit. European leaders understand this. Their immediate strategy is not to persuade him to embrace NATO’s old language of shared democratic purpose. It is to give him enough of a victory lap to prevent a walkout, troop withdrawal, or public humiliation of an ally.

This is alliance management under duress. It may work tactically. It does not restore the old bargain.

The second rupture is industrial. NATO is no longer talking only about defense spending; it is talking about production. The official NATO Defence Industry Forum in Ankara is focused on turning the 5 percent investment plan into increased production, cooperation, joint procurement, and what NATO calls “industrial deterrence.” The phrase is inelegant, but the concept is important.

As the U.S.-Israeli war with Iran made clear, deterrence now depends not only on forces in being, but on whether allies can replace missiles, drones, radars, air defense interceptors, and precision munitions once a war begins. The United States entered the crisis as NATO’s indispensable arsenal. It emerged with another warning: even that arsenal can be drawn down faster than industry can restore it.

On Sunday, The Wall Street Journal framed the challenge directly: NATO is beginning to fix its cash-flow problem, but it still has to convert money into munitions. Non-U.S. NATO countries raised budgets sharply, but defense industries remain constrained by production bottlenecks, fragmented procurement, personnel shortages, and duplicated national programs. More money does not automatically create more military power. It can also create inflation, competition for scarce components, and long queues for systems that arrive after the crisis they were meant to deter.

That is why Canada’s submarine decision matters. On the eve of the summit, Ottawa selected Germany’s ThyssenKrupp Marine Systems to supply up to 12 new submarines, the largest procurement in Canadian military history. Canada chose the German-Norwegian Type 212CD design over South Korea’s Hanwha bid, emphasizing Arctic operations, NATO interoperability, training, logistics, and long-term industrial cooperation with Europe.

This is not just a Canadian procurement story. It is a signal that a founding NATO member long tied to North American defense arrangements is now anchoring a critical future capability in Europe. Canada is not abandoning the United States. But it is behaving like a country that has learned to hedge against American volatility.

The Arctic makes that decision more consequential. NATO allies have promised Washington that they will do more to secure the region, but the Arctic remains a demanding theater where Russia retains major military assets and where climate change is opening routes faster than allied capacity is expanding. Canada’s submarines will not arrive quickly. The first deliveries are projected years from now. But the decision points to the alliance’s new geography.

The third rupture is political inside Europe. Viktor Orbán’s defeat in Hungary matters more than a routine change of government because Hungary had become an internal brake on Western policy toward Russia and Ukraine. Under Orbán, Budapest repeatedly obstructed European measures supporting Ukraine and complicated sanctions, assistance, and integration efforts. The new government led by Péter Magyar has signaled a different course, creating an opening for the European Union and NATO to act with less internal obstruction.

That does not mean Hungary instantly becomes a full-throated engine of Ukraine policy. Institutions, energy ties, public opinion, and Orbán’s political legacy will not vanish in a month. But the removal of a veto-wielding, Russia-friendly government is a structural development. For years, NATO and the EU had to plan around the likelihood that Hungary would slow, dilute, or block initiatives targeting Moscow. If that obstruction eases, the alliance’s internal map changes.

That change comes as the war in Ukraine is also shifting. The summit will almost certainly avoid triumphant language, and it should. Russia remains dangerous, adaptive, and capable of inflicting heavy costs. But the idea that Moscow holds an uncontested strategic initiative is harder to sustain than it was a year ago. Ukraine’s battlefield innovation, long-range strikes, attacks on Russian logistics, and resilience under pressure have forced NATO governments to think differently about time. Ukrainian adaptation often occurs in months; NATO procurement often operates in years. The mismatch is now impossible to ignore.

That is why the real Ukraine question in Ankara is not membership; NATO has already made clear that Ukraine’s long-term path into the alliance remains unchanged. The practical agenda lies elsewhere: whether Ukraine becomes more deeply integrated into NATO’s defense-industrial ecosystem. Joint production, maintenance hubs, drone development, interoperability, air defense integration, and munitions contracts may matter more over the next year than another carefully negotiated statement about Ukraine’s future.

The fourth rupture is Iran. The alliance may prefer to discuss Russia, spending, and Ukraine, but the Iran crisis will hang over near-term decisions. Several allies resisted or complicated aspects of the U.S.-Israeli war on Iran, including questions about basing and support. Trump has called for “loyalty” from NATO members and linked allied behavior to expectations of support. European officials also face pressure to contribute naval assets and help secure the Strait of Hormuz after the war formally ends.

This matters because NATO’s strategic bandwidth is already strained. If allies are being asked to rearm Europe, sustain Ukraine, harden the Arctic, prepare for hybrid threats, and contribute to Gulf maritime security, the 5 percent debate becomes more than a budget dispute. It becomes a test of whether NATO can absorb simultaneous theaters without the United States dictating priorities through crisis politics.

The fifth rupture is the American force posture question. The Trump administration has already examined reductions in U.S. commitments to Europe, including forces earmarked for crisis reinforcement. The Pentagon announced last month a review of U.S. forces on the continent, and earlier reporting indicated proposals to reduce European commitments by significant margins, including high-end assets such as strategic bombers, naval capabilities, and refueling aircraft.

This is the alliance’s hardest reality. Europe can increase spending quickly on paper. It cannot quickly replace American enablers. Long-range strike, air-to-air refueling, strategic lift, intelligence collection, missile defense integration, and nuclear deterrence remain areas where the United States is not simply another ally. It is the central provider. If that provision becomes conditional, Europe’s problem is not only political. It is operational.

That is why European leaders are trying to buy time. They need enough American commitment to keep deterrence credible while they build capacity that cannot be built on summit timetables. In that sense, Ankara is less a celebration of European rearmament than a negotiation over the pace at which the old alliance can become something else without breaking in the transition.

The sixth rupture is ideological. NATO was founded as a military alliance, but it also became a political community organized around democratic solidarity. That language has not disappeared, but it is no longer doing the same work. The summit’s practical vocabulary is now contracts, percentages, production lines, basing permissions, Arctic access, arms purchases, and force reviews.

That may be unavoidable. Alliances survive by adapting to power, not by reciting founding documents. But there is a difference between an alliance held together by shared strategic purpose and one managed through transactions among nervous governments. The first can absorb disagreements because members believe they are defending a common order. The second can still function, but it requires constant bargaining and produces less trust.

Trump did not create every pressure now bearing down on NATO. Russia did that first by invading Ukraine and then by sustaining a war that exposed Europe’s military limits. Iran has added another theater. China continues to pull American planning toward the Pacific. European domestic politics, including far-right gains and British instability, make long-term spending harder to sustain. Canada’s late military awakening reflects years of underinvestment, not only Trump’s pressure.

But Trump has made the new reality visible. His disregard for alliance sentiment, combined with his need for personal victories, has forced allied governments to confront what many had preferred to manage quietly: the United States may remain indispensable, but it is no longer predictably reassuring.

The Ankara summit will likely produce the necessary words. Leaders will affirm Article 5. They will cite spending increases. They will announce contracts. They will describe unity. Some of that will be real.

But the summit’s deeper meaning lies elsewhere. A year ago, NATO was still trying to preserve the old bargain by adjusting the numbers. Today, it is trying to build a new bargain while pretending the old one still governs the room.

That is the discontinuity.

Rather than publish a separate set of Friday items, I am using today’s edition to briefly frame why that distinction matters and identify the signals worth watching next.

The phrase “intelligence failure” carries a familiar meaning in American national-security debates. It usually suggests that the Intelligence Community — the CIA, NSA, and other agencies responsible for collecting and analyzing foreign intelligence — missed the warning signs, misread the evidence, or failed to communicate danger clearly enough to policymakers.

Iran points to a different problem.

Based on the available public record, the intelligence system did not appear to collapse into silence or affirmation. It assessed Iran’s capabilities, warned of regional risks, identified uncertainty around nuclear timelines, and complicated the public case for war. Senior officials reportedly challenged the most optimistic assumptions about what military force could achieve.

Then the president launched the war anyway.

Later assessments appear to have complicated the administration’s claims about Iran’s degraded military capacity, nuclear timeline, missile force, and ability to impose costs through the Strait of Hormuz and regional proxies

That is the central point of the Just Security piece. Iran may prove less an intelligence failure than a policymaker failure. The warning appears to have survived the system. What failed was the decision process around it.

I describe that failure as permission capture: a breakdown in which intelligence may be rigorous, relevant, and available, but loses force because the president values voices that permit action over voices that complicate it. That permission may come from a foreign leader, a political loyalist, an adviser, or anyone who understands that this president prefers the clean promise of action to the hard work of absorbing risk.

That distinction matters because “intelligence failure” can become a political escape hatch. It allows policymakers to imply that something essential was not known, not collected, or not pressed hard enough. Sometimes that is true. But when intelligence does warn, and when elected leaders choose to ignore or displace that warning, the label obscures more than it clarifies.

The better question after Iran is not whether the Intelligence Community can still warn Trump. The better question is whether warning still matters when it competes with permission.

Signals to Watch

• Whether Congress builds a record before the next crisis. Congress cannot force a president to heed intelligence warnings, and it should not become a shadow National Security Council. But it can insist on a clearer intelligence baseline after the use of force: what the Intelligence Community assessed before the decision, what confidence levels attached to those judgments, what dissents existed, and what assumptions supported the administration’s theory of success. Without that baseline, oversight becomes a retrospective argument over fragments.

• Whether dissent is protected or forced into leaks. One of the lessons of Iran is that the system may still be producing inconvenient analysis. That is the good news. The bad news is that dissent appears to matter less if it cannot shape decisions or create accountability. Watch whether Congress, inspectors general, and internal review mechanisms give analysts and officials protected channels to raise concerns about manipulation, suppression, or selective use of intelligence. If the only durable record comes from leaks, the oversight system is already failing.

• Whether “permission capture” migrates to the next crisis. Iran is not a closed case. It is a warning about the next one. The next test may involve terrorism, cyber operations, Cuba, Greenland, renewed escalation in the Middle East, or something not yet visible. The pattern to watch is whether the president receives sober assessments and then gravitates toward the voice that makes action sound cost-free. The danger is not that intelligence disappears. It is that it becomes just one more input competing against praise, loyalty, impatience, and the last confident person in the room.

• Whether the wrong lesson takes hold. The wrong lesson from Iran would be that the Intelligence Community simply needs sharper language, better collection, or more forceful briefings. Some of that may be true. But the public record points to a more immediate failure: intelligence appears to have done much of what the system asks it to do. The warning was available. The president chose a course and a narrative the evidence did not support.

  ---

  The Safehouse Briefing covers geopolitical risk and security strategy. If you found this useful, share it with someone who should be paying attention.

  ---

The security story is less obvious.

Britain is about to have its seventh prime minister in a decade at the same moment Europe is trying to rebuild its defense base, sustain Ukraine, reduce reliance on the United States, and prepare for a NATO summit shaped by ongoing doubts about American reliability. British politics has become accustomed to leadership turnover. European security is less able to absorb it.

Burnham is all but assured to become Labour leader and prime minister; Wes Streeting, the only serious rival, has said he will support Burnham rather than run. Unless another challenger emerges, Burnham could be in Downing Street by mid-July. That timing matters. The NATO summit in Ankara is scheduled for July 7-8. Starmer will likely attend as caretaker prime minister.

A shift in UK leadership will have limited impact on the geopolitical security front. NATO commitments will hold. Ukraine support will not disappear. But security policy depends on more than continuity of paperwork. It also depends on political authority, trust between leaders, and the ability to make commitments that others believe will survive the next domestic crisis.

The Foreign Policy Gap

As mayor of Greater Manchester, Burnham built his reputation on transport, public health, housing, regional governance, and the politics of devolution. Those are not small issues. What we know about Burnham’s foreign policy views is partial.

He is not anti-NATO. He is likely to maintain Starmer’s defense spending commitments and support continued assistance to Ukraine. He has also favored closer relations with the European Union, while staying inside the usual Labour red lines: no return to the single market, no customs union, and no reopening of Brexit as a formal membership question.

That gives Britain a direction, but it does not yet give it a strategy. Burnham will discover quickly that “fixing the country” includes decisions about missiles, procurement, intelligence sharing, sanctions, shipyards, and European defense finance.

The Calendar He Inherits

The first test is Ankara.

Britain is expected to present, or at least clarify, its Defence Investment Plan around the NATO summit. That document is supposed to connect last year’s Strategic Defence Review to actual money, but the plan has already slipped. A caretaker prime minister can restate commitments. He cannot credibly bind his successor to difficult funding choices. Allies will understand that. So will defense firms. So will the Treasury.

The second test is now less a date than a delay.

The UK-EU summit scheduled for July 22 has been postponed in response to Starmer’s resignation. The summit was supposed to be the next formal step in the reset Starmer began after years of post-Brexit friction. The meeting was expected to conclude talks on an agrifood agreement, an emissions trading arrangement, and a youth mobility scheme. It was also meant to give London and Brussels a platform for discussing whether the UK could move closer to the EU’s economic and security orbit.

The postponement changes the early test for Burnham. He will not have to walk into a major Brussels summit within days of taking office. He will instead inherit unfinished files, an uncertain timetable, and the burden of proving that Starmer’s reset was not just a feature of Starmer’s personal diplomacy.

The defense issue remains the hardest part. Britain is still outside the EU’s Security Action for Europe (SAFE) mechanism — the €150 billion loan facility for joint defense procurement. Talks over UK access have already been difficult, creating a gap between the political language of European security cooperation and the practical machinery now being built in Brussels.

That gap matters because Europe’s defense challenge is no longer only about national spending. It is about production scale, joint purchasing, industrial capacity, and the ability to reduce dependence on American systems over time. The EU is building the financial and industrial instruments Britain needs if it wants to remain deeply embedded in continental defense planning.

A Burnham government may try to move faster than Starmer on a more structured UK-EU security relationship. That could mean upgrading the existing security pact, seeking a regular role in EU foreign affairs and defense consultations, and finding a route into collaborative procurement without reopening the politics of Brexit. The strategic logic is clear, but the politics are harder. Brussels will not give Britain privileged access without conditions.

The postponed summit therefore becomes a useful measure of the transition. If it is quickly rescheduled and produces concrete agreements, Burnham will be able to claim continuity with Starmer while giving the reset a more practical purpose. If it drifts, the reset risks becoming another post-Brexit process with warm language and limited machinery.

Why the Gap Matters

Europe’s defense problem is no longer abstract. It is visible in production lines, delivery queues, stockpiles, and procurement delays.

Ukraine has forced European governments to measure their defense industries against wartime demand. As I have noted in many recent Tuesday newsletters, ammunition production remains uneven; air and missile defense capacity is stretched; and Patriot air defense demand exceeds supply. European alternatives are improving but still incomplete. Strategic enablers — air-to-air refueling, lift, ISR (Intelligence, Surveillance, and Reconnaissance), command-and-control, logistics, satellite capability, and high-end missile defense — remain heavily dependent on the United States.

This is the context in which Britain’s leadership gap matters. Since Brexit, Britain has tried to define itself as a reliable European security actor even while sitting outside the EU, but that role depends on credibility. And credibility depends on the ability to make decisions and sustain them.

A summer of transition does not destroy British credibility. It does weaken the signal at a poor moment.

What to Watch

The first indicator is the foreign secretary. If Burnham appoints someone with real diplomatic and security experience, it will suggest he understands the gap in his own background and intends to compensate for it.

The second indicator is the Defence Investment Plan. If it appears before September with credible funding assumptions, the transition will look manageable. If it slips again, the gap between Britain’s defense language and its defense choices will widen.

The third indicator is the postponed UK-EU summit. Burnham does not need to produce a grand bargain with Brussels in his first week. He does need to show that the reset has survived Starmer’s departure.

The immediate question is whether the summit is quickly rescheduled and whether it produces more than process language. Agrifood, emissions trading, and youth mobility matter in their own right, but the larger test is whether Britain can build a practical security and defense relationship with the EU while remaining outside its core institutions.

The fourth indicator is President Donald Trump. Burnham’s past warnings about “poisonous” American-style politics and Trump-driven instability suggest he will not instinctively flatter Washington. That may help him domestically, especially against Reform, but the test is whether he can separate political distance from strategic management. Defense spending, Ukraine, intelligence sharing, tariffs, and any future Iran-related request for British support will all run through the U.S. relationship. Starmer’s relationship with Trump deteriorated over Iran and the use of British facilities. Burnham’s task is not to like Trump. It is to keep the U.K.-U.S. relationship usable under a president who treats allied friction as leverage.

The fifth indicator is Ukraine. Do Burnham’s early statements and appointments suggest that he sees Ukraine as a discrete foreign policy issue or as the central test of Europe’s security order? The latter is the right frame. Ukraine is buying Europe time to rebuild deterrence. Britain has been one of the countries most willing to understand that.

The Broader Pattern

Burnham’s record in Manchester suggests he understands delivery, public administration, and the political value of showing visible results. Those skills matter. Defense policy often fails not because leaders misunderstand the threat, but because governments cannot turn strategy into programs, contracts, production, and sustained funding.

But the constraints are real. Reform UK is rising. Labour’s left is restless. The fiscal room is narrow. Public patience is limited. A prime minister elected on domestic repair will face pressure to treat foreign policy as a secondary concern.

The world is unlikely to cooperate.

The Safehouse Briefing covers geopolitical risk and security strategy. If you found this useful, share it with someone who should be paying attention.