Markets now expect roughly 60 basis points of tightening over the next year — far less than the average 250 basis points priced during 2022 — yet the sharp rise in long‑term yields suggests the tightening cycle could shift from initially shallow to a potentially deeper one.

Figure 1: Odds of a faster hike (%)

Source: CME, Kalshi

Forward Yields

The forward market — the yield discounted far out into the future — moved first, but the 10‑year has now caught up. The 10‑year forward yield, discounted 10 years ahead, has barely changed, at about 6.1%, since early August’s oil‑price lows, while the spot 10‑year has climbed 18 basis points.

Crucially, the forward yield peaked back on August 18, and because forward rates typically lead spot rates, that early turn has helped fuel the equity rebound (see Figure 2).

Even so, the combined move in forwards and long‑term spot yields now implies a rate‑hike path of roughly 120 basis points—based on the current spread—playing out over a longer horizon.

Figure 2: Forward yields have “peaked” (%)

Source: CME, US Treasury

Rates in a “good place”

NY Fed’s Williams struck a deliberately neutral tone on policy—likely mindful of the recent jump in yields—which he attributed largely to stronger underlying fundamentals (e.g., today’s factory/durable ex-transports orders at 0.6% and 1.1%, the best since April). The economy’s strength therefore dictates financial conditions, as the investment surge is driving up rates.

At the same time, he echoed Warsh in saying the Fed must ensure inflation remains on a sustainable path towards the target; if not, he signaled alignment with Warsh on the phrase that the Fed “has more work to do on inflation” (a ~wording for rate hike(s)).

This is a signal that, in my view, Williams—who has consistently described policy as moderately restrictive and as remaining data-dependent—sees the current stance as in a good place, but that could change if inflation does not stay on the right path.

Our model scored his interview as moderately hawkish, putting Williams among those who see a conditional case for a hike if inflation doesn’t cooperate. If he tilts at this month’s FOMC meeting—and with Friday’s jobs report and the September 11 CPI pivotal—the first hike becomes almost certain, with room for more thereafter.

Table 1: Williams LLM scoring of CNBC interview

Source: FedWatch Advisors LLM

Oil markets in charge

Yield moves reflect energy, tariffs, AI‑driven capex, heavy issuance, and a stronger economy. Yet the term premium remains subdued, with only moderate expectations of a hike, and TIPS breakevens are steady as real yields rise alongside nominals — a synchronized signal of firmer GDP expectations rather than an inflation scare.

Oil markets are back in the driver’s seat, and the initial spike already pushed up the odds of faster hikes as the energy‑and‑food shock fed through inflation. If crude makes another run at $100 — now within reach — it would quickly lift the probability of accelerated tightening, given how closely oil and rates tend to move together.

Figure 4: Oil and odds of faster hikes (%)

Source: ICE, CME

Correlation Driver

Thus, the correlation between oil and yields, currently at 0.72, is the highest since the GFC when oil prices spiked. This correlation again serves as a broad market indicator: if the correlation rises, risk assets are under pressure, and vice versa. The “peak” in forward rates, which reflects expectations for faster hikes, could lower the correlation and thus offer relief to markets.

Figure 5: Oil and yield correlation is the highest since the GFC

Source: Bloomberg, ICE, Tradeweb

Yesterday’s Dallas manufacturing survey was extraordinary, reflecting Texas’s breakneck data‑center boom, with Dallas–Fort Worth now the world’s top market—surpassing Virginia and Atlanta—driven by AI/HPC demand and ERCOT’s fast‑track grid interconnections.

At the same time, Texas manufacturing firms reported margin pressure from sharply rising raw‑material, labor, health‑insurance, and other input costs, noting that they can no longer pass through price increases and are instead absorbing inflation that threatens profitability.

With ISM manufacturing due this morning, services ISM on Friday, and global PMIs in between, yields are reacting acutely to price signals in these surveys, especially “prices paid,” which the Gulf conflict is lifting.

The Strait of Hormuz shock accounts for a material share of the ISM Prices Paid jump—via higher energy, metals, freight, and insurance costs—while regional PMIs, driven by local demand and labor conditions, show the impact in their input-cost subindices.

Figure 1: Fundamental drivers: 30Y yield and prices paid

Source: Dallas Fed, ISM, US Treasury

Yields are tracking Prices Paid because it signals how long domestic input‑cost and goods‑inflation pressure could persist if the conflict drags on—a key condition several FOMC voters and non‑voters have cited as grounds for considering a rate hike.

To that effect, the bond market is putting the Treasury and the Fed on the same page. As Warsh said, [the fundamental reason for yields rising is that] the global “savings glut” that suppressed interest rates for decades is turning into a “global investment surge.” He flagged in his speech at Jackson Hole that the recent rise in overall commodity prices bore watching to see whether trends indicate upside inflation risks.

Hike odds are pushing toward 60% in futures and prediction markets. Still, Fed communication—especially Warsh’s—puts the probability north of 80%, reflecting a much stronger policy‑tightening bias than markets alone imply (Figure 2).

With core PCE running roughly 1.5 standard deviations above the Fed’s 2% target, bonds aren’t second‑guessing the Fed’s reaction function: elevated commodity prices and rising “prices paid” are exactly the inflation signals policymakers have warned would justify a hike if they persist.

Figure 2: Rate hike probabilities (%)

Source: CME, Kalshi, FedWatch Advisors Fed Sentiment Index

But as the sell-off in Treasuries unfolds and the global aggregate bond index yields 3.72% at its yield-to-worst, the UK and Japanese bond markets remain in the lead. Their dynamics differ from those of Treasuries, yet, through correlation and hedging, they affect Treasury yields.

The UK is leading today, with its 50Y bond trading close to 25 cents on the dollar (Figure 3). The reason is not technical hedging of corporate bond issuance, but macro fundamentals.

Figure 3: UK long-end, worst performing (price, cents)

Source: UK DMO

Japan is highly reliant on the Strait of Hormuz (~92% of oil imports). In comparison, the UK is indirectly reliant through Qatar and North Africa (~15% fertilizer/chemicals imports). Since each country is a twin-deficit (fiscal & trade), investors require higher yields to fund both in the wake of commodity-push inflation. Those trends are now also driving US Treasuries.

Figure 4: 10Y yield ten years from today (%) leading the US

Source: MoF, UK DMO, US Treasury

So, is the US the best performing bond market? Over the past month (Aug 1^st-Sep 1^st), Treasuries returned 0.30%, while Europe and Japan were down 0.5%. Yes, the US bond market is the best-performing so far, but US junk bonds top the charts (see Figure 5).

Figure 5: Best-performing bond markets

Source: Barclays Indices

Bond prices reflect economic fundamentals, which are affected by the Hormuz supply shock. The latest Global US PMI shows rising employment growth, while input cost inflation eased; the reading remains well above pre-pandemic levels.

These are the trends directly linked to Hormuz that are setting the market up for several central banks, including the Fed, to raise rates this month. Ultimately, in my view, rate hike(s) will be a bull case for stocks and bonds.

Figure 6: Commodity trends

Momentum began to recover in late July, and after the brutal selloff that followed, traders leaned straight into the concept, doubling their Texas hedge by buying call options on momentum ETFs (Figure 1).

September opens as a seasonally difficult month, with the S&P posting an average decline of roughly 2.5 percent since 2021 — a reminder that even during periods of strong momentum, the index can struggle as fall begins.

Yet those same seasonal headwinds can create opportunity when underlying momentum remains intact. When momentum is strong enough to press down on volatility, the market can turn what is typically a weak month into a window for value.

Figure 1: Momentum ETF call options volumes and Momentum Total Return (%)

Source: MSCI, Invesco, Virtus, Factors to Watch by Bloomberg

Momentum Earnings

Going into this week, there could be more of this hedging activity as a big lineup of economic data and earnings takes shape. Companies like Dell, Broadcom, DocuSign, Zscaler, Snowflake and Planet Labs are prominent holdings in Momentum ETFs and related indices.

Dell and Broadcom remain the clear momentum standouts, delivering 30–80%‑plus year‑over‑year revenue growth against the S&P’s mid‑single‑digit baseline, and both continue to show a strong pattern of beating guidance and lifting AI‑linked outlooks.

For the momentum index and its ETFs, Broadcom is the only name with meaningful S&P earnings weight, while Dell operates primarily as a pure momentum‑factor driver.

Both stocks, however, have effectively been “Texas hedged” for several years, with call‑option volumes steadily rising and short interest consistently falling. This reinforces their directional exposure rather than offsetting it (Figure 2).

Taken together, these dynamics give Dell and Broadcom outsized influence over momentum in September, creating a potential bedrock for an upside surprise in overall stock‑market performance.

Figure 2: Broadcom and Dell: momentum is back

Source: NYSE

Key factors for momentum are bond issuance and yield direction. There is an estimated $215 billion in Investment Grade issuance in September, of which ~30-40% could be from the hyperscalers.

Figure 3: IG Issuance ($, billion)

Source: Markitt, Bloomberg Data

Markets now understand what that means: the issuance will be hedged with Treasuries, pushing up long-term rates, while the proceeds are used to invest in data centers that affect the economy and inflation, pushing up rate-hike probabilities.

At the same time, proceeds are used to buy equity stakes in other hyperscalers and chip companies, which in turn leads to CDS hedging (aka the Texas version).

Taken together, these have created a unique environment: despite rising yields and widening CDS spreads, the long momentum indices have not pared back (Figure 4).

In fact, if the Fed goes for a hike, long-term rates may peak and underpin stock markets by the token of momentum. Hence, September is the month of the momentum trade comeback.

Figure 4: Momentum Index and Credit/30Y Treasury ETFs (normalized scale)

Source: Morgan Stanley, iShares, F/M Investment

At the center of the meeting are China’s interests—access to SWIFT, secondary sanctions on Chinese banks, secure energy supplies—against the US’ interests of constraining Iran’s oil revenues. Kharg Island is a critical node, and China’s oil purchases financed through its banks, undermine sanctions. Maritime enforcement in the Strait must tighten.

The US may have the upper hand here to some extent, given how reliant Chinese banks are on revenues from oil sales to Iran and on SWIFT payments in dollars (see Figure 1).

Figure 1: SWIFT payments in USD and CSI 300 Bank Index

Source: Shenzhen, SWIFT

Market Reaction

For markets, these opposing interests are key issues. Seizing Kharg Island could spike oil prices and tank Chinese banks and equities. At the same time, cooperative enforcement in the Strait could increase oil flows, and cutting SWIFT access may negatively affect dollar liquidity and Treasuries.

Markets reacted with pressure to the overnight U.S. strike on two Iranian rocket launchers on Larak Island—an IRGC forward base sitting directly on the Strait’s narrowest point—while the President simultaneously threatened to bomb Kharg Island and retaliation to Iran’s attack on US forces, producing a twist‑steepening in yields, a pullback in equity futures, and a wider Brent spot premium.

But it is a reaction to how oil and inflation risk premiums—a combination of secondary sanctions and the Kharg seizure —are priced in at the tip of the G20 meeting, driving yields and oil once again in tandem, weakening the dollar and Gold, and driving up copper and agricultural commodities.

Given the G20 is an agenda of intervention, the ‘soft lines’ of recent interventions in the Yen at 163 to the dollar and 5.33% on the 30Y are to be tested as $200 billion investment-grade issuance hits in September, while the Fed’s new projections may show several rate hikes in 2027.

Figure 2: Yield curve and oil prices

Source: US Treasury, Comex

Repricing Risks

With one move, the President has effectively unwound Friday’s market setup, when investors were balancing expectations of expanding oil supply and a firmer Fed against the risk that a fresh Iranian supply shock could lift near‑term inflation.

The on‑again, off‑again war dynamic is showing up most clearly in energy, which is up 2–4% (refiners +2%, offshore drillers +5%), while consumer discretionary, financials, and tech are off about 0.7% on average.

Markets are pricing the Iran premium (Figure 3) as a mild growth drag—higher energy costs and tighter policy—but even the threat of seizing Kharg is not yet being treated as a global economic break point. At the same time, economic pain is materially felt in Iran, including internet/GPS disruptions.

Figure 3: The Iran gasoline premium ($/gallon)

Source: Axios

Lost Decade

Yet, the market will keep trading the war-on/war-off” risk environment, at least for the immediate future. For the secular horizon, a lost decade in bonds is well underway (see Figure 4).

Inflation can only ease if the Fed and other central banks tighten more aggressively, a path that risks stoking debt concerns. Treasury’s daily data already show gross interest outlays at about $1.05 trillion through June 2026, up from roughly $900 billion over the same period last year.

While that may provide an ongoing bid for “AI and data center bonds” away from Treasuries, yields remain at risk for a more significant breakout.

Figure 4: Lost decade for bonds

Source: Ashwath Damodaran, NYU Stern

Heavy‑sour barrels flowing into the Atlantic Basin narrow discounts and ease supply‑risk premia, compressing Brent’s geopolitical premium as traders price U.S. access to 17 Venezuelan fields.

Venezuela’s slow production recovery—still only ~1.25 mb/d and years from meaningful expansion—limits the pull on headline oil prices, even as the deal nudges Caracas toward considering an OPEC exit that would have little near‑term impact on global supply.

At the same time, Venezuela’s rising exports highlight the U.S.’s role as a price‑responsive producer, expanding when prices rise and slowing when capital tightens.

The deal reinforces this dynamic by extending U.S. influence over the Western Hemisphere’s heavy‑sour supply base, shaping the discounts that feed directly into Brent pricing and, ultimately, U.S. gasoline.

Figure 1: Venezuelan oil exports to the US, and US total oil exports to the rest of the world (mb/d and kb/d)

Source: Department of Energy

Market Indigestion

Announced after the close Friday, the news of the deal first hit HyperLiquid, where Brent slipped 1.5%. However, the larger move was the week’s 5% drop as improving Hormuz flows bled out the geopolitical premium, Warsh’s hawkish turn weighed on demand expectations, and new Venezuelan supply optionality further softened expectations in the oil price curve.

That pressure spilled into the long bond, which—despite Warsh signaling tighter policy—still finished 1.5 basis points lower and sits roughly 15 basis points below its level at the debt‑buyback announcement. Markets are effectively pricing a mix of expanding oil supply and firmer Fed policy, a combination that pulls energy prices down without tipping the economy into a deeper slowdown signal.

A new GDP Now

Warsh’s seven economic indicators—reconstructed here as a synthetic GDP Nowcast—point to growth running roughly two percentage points below what the Atlanta Fed’s live model currently shows.

Should the Fed follow through with a single 25bp hike by September later this year (December), the real bite on GDP would only be felt in 2027; the 12–18 month transmission lag means 2026 escapes largely unscathed, with any added drag capped at a modest 0.1–0.2pp.

Figure 2: Warsh’s GDP vs. Fed’s GDP (Y/Y%, nominal)

Source: Atlanta Fed GDPNow, FedWatch LLM calculations. Nominal GDP Now = real GDP Now + nowcast of PCE

How Restrictive can the Fed get

Fed policy is only slightly restrictive on Warsh’s 12‑month core PCE gauge (+0.3%) and outright accommodative on real‑time PCE (‑0.16%), leaving it far from the kind of sustained restrictiveness seen in 2022–23 when policy had to tighten materially to drive annualized PCE from 5.5% to 3.5% (Figure 3).

If we treat the 2022- 2023 tightening cycle as a rough guide, according to the San Francisco Fed, the implied (very approximate) “disinflation elasticity” was about 0.6 percentage points of core PCE per 100 basis points of tightening.

To use that same linear rule of thumb, the Fed would need to tighten by the equivalent of about 200–300 bp in effective policy—pushing the real stance into clearly restrictive territory (on the order of +1.5–2.0% on Warsh’s metric) to have a reasonable chance of pulling PCE from 3.7% to 2% over the next year.

Because the short end isn’t pricing meaningful tightening while the long end already embeds a higher policy path, falling oil prices and a firmer Fed policy stance should pull 10Y–30Y yields lower by roughly 40–80 bps, driven mainly by term‑premium compression rather than shifts in expected short‑rate trajectories.

Figure 3: Fed policy restrictiveness and LLM score

Source: Federal Reserve, Cleveland Federal Reserve, FedWatch LLM

Obviously, Warsh is going to say something, but it will likely be a brief speech; potentially the 2022 hawkish version without pain and with a length well below the average historical word count for JH speeches of around 1500- 2500 words.

For markets, the speech’s brevity becomes its own signal—a Greenspan‑style “briefcase indicator.”

Fewer words imply he’s quietly urging colleagues to dial back hawkish guidance. But when he stretches out the economic narrative, he’s aligning with the traditional Jackson Hole mode—effectively siding with the committee by leaning into the familiar “economic outlook” frame.

Over time, Jackson Hole speeches have become their own market signal: the old link between 10‑year yields and word count held until 2022, but now it plays more like a Mad Libs—the shorter the speech, the higher the yields, as brevity reads more hawkish than lengthy remarks that allow for plenty of room to be dovish.

Figure 1: JH speech word count and the 10Y yield (%)

Source: Federal Reserve, US Treasury

While Warsh has been tightly lipped on guidance, he has made price stability his key objective and could play a significant role today. For the Fed, and any central bank, price stability is the foundation of a healthy economy, and preserving it requires credible communication and, when inflation drifts, the willingness to tighten policy and accept short‑term pain to secure long‑term stability.

Over the past thirty years, there have been five hallmark speeches on price stability, with Greenspan (and Powell) most notably, especially with a strong nod to the link between price stability, credibility, and political independence. The market reaction was positive for stocks, but initially negative for bonds (Figure 2).

The 1996 JH speech by Greenspan, “Achieving Price Stability,” which consists of 1935 words, can provide a framework for Warsh’s speech.

Greenspan argued that price stability is achieved only when inflation expectations no longer shape decisions, even as defining the price level becomes harder in a modern, innovation‑driven economy.

Warsh could recast Greenspan’s point for today by warning that in an AI‑driven, fast‑changing economy, price stability still requires inflation expectations to stabilize, but because innovation blurs the very idea of a general price level.

If elevated expectations erode credibility faster, it demands clearer communication, tighter discipline, and a firm commitment to the 2% anchor. Such framing is sufficiently hawkish to push Treasury yields lower.

Figure 2: JH speeches focused on price stability and market reaction

Source: Federal Reserve, Kansas Federal Reserve

Indeed, that is what happened in 1996 (see Figure 3). Greenspan’s speech was perceived as hawkish, such that the words “price stability” triggered a year-end 12% rally in the S&P, a stabilization of the dollar, and a tightening of credit spreads.

As for the 10Y, yields peaked around 7% in 1996 (the long-run average since 1850) and fell by a percentage point. The emphasis on price stability is read as “hawkish credibility,” which can materially loosen financial conditions.

Figure 3: 10Y yield 1996 vs. 2026 (%)

Source: Bloomberg, US Treasury

The standout is Kansas City Fed’s Schmidt — a non‑voter who cannot formally dissent, yet openly favored a hike in July, making him the second non‑voter after St. Louis Fed’s Musalem to signal a preference for tightening. Hammack, who anchors the hawkish flank of the FOMC alongside Kashkari, went further still, arguing that now is the right moment to act.

The Fed’s near‑term reaction function—its Now‑Casting of whether to move in short order—shows up clearly in the Fed Sentiment Index, which aggregates all Fed communication, including market commentators (Figure 1). Markets price the odds of a hike at roughly 40% (prediction markets at 55%, historically better forecasters), but the Fed Sentiment Index implies a much higher probability of about 74% (Figure 2).

With a growing chorus of Fed officials calling for a hike — three Board members conditionally and five regional presidents unconditionally — the majority of the FOMC, non‑voters included, is now leaning toward tightening.

That collective stance suggests forward guidance may strengthen from here, even though Warsh is likely to stay ambiguous on policy direction.

Figure 1: Fed Sentiment Index (LLM score)

Figure 2: Odds of a hike by three methods (%)

Source: FedWatch LLM **Implied Hike Odds≈50%+20%×FSI total score

To that extent, 2Y yields, the proxy for the Fed’s tightening cycle over the cyclical horizon, have broken away from oil prices, possibly reflecting that the Fed (regional) presidents have been clear about what to expect from policy.

Beth Hammack was specific, saying that the Fed’s responsibility is to communicate urgency, reinforce credibility, and prevent the public from believing that high inflation is permanent.

Figure 3: 2Y yields (%) and oil prices ($/pb)

Source: US Treasury, ICE

Warsh in play

Inflation expectations, measured by the cleanest market-based measure (the inflation swap), have begun to rise in tandem. The underlying inflation trend—measured by the average of the NY Fed’s multivariate trend CPI, Dallas Fed trimmed-mean core PCE measures (including alternative measures), and super core PCE—has resumed upward, as has the nowcast core PCE (Figure 4).

Based on Warsh’s own reaction function, as he briefly explained during the July press conference, the Fed should tighten policy when the underlying inflation trend is rising.

His speech tomorrow may not focus exactly on this, but with eight Fed members in favor of tightening by acting now or soon, the underlying inflation trend is strengthening Warsh’s case for a hike as well.

Figure 4: Underlying inflation trend, Nowcast core PCE and inflation swaps (%)

Source: CME, NY Fed, Cleveland Fed, Dallas Fed

Tech companies now run payments, credit, deposits, underwriting, and the rails that move money at system‑level scale, with Apple Pay alone handling $6 trillion a year and Google Pay and Meta routing similarly large flows.

They finance assets through SPVs, operate alongside fintechs that receive bank charters, and control transactions, fees, credit, and capital allocation without bank-level oversight. This combination has led regulators to label them shadow banks.

Their platforms even set internal interest rates: Apple Pay Later, Amazon seller loans, Meta’s ad credit terms, and Google’s cloud financing all create closed‑loop systems where the platform, not the market, determines the cost of capital.

And among them, Apple stands out as the “Fed of Technology.” It commands the largest consumer payment rail, dominates platform credit, and shapes liquidity through Apple Savings and settlement timing, making it the closest central‑bank analog in the digital economy.

The Tech vigilante is a closed‑loop financial system embedded inside Big Tech platforms that can move trillions in payments, extend credit, control settlement timing, allocate capital through SPVs, and effectively determine interest‑like rates within its ecosystem.

Apple’s short-term rates being determinative, effectively set the term premium for Tech; a risk premium that expresses how debt-financed investment in equity stakes may generate productivity. This premium is now tightly correlated with Treasury Term premium (Figure 1).

Figure 1: The Tech Term Premium vs. Treasury Term Premium (%)

Source: New York Federal Reserve, FedWatch Advisors calculations

If Apple operates like the Fed, bond markets would price in expectations for the actual Fed based on Apple’s yield curve. Such has indeed been the case, as Apple yields match Treasury and even (implied) Fed Funds futures yields.

Rate expectations are increasingly determined by developments in Tech, given the sector’s impact on GDP (about 5.6% of U.S. GDP, $1.6–$1.7 trillion in economic value).

In terms of inflation, tech has been a material factor. AI investment has pushed tech‑sector inflation sharply higher, with semiconductor prices up 26% (the highest since 1984), computer equipment up 7.8% (the highest since 1995), and software PCE up 14.5% (the highest since 1977).

It is a surge that added 15 bps to headline PCE and two‑thirds of a percentage point to core PCE in just four months. As short-term Tech bond yields are about 3.75% with core PCE at 3.3%, the “Tech policy rate” (so to speak) is more restrictive than the Funds rate (by about 10 basis points).

That feeds into the Tech term premium, which in turn affects Treasury yields, as Tech bonds reflect the impact of AI investment on GDP and inflation.

Figure 2: Apple, Treasury, and Fed Funds curve (%)

Source: Apple, US Treasury, CME

Bringing this concept to what is happening right now- Nvidia’s unexpected positive forward guidance for accelerating sales by 2028, and Apple’s 2027–2028 outlook revenue growth—mirrors the IMF’s model of credible, state‑contingent forward guidance.

Tech guidance, however, isn’t a rate‑hike trigger on a standalone basis. Still, the guidance clearly argues for higher‑for‑longer rates, showing that AI demand is booming, Big Tech thrives, and financial conditions remain loose even as the Fed must still respond to broadening inflation.

With Kansas Fed’s Schmidt, the eighth Fed member favoring a hike without dissent, the chances of a September rate hike remain high. While prediction and futures markets show about 40% odds—historically the zone where the Fed has been willing to move on rates—the Tech guidance implies a rate hike should follow shortly (Figure 3).

Figure 3: Odds of the next policy move (%) and Tech EPS guidance ($/share)

Source: CME, Nasdaq, NYSE, FactSet

Year-to-date, Nvidia’s 2026 equity moves include $30B into OpenAI, $3.2B in Corning warrants for 18M shares, $2.1B for rights to 30M IREN shares, and $2B for ~23M CoreWeave shares, forming a $40B public‑plus‑private equity portfolio. The average stock purchase as a % of outstanding float exceeded 15 percent.

As equity risk found its way into Nvidia’s debt, this idiosyncratic risk is hedged with Nvidia’s single-name CDS, the 5Y maturity of which was most actively traded. The CDS curve has steepened, affecting the 2Y-7Y segment of the Treasury curve (Figure 1).

Conversely, Nvidia’s hyperscale issuance, dominated by 20Y to 30Y maturities, has been primarily hedged using the long end of the U.S. Treasury curve — specifically the 30‑year and 10‑year maturities — because these match the duration profile of Nvidia’s longest tranches.

Nvidia’s CDS have largely moved in step with the front end of the Treasury curve, reflecting how the company’s role in driving the AI cycle feeds directly into near‑term Fed expectations.

By contrast, its long‑dated bond issuance pushed a hedge‑driven risk premium into the long end of the Treasury curve, amplifying concerns about fiscal deficits. The more Nvidia beats earnings, the more capacity it has to issue new debt to buy equity stakes, influencing the Treasury curve across maturities.

Figure 1: Treasury yield curve (%) and Nvidia’s CDS (bps)

Source: Markit, US Treasury

Core PCE

The hotter monthly print reflects that its reference period fully captured the energy-price surge, while CPI only partially did so. Energy’s 15.3% contribution continues to hold headline inflation up.

Services remained the driver: a roughly +3% m/m jump in air travel, medical, and financial services added $86.2B to spending, lifting PCE to 0.2%. Rising healthcare costs and wage‑driven service prices pushed core to 3.3% YoY, and strong dividend and income gains kept demand firm enough to prevent goods disinflation.

PCE excludes AI infrastructure, chips, GPUs, HBM, servers, and data-center capex because it measures household consumption. Still, indirect AI channels — stronger asset income, tech‑linked wage spillovers, and early productivity effects — may be quietly supporting demand and keeping services inflation sticky.

Nvidia’s influence can indirectly add inflationary pressure, meaning its equity and bond yields increasingly behave like a tightening—or loosening—force on financial conditions.

Figure 2: Nvidia yields and Fed Funds adjusted for real-time core PCE (%)

Source: Nvidia, Federal Reserve, Cleveland PCE NowCast

Nvidia Inflation

Nvidia told major customers that AI server prices will rise by more than 15% next year, driven by soaring HBM and DRAM costs, which now account for 25–30% of the system BOM. Contract manufacturers for Microsoft, Google, and Oracle have already begun passing those increases through.

Nvidia’s planned AI‑server price hikes could lift inflation expectations by signaling that **AI‑related cost pressures—especially soaring HBM and DRAM—are structural rather than temporary, with broad and immediate pass‑through already underway across major cloud providers.

Nvidia’s options‑implied earnings move has tightened alongside 2‑year inflation swaps, as both now respond to the same forces—the company’s outsized role in the AI cycle and the energy‑ and data‑center‑driven price pressures those swaps capture.

Nvidia has effectively become a macro force: its earnings, debt issuance, equity purchases, CDS hedging, and AI‑server price hikes now move both Treasury yields and inflation expectations, making the company a new transmission channel for financial conditions.

Figure 3: Nvidia implied move on earnings and short-term inflation expectations (%)

Source; NYSE, CME, Nvidia

Iranian sanctions, for now, are not seen as a major disruption because China has responded by warning it will retaliate if Washington expands secondary sanctions.

By invoking its “blocking rule” for the first time to order Chinese firms to ignore foreign sanctions, China is insisting it will keep buying the ~80% of Iran’s crude it already imports despite U.S. pressure.

Chinese banks are particularly vulnerable to Iran sanctions since they play a key role in oil sales through the Society for Worldwide Interbank Financial Telecommunication (“SWIFT”).

Figure 1: China H-share Financials versus S&P 500 Financials

Source: Shenzhen, NYSE

SWIFT is a global messaging network that transmits standardized instructions between banks, enabling the movement of funds through correspondent‑banking channels

Through these channels, SWIFT coordinates the use of Nostro and Vostro accounts to move dollars across institutions, carrying out instructions that enable cross-border liquidity and settlement, even though the actual money moves outside SWIFT itself.

This matters for U.S. Treasuries. When a foreign institution purchases a Treasury, the security settles on Fedwire Securities. Still, the cash leg of the transaction requires sending dollars to a U.S. custodian or clearing bank—typically routed through correspondent banks using SWIFT messages such as MT103 or MT202.

In practice, SWIFT‑enabled dollar transfers are the connective tissue that allows foreign buyers to fund Treasury settlement, meaning the Treasury market depends on SWIFT for the cross‑border dollar movement that underpins global participation.

Weaponizing the core wiring of global finance tightens dollar liquidity worldwide, prompting foreign institutions to pre-emptively hoard dollars or Treasuries, thereby amplifying volatility across the U.S. Treasury market. Such was the case in 2022 when Russia was cut off from SWIFT.

Figure 2: SWIFT payments in US dollars (% of total), Treasury Volatility

Source: SWIFT, Merrill-Lynch

When Russian banks lost SWIFT access in March 2022, cross‑border payments were immediately jammed as the removal of SWIFT’s messaging layer created severe friction in settlements and trade finance, delaying energy payments, forcing deep discounts on Urals crude, and disrupting term contracts and supply arrangements.

The cutoff pushed Russian and other non‑Western banks into slower, opaque correspondent-bank workarounds. This injected uncertainty into trade finance, commodity settlement, and cross‑border flows, lifting term premium and producing sharp volatility in long‑dated Treasuries.

That instability was then reinforced by freezing $300 billion in Russian reserves. This prompted global reserve managers to reassess the political risk of holding Treasuries and begin gradually diversifying toward gold, RMB assets, and non‑Western custodians, embedding a structurally higher term premium over time.

Druckenmiller warns that Treasury buybacks risk muting the market’s price signals. Still, the Russia SWIFT cutoff showed something even more fundamental: when payment plumbing breaks, the bond market can react more violently.

Seen through Druckenmiller’s lens, the implication of Bessent’s announcement cutting off countries from SWIFT is far greater for the Treasury market than debt buybacks.

The lesson is sharper: the integrity of Treasury pricing depends not only on avoiding artificial buybacks but on protecting the global payment infrastructure that allows capital to reach the market.

The dollar’s hegemony and Treasuries’ risk-free status, both pillars of SWIFT and the global system, are now questioned by the Treasury’s own announcements, leading to further erosion of the dominance of dollar assets.

Figure 3: SWIFT payments in Yuan and US dollar as % of foreign reserves

Source: IMF COFER, SWIFT

Currently holding close to $1 trillion in excess cash, the Treasury maintains excess cash at the Federal Reserve and closely coordinates the issuance of T-bills when the balance is increased or spent down during government shutdowns and debt ceiling crises (a process known as “extraordinary measures”).

The TGA is the Treasury’s operating cash balance at the Fed. All tax receipts, debt issuance proceeds, and other inflows go into the TGA; all government payments (Social Security, defense, contracts, payrolls, etc.) go out of the TGA. When the balance rises, liquidity is drained from the private sector; when it falls, liquidity is injected into it.

Treasury officials confirmed that Bessent has purposefully accumulated roughly $950B in the TGA, creating a large “war chest” of cash. This is not an accidental spike—it’s a strategic choice.

Figure 1: TGA and T-bills ($, billion)

Source: Sifma, US Treasury

When Treasury uses the TGA to buy back outstanding debt, there are three mechanical channels through which this affects markets and yields.

When the TGA is drawn down, liquidity flows back into the banking system as reserves rise, which can pull Treasury yields lower as banks deploy those reserves into bonds and can lift bank stocks because higher reserves effectively strengthen their capital buffers.

That said, using TGA funds for buybacks leaves net government debt unchanged, since the drop in the cash balance offsets the reduction in outstanding securities.

What does change is the mix of debt: buybacks ease near‑term bill‑issuance pressure while leaving long‑end yields untouched unless the buybacks themselves are large enough to move the market.

To that end, the Treasury does have a material toolkit at its disposal: $6.5 trillion in outstanding 10Y to 30Y bonds to buy back, 35% of which the Fed holds on its balance sheet.

The impact could be more significant this time than during Operation Twist in 2011–12, when the Fed sold bills and bought long-dated Treasuries, stripping roughly $650 billion of 6- to 30-year duration — essentially the same amount in 10-year-equivalent terms — from the market.

In the current scenario, if Treasury were to use the full $935 billion TGA balance to buy long‑end bonds, it wouldn’t remove duration risk, but it would still matter because funding those purchases would require bill issuance that is likely to run at roughly twice the pace of long‑end issuance.

If Treasury wants to maintain a high cash buffer, any TGA‑funded buybacks require future debt issuance to restore the balance. Treasury officials expect they have a window before the next debt‑ceiling constraint to rebuild the TGA if they draw it down.

Figure 2: Fed’s Treasury holdings and total amount of Treasuries outstanding ($, million)

Source: NY Fed, US Treasury

In 1986, Bill Gross cast the bond market as a battlefield where innovation and leverage were the new weapons of survival. Still, by 2013, he warned that quantitative easing had shattered the old order, pulling investors into a war of violent repricing as central banks held the commanding heights.

That imagery resurfaced over the weekend when the President suggested that the ultimate form of bond‑market intervention could involve the military itself—rhetoric that evokes wartime financial mobilization, when governments have forced private assets into public debt, suspended trading, and tightened capital controls.

In that scenario, the state replaces market discipline with government command. It turns price discovery into an instrument of national strategy—though under U.S. law, only a formally declared war or a congressionally authorized national emergency unlocks such extreme financial powers.

But in a scenario where the Iran war becomes prolonged and threatens the government’s financial stability amid illiquidity and volatility in the Treasury market, the state could use non-market power to maintain stability because financial tools are insufficient.

The Trading with the Enemy Act (TWEA, 1917) allows the President to control or prohibit financial transactions, freeze or seize assets, and direct capital flows. This law applies only during a formally declared war and has historically been used in WWI and WWII for full financial mobilization.

The New York Fed wrote a paper about the late 1930s and WWII, when Treasury officials insisted on stable long-term borrowing costs. The Fed adopted yield-curve control to guarantee predictable rates across maturities, ensuring the government could fund the war at fixed, low interest costs.

T-bills were anchored at ⅜‑percent, and long-term yields were “capped” around 2.5% and labeled as “war bonds,” issued at a predictable rate.

Figure 1: Treasury yields during WWII (%)

Source: New York Federal Reserve

Fixing the wartime Treasury curve forced the Fed to absorb unwanted bills, pushing investors into longer maturities. The regime broke when the Fed dismantled the ⅜‑percent bill peg, lifted caps on short‑term coupon yields as inflation rose and investor demand shifted.

Once rate fixing ended in 1947, short‑term yields jumped almost immediately—bill rates vaulting from ⅜ percent to nearly 1 percent within months—while longer‑term yields, freed from artificial support, drifted higher as investors unwound their wartime preference for bonds, producing a broad upward repricing across the curve (Figure 2).

Figure 2: Treasury Yields Post-Fixed Rates, 1945-1951 (%)

Source: New York Federal Reserve

What stands out from last week’s debt buyback announcement and the President’s comments is the symbolic escalation: policymakers openly acknowledging that conventional financial tools may be insufficient to stabilize a Treasury market under wartime stress.

Buybacks are a technical response to the President’s language, which gestures to the idea that sovereign credit, in extremis, becomes an instrument of national defense rather than a product of market negotiation.

That already seems to be happening: despite doubts about the impact of Bessent’s buybacks, 30‑year yields slipped about 3.5 basis points while T‑bills and short‑term notes climbed 2 to 6 basis points over the week — a small but telling sign the market is responding to intervention.

The goal of the debt buybacks could be, as in the 1940s, for the Fed to absorb unwanted T-bills that fund the buybacks and eventually adjust its balance sheet towards a long-maturity Treasury portfolio.

It is a different version of “Operation Twist,” but with a similarly potent effect on risk assets, and the markets are likely to continue to treat it as such.

Figure 3: Fed’s balance sheet composition 1938-1951 by maturity ($, billion)

Source: New York Federal Reserve

The fact that Treasury says it can do more, and the Fed says it can do more on the very same day, signals that some form of information asymmetry is at work between the two institutions.

For investors, it is about liquidity versus risk. Before the debt buyback announcement, the term premium climbed by 40 basis points in July as the FOMC press conference raised independence concerns, a shift that the Treasury’s latest move may have amplified.

Liquidity can fix a lot, and heading into the debt-buyback announcement, the 30‑year was suffering from severe illiquidity after its poorly received auction, with the on‑the‑run/off‑the‑run spread—yield difference between benchmark and previously issued-- collapsing and pushing the term premium higher (Figure 1).

This is why Treasury stepped in: the sudden bout of 30‑year illiquidity—driven by the post‑FOMC fallout and heavy hyperscaler‑related hedging—was choking long‑bond sales and risked stoking debt‑sustainability concerns.

Figure 1: Why the Treasury intervened: dislocation (basis points)

Source: US Treasury, New York Fed

Jackson Guidance

Warsh may bristle at the idea of Treasury actively intervening at the long end, because a yield curve showing a twisting flattening reflects not genuine policy expectations or economic data, but rather a forced transfer of liquidity from the short end to the long end.

It could muddle Warsh’s reaction function, which investors want even more clarity on. But to Warsh, he already explained, and why next week’s Jackson Hole speech may not reveal much more.

In the July press conference, he said that any central banker with labor markets near equilibrium is inclined to tighten when underlying inflation rises and to loosen when it falls, and although that reaction function is straightforward, as he explicitly said: “That’s my reaction function.”

Adding that, in crisis mode, offering highly explicit forward guidance—even at the risk of tying our own hands—can be prudent. Still, under more benign conditions, that level of clarity is worth reconsidering.

It will be a dominant theme in his speech precisely because conditions are benign.

As the Minutes noted — “some participants commented that financial conditions might not currently be sufficiently restrictive to facilitate a return of inflation to 2 percent” — Warsh may argue that financial conditions remain too loose to align with his reaction function for underlying inflation pressures.

As a result, markets now see the odds of a single hike—rather than a series—converging between September and December in both futures and prediction markets, reflecting expectations for a reaction function largely untouched by Treasury’s long‑end interventions.

Figure 2: Hike Odds converging (%)

Source: CME

Underlying inflation

The President’s move to drop the out‑of‑quota beef tariff marks a clear shift toward using trade policy to ease inflation, especially with food carrying a heavy 13.7% CPI weight. Beef fits a broader pattern: the November 2025 rollback on multiple food items and fertilizers.

Larger food tariff cuts—like removing a 20–40% out‑of‑quota tariff—can still reduce import prices 5–10% and retail prices 2–4%, implying another 0.2–0.4 ppt of potential food‑at‑home disinflation if remaining food and fertilizer tariffs are trimmed.

This could eventually feed into Warsh’s reaction function, since food shapes underlying inflation, but the impact is likely to show up only gradually as we move into 2027.

Figure 3: Beef with Beef inflation (Y/Y%)

Source: USDA

Bitcoin shorts were heavily crowded going into the White House Crypto Summit and before the Treasury debt buyback announcement. Those positions were then washed out in a historic squeeze, with BTC and ETH together seeing nearly $2.7B in shorts unwound, making up 92% of all liquidations.

The shorts are now less crowded, but new shorts may rebuild if BTC stalls near resistance (72–75K) amid new uncertainty around the Clarity Act’s passage. However, there is now some progress in the Senate.

Figure 1: Crypto squeeze

Source: Coindesk

Senate Majority Leader John Thune filed cloture on the motion to proceed on August 8, 2026, teeing up the first procedural vote on the Act for September 15 — the day senators return from recess.

Once the Senate returns, it has only about fourteen working days before midterm campaigning stalls any real progress, and any bill that does advance will still need to be reconciled with the House version and the Agriculture Committee’s parallel measure, S. 3755.

This vote is procedural, not final passage, but it determines whether the Senate can even begin debating the bill. Odds of passage in 2026 have recently fallen sharply on Kalshi, as profiting from crypto is a key political sticking point.

Other points of contention include how the bill addresses illicit finance, law-enforcement authority, and oversight of crypto network developers, as well as an unresolved dispute between crypto firms and traditional financial institutions over whether yield‑bearing deposits should be allowed.

The Clarity Act is therefore alive but fragile. The September 15 cloture vote is the decisive gateway: if it fails, the bill is effectively dead for 2026; if it succeeds, the Senate enters a compressed, contentious legislative sprint with ethics, AML, and stablecoin‑yield fights still unresolved.

The crypto industry is bullish about the bill’s odds as the Treasury rolled out proposed regulations on Monday to quickly implement last year’s crypto law, which established a framework for regulating payment stablecoins. Prediction markets see an 80% chance the Senate will vote on the Act by October (Figure 2).

Figure 2: Odds of Senate voting in October and January (%)

Source: Kalshi

CFTC Chairman Mike Selig warned that if Congress fails to pass the Clarity Act, the agency will move ahead on its own by creating a formal ‘crypto asset market’ regulatory category under existing authorities.

Selig said he’s directed CFTC staff to craft a crypto‑asset market structure under existing authorities — including rules that allow developers to legally offer protocols in the U.S. — even as the SEC advances its own Regulation Crypto Assets and both agencies continue to coordinate on how different digital assets should be classified.

Bitcoin is entering a moment where the regulatory fog that has defined the past several years is finally lifting — a shift that supports a move toward higher, more durable levels if the September 15 cloture vote doesn’t break sentiment.

Flows into Bitcoin ETFs — $164 million over the past 24 hours — and BTC options volume over the past 24 hours reaching $1.52B, with calls making up about 61% of activity, underscore this renewed market optimism.

Figure 3: IBIT fund assets and call volume ($, million)

Source: iShares

A strong economy, driven by both fiscal borrowing and investment, leads to higher rates as a matter of economic law. To change the dynamic of rising nominal GDP and rising nominal yields, the law must be rewritten.

So, it is worth noting that the intervention route is used to try to alter the laws of economics.

Yet, when FX intervention—used when exchange rates turn disorderly—and bond‑market intervention—used when market functioning breaks down and term premia overshoot—collide, the attempt to reshape investor behavior sends conflicting policy signals.

Mary Daly noted that long‑end rates reflect structural forces—geopolitics, fiscal sustainability, and now the debt‑financed global AI build‑out—so they offer no guidance on what the Fed should do with policy.

Yields matter as inputs, she said, but price stability is anchored in the short end, where markets clearly understand the Fed’s reaction function, leaving policy “in a good place,” a phrase for a neutral stance often used by Williams and consistent with her strong support for a July hold.

Thus, when Treasury intervenes through buybacks partly financed by additional T‑bill issuance—right where market expectations of the Fed intersect because bills price off Fed Funds—it effectively inserts new bill supply into the very channel that shapes the reaction function, altering how policy signals are received (Figure 1).

Figure 1: Fed expectations and T-bill issuance ($, trillion, cumulative)

Source: CME, Federal Reserve, US Treasury, Sifma

Treasury debt buybacks are not new; they have been running for months, with the long end steadily taking a larger share. But yesterday’s announcement effectively put a spotlight on the long end, turning it into the primary target (see Figure 2).

The tools that Treasury has at its disposal to address the Yen—the currency stabilization fund and the foreign repo facility—have limited firepower, estimated at ~$300 billion.

For debt buybacks, however, the Treasury has $6.5 trillion of current 10Y to 30Y Treasury bonds available, of which the Fed owns about 30 percent. Bessent could increase debt buybacks on the fly and that could “cap” yields around current levels, but there is not a free lunch.

Figure 2: Rolling Buybacks ($, billion)

Source: Lighthouse Macro, US Treasury

T-bill issuance could increase because debt buybacks are treated as cash outlays. Based on the total current market value of outstanding long-end Treasuries accounted for the Fed’s holdings, T-bill issuance could increase by $1.2 trillion should Bessent choose to buyback the entire long-end outstanding.

But such a pop in Bill issuance pushes the share of T-bills of the total debt above 25 percent. That is far above the Treasury Advisory Borrowing Committee recommendation (share <20%).

That could make the case the Treasury must consider increasing the size of 10Y and 30Y auctions; a likely maturity extension of the debt. Without a Congressional mandated sequestration of fiscal spending, the only way to control the deficit and $40T in debt is by extending maturity.

It is a sharpened sword of Damocles hanging over the Treasury market, because maturity extension—attempted unsuccessfully in 2023 and still not priced—would upend Bessent’s intervention strategy and risk a sharp backfiring in yields.

The reality is that Treasury buybacks simply retire older issues and replace them with new ones, which is liquidity housekeeping, not an outright purchase program. Unlike the Fed, Treasury doesn’t create money supply in the process.

Yet the market seems to interpret the announcement as a first step toward YCC, with the yield curve flattening by 10 basis points between 2Y and 30Y yields.

A key driver of the market’s reaction is the sharp rise in yields — the same dynamic that previously prompted repeated jawboning by Bessent and the President during the war and again in early 2025. With those bouts of verbal intervention proving ineffective, Bessent may now be turning to a new tactic to try to rein in long‑term rates.

Treasuries are trading this as a “twist flattener,” which conveys a macro message of short‑term yields rising on tighter policy or front‑loaded inflation risk, while long-term yields fall as markets price weaker growth and lower terminal policy rates, or a future policy reversal.

Figure 1: Yield curve twist

Source: US Treasury, Tradeweb

For traders, the math is straightforward: between September 9 and November 4, the Treasury will conduct $64 billion in buybacks of 10‑ to 30‑year maturities—expiring just after the Midterms—while simultaneously issuing $168 billion in new long‑end debt over the same period to replace what it retires.

The 2026 long‑end buyback ratio (~35–40%) is much larger than anything Treasury has done in prior buyback cycles — both the 2000–2002 program and the 2024–2026 liquidity‑support pilot, which had a ratio of 8- 12%.

Even though buybacks aren’t QE or YCC—which involve the Fed—the sheer ratio of long‑end buybacks to new issuance helps explain why 30‑year yields reacted so sharply.

For markets, there are two important considerations.

The Treasury is committed to “cap” interest rates, even bringing them down through intervention methods: the Yen and now buybacks. For a period, this strategy may succeed if the Fed does not proceed with QT, as its stake in outstanding long-maturity bonds remains above 35%.

Finally, the Treasury signals a “weaker” US dollar policy by propping up the Yen and bringing long-term rates down (the Dollar index is down by 0.7%).

This could positively affect risk markets, especially emerging markets (Brazil up 1.5%), hard commodities (Gold +$150) and crypto (+6%), as well as Tech/semis with substantial overseas earnings.

Figure 2: 30Y yield intra-day (%)

Source: US Treasury, Tradeweb

The Treasury’s move to double long‑end buybacks from $2 billion to $4 billion was purely a liquidity‑management adjustment for outstanding 30‑year bonds. Yet, markets treated it as a broader easing signal, with yields falling 10 basis points and semis/tech rallying sharply.

This highlights how tightly rate dynamics are now correlated with the AI‑and‑tech complex rather than traditional QE interpretations.

For expectations heading into today’s FOMC Minutes, this shift in correlation is crucial. The anxiety about bonds pushing toward multi‑decade highs has been anchored in fears of runaway fiscal deficits.

But a closer look shows the recent rise in yields hasn’t come from inflation or Fed policy repricing — it’s been driven by exogenous forces like the Hormuz shock and heavy long-end issuance by hyperscalers and semiconductor firms (see Figure 1).

The term premium has climbed about 40 basis points since late June, driven largely by the extra compensation investors demand to hedge the surge in hyperscaler and semiconductor issuance.

As supply goes global — with Google’s Kangaroo bonds pricing between 5.5% and 6.98% — issuers hedge these deals in local sovereign markets, tightening the link between 30‑year yields and semiconductor benchmarks.

Put differently, the Fed may be reacting to inflation for the wrong reasons. Long‑end yields are aligning more closely with the surge in AI-driven investment than with the Hormuz energy shock, suggesting financial conditions are tightening for reasons that have little to do with traditional inflation dynamics.

Figure 1: Breakdown in changes in yield (basis points)

Source: Markit, MarketAxess, Tradeweb, CME

But now the IPO by Unitree Robotics, priced on the Shanghai STAR market, could kick off the next round of global AI bond issuance, financing specifically robotics.

The STAR Market is effectively China’s launchpad for strategic tech IPOs, especially in areas where Beijing wants rapid domestic scaling.

China’s fully integrated robotics supply chain—controlling ~90% of global Neodymium–Iron–Boron (“NdFeB”), the alloy used to make neodymium permanent magnets, which are the strongest commercially available magnets and a critical component in humanoid robots because they power high‑torque motors and actuators.

U.S. robotics remains structurally underdeveloped, dependent on low‑cost Chinese platforms that are now restricted because critical components—motors, actuators, sensors, and especially NdFeB magnets —must be sourced domestically at far higher prices.

This cost pressure is amplified by a tightening legislative regime.

The American Security Robotics Act blocks federal procurement of Chinese humanoids. The FCC’s 2026 foreign‑robot import ban erased an $8B consumer robotics market and forced startups toward U.S. alternatives costing 3–5× more, and the NDAA Section 163 restricts Pentagon use of “adversary‑linked” robots without offering any industrial policy to expand domestic supply.

Layered on top, the Pentagon’s 1260H component restrictions—already applied to Unitree—extend from direct Department of Defense contracting bans in June 2026 to full component‑level prohibitions in June 2027, further raising BOM costs and constraining inputs.

Even as U.S. robotics production is constrained, hyperscalers are expected to push into robotics because embodied AI is the next compute platform — and they can finance it through debt issuance that startups cannot match.

Figure 2: China Tech versus US Tech issuance ($, billion, monthly data)

Source: Bloomberg Intelligence, Shanghai STAR

“Robo bonds” are, for now, non-existent, although Alphabet funds Everyday Robots (now folded into DeepMind). Microsoft supports integration in cloud and industrial automation, and Meta finances embodied‑AI research and robot learning datasets.

Finally, Tesla’s Optimus humanoid development is financed, for now, through general capex, but the company has a history of issuing convertible bonds.

The China Robotics Index is already moving in tandem with the global credit index—which captures hyperscaler and semiconductor issuance—underscoring that robotics is shifting into a capital‑markets asset class tied to national security, physical‑AI buildout, and ultimately macro variables like Treasury yields.

As robotics becomes both a strategic technology and a commercial platform, its influence on yields extends beyond traditional economics, even as fiscal‑deficit risks remain part of the backdrop.

Figure 3: China Robotic Index (normalized) and Global Credit Index (Total return index)

Source: Barclays and Soleactive Indices

Today, however, shorting JGBs is the most profitable trade of all. Since the Bank of Japan began normalizing rates in 2024, Japanese bonds are down 10 percent, while Treasuries and other G10 sovereigns are marginally up 5 percent due to rate cuts.

Yet, the story behind the big JGB short runs deeper.

In 2023, on a panel at the ECB conference in Sintra, Governor Ueda signaled there was a good reason to “shift policy” if it became “reasonably sure” inflation would accelerate the following year.

In that same year, Japan began its shift to a new solvency framework aligned with the global Insurance Capital Standard by moving life insurers and pensions, the biggest passive buyers of JGBs, to a market-based valuation of assets, with full rollout in 2026.

Japan’s 2023 solvency shift sharply reduced structural demand for ultra‑long-JGBs, heightened mark‑to‑market sensitivity, and pushed insurers and pensions toward shorter maturities and derivative hedging just as the BOJ began tightening.

In 2026, the full J-ICS rollout will further erode long-dated JGB demand by forcing liabilities to be fully mark-to-market, raising capital charges on long‑duration assets, and pushing insurers and pensions even more into shorter maturities—all while the BOJ signals faster rate hikes.

JGBs now carry a heightened “widowmaker” aura—a trade whose risks have escalated and whose repercussions increasingly spill over into global bond markets.

Figure 1: Widowmaker: JGBs are the only bonds in decline (index, normalized scale)

Source: Barclays Indices

A way this may happen is plain-vanilla Treasury Bills. Yesterday’s Treasury International Capital (“TIC”) data showed a substantial drop in Japan’s T-bill holdings in June. This may have been ‘planned’ ahead of a Yen intervention to avoid sales of longer-maturity Treasuries (Figure 2). (See Peter Boockvar for more TIC analysis).

China and Japan, followed by the UK, are the largest holders of Treasury Bills, with holdings totaling $460 billion, according to TIC data as of June. While this represents about 6 percent of total outstanding T-bills ($7T), which are mainly held by US domestic money market and bond funds and ETFs, the sudden drop in Japan’s holdings nonetheless reverberates.

The US deficit’s survival relies heavily on T-bill financing; with foreign holders reducing exposure, there is a risk of a “T-bill sudden stop.” With fewer foreign buyers of T‑bills, markets will begin to price an additional risk premium into Treasuries from the twin deficits—fiscal and trade.

Figure 2: T-bill sudden stop ($, billion)

Source: US Treasury TIC data

As Japan’s ALM overhaul and BOJ tightening take hold, 10‑year JGB yields have pushed up against a key resistance at 2.96% — the 38.2% Fibonacci retracement from the 1990 peak at 8.23% to the −0.28% trough in 2016/2019 (Figure 3) — a level that now carries technical and macro significance.

Japan also has a debt clock, which has crossed over $10.25T, an equally significant level as the US’ $40T mark. When JGBs break the 38.2% retracement, the historical correlations with Treasuries and other global duration markets suggest a broader unwind in long‑maturity positioning across Japan and beyond.

That move would likely weaken the yen further—already trading around 159.78—and could prompt additional Japanese T‑bill sales, adding pressure to long‑dated bonds in economies dependent on twin‑deficit financing.

Figure 3: 10Y JGB yield retracements (%)

Source: BOJ, Bloomberg

The driver is a run of softer data—payrolls, CPI, retail sales—that has steadily pulled the hawkish narrative down. And with no meaningful official action in either the yen or oil market headlines, conviction has grown that September won’t bring a hike.

In effect, the soft data has become the referee. And with the crucial yield‑oil correlation breaking down, markets see that data as a warning shot—one that could make a Fed hike look misguided.

Tom Lee at FundStrat makes a few salient points, including that a hike would undermine the Fed’s credibility.

A one-off hike looks more like a reversal of earlier cuts; tightening won’t dent AI investment while high long-end yields are already cooling housing and durables; and core PCE and CPI are easing year‑over‑year.

A hike would make the Fed look like it’s bending to pundit pressure—even though those “pundits” now include several voting members. Yet, based purely on member weightings—Chair counts three; voters two; regionals one—the Fed LLM shows a dovish tone is emerging as softer data is being acknowledged (Figure 1).

Figure 1: A dovish tone emerging (score, % odds)

Source: FedWatch Advisors LLM, Claude, Federal Reserve

It marks a meaningful shift on a day when prediction markets are pricing a real change in the war’s trajectory.

The market on whether the U.S.–Iran 60‑day negotiation window gets extended has collapsed—down 47 points to 16%—while the sister market on a U.S. invasion before 2027 surged 67 points to 18%. Invasion tail‑risk is being treated as a credible scenario rather than a fringe one.

Such an invasion scenario could easily tilt future inflation outcomes—even as some Fed voices turn more dovish and the President threatens strikes on Oman. Global inflation uncertainty remains high, with 10‑year yields discounted ten years from now in the US, Japan, France, and the UK ranging from 5% to 6.5%.

Figure 2: Odds of Invasion (%)

Source: Polymarket

The FOMC minutes should mirror that uncertainty. Hammack, Logan, and likely Schmid should appear with full arguments on persistence of inflation and credibility risk. Risk language should tilt clearly toward inflation, labor concerns should be downgraded, and AI demand plus energy pass‑through should be elevated to explicit upside risks.

Open dissent gives Warsh room to hold steady, speak more forcefully, and lean rhetorically tighter—even without a hike.

And the Minutes will likely read hawkish, given the strong link between the Staff Economic Projections’ diffusion index of Fed members seeing upside PCE risks and the LLM‑scored hawkish language embedded in the Minutes themselves (see Figure 3).

Figure 3: Minutes hawkish sentences and SEP upside PCE risks

Source: Federal Reserve

Bond yields jump on headlines from Oman because they signal a near‑term oil spike that could blunt the recent softening in inflation.

That keeps the bond risk premium climbing, reflecting renewed inflation worries and a Fed still leaning hawkish. Japan’s push toward faster, more frequent hikes adds another layer of global rate pressure, even as markets price no Fed move and a dovish tone emerges.

Figure 4: Bond risk premium (%)

Source: Japan MoF, US Treasury. Risk premium: 10Y yield ten years forward - 10Y yield (%)

It is a striking milestone not only because of the size of the debt, but because of the growth that has accompanied it. Since 1970, nominal GDP has expanded by an average of $470 billion per year, while federal debt has climbed even faster, rising about $760 billion annually over the same period.

The gap between the government’s borrowing speed and the economy’s growth speed appears in what the IMF calls the equilibrium interest rate—the borrowing rate that keeps the debt‑to‑GDP ratio stable, given nominal GDP growth and the primary fiscal balance.

This equilibrium rate determines whether debt dynamics remain stable, and it differs from the neutral interest rate, which reflects the level consistent with full employment and inflation at target.

With nominal GDP growing at 9 percent, a primary deficit of 2.6 percent of GDP, and debt at roughly 138 percent of GDP ($40 trillion on $29 trillion GDP), the equilibrium nominal interest rate is about 7 percent—far above the nominal neutral rate near 2.5 percent.

That contrast is why the $40 trillion moment may still feel manageable for markets: the U.S. is currently borrowing at a weighted‑average rate near 3.5 percent, well below the 7 percent threshold where debt-sustainability models suggest real stress would begin.

Figure 1: Equilibrium interest rate and neutral rate (%)

Source: IMF, Federal Reserve, US Treasury

The stock market ended the week essentially flat to slightly positive, even as grim headlines about the highest yields seen at recent 10‑year and 30‑year auctions failed to rattle Treasuries. Yields finished lower on the week, a sign that markets continue to brush off debt‑related anxieties.

At the same time, expectations for higher interest rates are building. Both the University of Michigan and the Conference Board surveys show a clear rise in the share of respondents anticipating rate increases over the next year—from 44 percent to 53 percent.

The Treasury forward market also reflects this shift, with discounted nominal yields pushing the 10‑year rate toward the critical 5 percent threshold (Figure 2).

Forward expectations further out price the T-bill and 10Y around 6.25% in the next ten years. The equilibrium interest rate to keep debt stable in the next ten years is at about the same level.

One reason is that if this 9% annual GDP growth is sustained, the debt-to-GDP ratio would drift toward about 92 percent. This is a factor why markets in general are not ‘worried’ right now about what looks like exploding Federal debt levels.

Figure 2: Interest rate expectations (%)

Source: University of Michigan, Conference Board, CME

Despite strong nominal growth, bond‑market anxiety persists because a deeper strain is building beneath the risk‑free T‑bill: as the deficit is increasingly financed through short-term bills, the accumulated interest paid on the debt has now surpassed $1.1 trillion, even though the average interest rate on Treasury debt remains just 3.5 percent (Figure 3)

Figure 3: T-bills and interest costs ($, trillion)

Source: US Treasury, CBO

A “risk-free crisis,” however, could emerge when $7.9 trillion in money‑market cash begins rotating out of safe, high‑yielding T‑bills just as real rates fall due to stickier inflation, colliding with $11 trillion in bond‑fund assets still delivering negative duration compensation.

As that simultaneous shift drains liquidity from the front end while forcing long‑duration portfolios to absorb losses, it creates a negative feedback loop in which outflows, rising term premiums, and tightening financial conditions destabilize even the traditionally “risk‑free” parts of the sovereign market.