While some believe that KKR’s investment in Atlantic Aviation as the K-shaped economy on full display… I had almost the opposite reaction. I thought the investment and results was kind of genius.

I don’t look at it like ‘oh, more wealthy people are flying private’, and I don’t view this as private equity executed another roll-up. Honestly, what caught my attention was the structure of the bet. KKR did a roundabout thing here… they didn’t buy the aircraft manufacturer, the fractional operator or the charter company, they bought infrastructure like fuel, hangars, ramp space, ground handling and airport access. It might have even bought time.

KKR acquired Atlantic from Macquarie Infrastructure in 2021, and at that time, Atlantic operated 69 locations and generated $276 million of EBITDA in 2019. KKR paid 16X times that EBITDA, so frankly, this wasn't some distressed asset picked up on the cheap at all. Atlantic also came with something less obvious on a financial statement: these are businesses built on long-duration, hard-to-replace leases. Macquarie's own disclosures, going back to when it first built out the platform, had described the structural barriers surrounding the FBO business. The thing that gets overlooked is that airports have finite physical space, airport authorities control development and competition, new facilities require substantial capital, and another operator cannot simply open across the street when the economics get attractive. That didn't stop KKR paying a premium for scarcity. The interesting thing here is that they spent the next 5 years making scarcity productive.

Read more

It’s a weird thing to get excited knowing what is happening behind the scenes in the workforce, I get that… but at the end of the day, the workforce will tell you more than financials will.

Yes, I saw you roll your eyes when you read that, but the financials are more like a crime scene to me… even capex. Financials primarily record the financial consequences of decisions… what management PLANNED to do with it and plans change all the time.

Me? I like getting there while they’re still buying the shovel. My focus is in who started digging months ago. In this case…One stands out for a mile.

So the framework agreements that were signed by Lockheed and General Dynamics with the Department of Defense today initiated 7-year multiyear procurement contracts to effectively expand the production of the components that will quadruple the THAAD and triple the PAC-3 MSE interceptor program capacity. Today’s agreements are the component layer: motor cases, seeker housings and midsections, shroud deployment systems. The interceptor-scale signal came earlier. Lockheed signed PAC-3 and THAAD frameworks in January, took a $4.7 billion PAC-3 UCA in April, a ~$35 billion THAAD UCA in June, and a late-July PAC-3 modification that took the multiyear interceptor contract to about $58.6 billion. This is obviously a lot of missiles. This also gives manufacturers visibility to invest in plants, tooling and suppliers… and of course people.

One of the most important statements I saw was from the Under Secretary of War for Acquisition and Sustainment, Michael Duffey- “Today’s announcement represents an example of the evolution of how we partner with industry to expand the Arsenal of Freedom, outpace emerging threats, and ensure the Warfighter never faces a fair fight. General Dynamics and Lockheed Martin have answered the call. We are cutting red tape, shortening timelines, and rapidly scaling our domestic manufacturing capacity to deliver the advanced air and missile defense capabilities.”

Note the words he used- outpace, shortening timelines, rapidly scaling. The real question here is if Lockheed and General Dynamics are entering the 7 year window from the same starting line… and as you may have guessed it… they are not.

Read more

At the start of this year, 3 major airlines were adding people considerably faster than they were losing them. By July, that was no longer true.

Let me kick this piece off saying that jet fuel has become a much larger problem than crude alone would suggest. The U.S. Gulf Coast jet fuel/WTI crack has risen roughly 210% year-to-date. Jet fuel itself has just about doubled. More recently, the divergence has become particularly notable because what we have seen over the last 3 months is that WTI has fallen by about 12% and jet crack rose by near 50%.

In a cost shock, the winner isn’t who hurts least… but rather who makes everyone else hurt more.

So sure, crude has eased up a bit, but the product airlines actually burn has not and that is creating a fairly unusual operating issue. American, Delta and United cannot control the crack spread. What they can control, within limits is the organization makeup as a whole. Earlier this year, workforce formation across all three carriers was overall pretty positive. I saw more starts than exits. It looked like summer 2026 might have been one for the books. The last few months, that all began to change and workforce flows changed… they favored replacement and, in some cases, contraction. My initial thought was that the airlines are simply becoming cautious as fuel becomes more expensive, but that didn’t feel like the whole story.

I’ve said it once, and I’ll say it a thousand times… anyone can hire when times are good… and the real gossip is who survives when they’re not…. let alone, who might be facing layoffs.

Read more

I’ve been mapping this for weeks and much to my surprise, the map got bigger than I expected. So let’s start with the announcement that kicked all of this off.

Formula E and PIF expanded Driving Force, an education program built around STEM, sustainability, and electric motorsport. The latest version is actually pretty neat… it uses Formula E’s new GEN4 car to teach kids not just how the technology works, but what the careers attached to it look like.

And sure, this is easy to file away under it’s a kids program, it’s not Formula One, etc.

But I kept lingering on one question:

Why is one of the largest sovereign wealth funds on earth interested in teaching children about careers in electric mobility?

So down the rabbit hole I went. When you put Driving Force next to what PIF is doing across automotive manufacturing, EVs, autonomous mobility, gaming, esports and motorsport, it stops looking random. Then I got into my favorite… the workforce data. That’s where this got really good.

Everything I can see says PIF is building the human capital underneath an industry before that industry is anywhere near scale. That was my thesis going in, but I did not expected there would be such a big mechanism sitting underneath it:

PIF is using portfolio companies to import industrial knowledge into Saudi Arabia.

Not capital alone - Not knowledge alone. All of it, together, within a very unique sequence. These companies are telling us in all their press releases. It just seems like nobody is really paying attention.

Because I love digging, I read PIF’s 2026–2030 strategy that came out a few days ago and it’s quite useful… unusually plain-spoken for a document of its kind as well.

The fund calls the next phase Value Realization. After years of rapid expansion and capital deployment, it’s pivoting toward integrated economic ecosystems, even more strategic-asset optimization, and considerably more mature domestic value chains. PIF describes itself as an architect and steward of platforms that let others scale.

But the line that really stood out to me is that its strategic ecosystems are meant to accelerate capability transfer across industries. That is remarkably close to what the workforce data is showing me independently.

To catch you up on what is happening:

PIF invested in Lucid.

PIF created Ceer with Foxconn.

PIF put $3.5 billion into Uber back in 2016.

PIF built Savvy Games Group.

PIF got deep into electric motorsport.

And PIF just led the consortium taking Electronic Arts private in a $55 billion deal alongside Silver Lake and Affinity Partners.

Every one of those can be analyzed on its own. Most people stop right there without realizing what is moving between these assets.

Capital, of course… but also technology, customers, suppliers and manufacturing processes… oh and don’t forget people. It really comes down to people. It isn’t just a case of hired roles… it’s the pedigree, and talent caliber that comes into place.

We’ve all seen the news. Lucid is in the middle of a big changeup…. cutting costs, trimming inventory, moderating production, chasing $1.4 billion in cash-flow improvements. And look, there's been a lot of noise around this company lately… including chatter about bankruptcy that has no business being in the conversation. A company sliding does not commission a stamping line...it doesn't add senior quality engineers and it doesn't keep the delayed-draw facility mostly untouched. Lucid is running a hard reset, and a hard reset is what a company does when it intends to still be here.

The workforce shows it plainly and makes me even more excited at what is in the works. In the dataset I reviewed, Lucid’s observed headcount fell roughly 10% between December 2025 and August 2026, and engineering took the worst hit.

April was a bit more brutal. Engineering exits directionally looked to outnumber starts, with simultaneous double-digit declines across operations, business management, sales support, marketing, HR and IT. Hiring then nearly stopped. By midsummer, engineering starts were running at a small fraction of exits.

All in all, headline headcount isn’t what caught my attention. While the broader organization gets cut back, Lucid keeps putting resources behind its Saudi operation.

AMP-2 has moved from construction into industrialization…. stamping, body, paint and final assembly systems all being installed and commissioned ahead of production trials. Lucid has also been expanding AMP-2 toward completely built-up vehicle production and substantial additional capacity. Midsize production is still planned to start there first.

So I stopped asking where Lucid is cutting and started asking something different:

What is Lucid refusing to cut? And then I dug into the people.

Read more

ORGANIZATIONAL ALLOCATION SERIES · JULY 2026 · SHAZ US

Workforce Intelligence · Pre-Earnings Execution Read

Sharon AI slowed up hiring this summer.

It closed a US$1.6 billion financing during the same period.

For the six months prior, the company grew its workforce by 175%. It recruited infrastructure leadership out of Microsoft, AWS and NexGen Cloud. It appointed a chairman who ran Telstra and a chief financial officer who spent two decades at Macquarie. By every observable measure, this was an organization being assembled at speed. Then the capital landed, and the curve went flat.

Organizations disclose what they believe through what they staff. This one has told the market it will have more than 62,000 NVIDIA GPUs deployed by the middle of next year, and that revenue against a US$1.32 billion agreement begins in the first quarter of 2027. That is the gap that does not appear in a price target.

The five analysts that this frameworks watches that cover SharonAI Holdings (NASDAQ: SHAZ). All five rate it Buy or Overweight. The average twelve-month target is $126.06 against a last price of $46.94… implied upside of roughly 169%.

None has cut. None has moved to Hold. The targets were published between 16 and 23 July, after the major contract announcements, so they are current. Meanwhile the stock fell about 20% on 28 July alone, and the options market is pricing implied volatility above 110% into the 6 August earnings call.

Two groups of professionals are looking at the same disclosures and reaching opposite conclusions. The conventional response is to build a third valuation model and adjudicate. That produces no edge: every covering analyst already has a model, and the models are not the disagreement. The assumptions underneath them are.

The framework here rests on one proposition: organizational commitments precede financial results by two to four quarters. Where narrative and behavior align, the announcement has operational substance behind it. Where they diverge, the divergence is the signal.

Three behavioral measures apply here:

The framework establishes direction, not outcome. It cannot forecast a quarter. What it can do is determine whether management behavior is consistent with the strategy being sold to investors a different question, and an earlier one.

Signal one: velocity

Indexed to July 2025 = 100:

The organization grew roughly 175%, and effectively all of it occurred in a six-month window between December 2025 and May 2026, bracketing the February Nasdaq listing.

The series then flattens. Two consecutive months of zero net additions, beginning the month the company closed an oversubscribed US$1.6 billion private placement anchored by Situational Awareness and Oaktree-managed funds.

Four readings are consistent with the data:

Read more

ORGANIZATIONAL ALLOCATION SERIES · JULY 2026 · PLNT US

Workforce Intelligence · Equity & Credit Cross-Read

The analytical gap under examination: the constructive valuation case still requires confidence that membership growth, club productivity and development eventually stabilize. The workforce does not yet show a broad organizational rebuild behind that stabilization… scarce hiring capacity is being directed toward a slower-growth, digital-retention, franchise-management model. Whether that resolves into contained repair or sustained reacceleration should become visible in the workforce before it is fully visible in reported results.

The divergence is strong. Here is what’s priced versus what’s being built.

The starting point is the divergence itself. The equity is down by half this year, and the most-followed sell-side shop stays Overweight. That target does not assume a full return to the old expansion algorithm… it is built on roughly 3–4% membership growth, about 4% annual development and mid-single-digit revenue and EBITDA growth. But even that slower path requires believing membership, club productivity and development stabilize. The framework focuses on the data and if it speaks to whether the organization is being built to deliver that stabilization. It is a read on behavior, not valuation, and it does not yet corroborate the constructive case.

The bull case is that this is a fixable marketing stumble and the compounding resumes. It may be. But the organizational build the framework would expect ahead of reacceleration… which looks like recruiting, people operations, field leadership, program management, but it is not yet visible. (It should be.) Instead, scarce hiring capacity is being directed toward a turnaround of the current model and toward capabilities relevant to a franchise system under stress. Read forward, the workforce is more consistent with a slower, more capital-intensive and more digital recovery than a clean return to the prior algorithm.

Ok, so why does this one stand out? Simple. Markets price stories. Long-term returns come from understanding execution. Traditional analysis assumes management allocates capital efficiently… The workforce dataset assumes something narrower and more useful, that management allocates capital honestly. Not in what it says. In what it does.

Hiring is among the most forward-looking discretionary budgets management sets. Expansion hiring often reflects the capabilities management expects to require over the coming twelve to twenty-four months; replacement hiring largely reflects the organization it already has. Distinguishing the two is the whole exercise.

And hiring is typically legible… a senior hire arrives with a resume that indicates what capability was added and from which operating playbook. Earnings report what happened. The allocation of hiring capacity indicates what management is preparing to require. That distinction matters.

The pace of expansion is decaying.

Read more

Read more

The real money is made earlier… when capital is quietly being forced to move and most people still think nothing has changed. That is the work. I track where capital is forced to move before markets fully price why, and workforce …

Read more

The earliest stage of any defense production cycle is not procurement, it’s staffing.

Multiple geopolitical tensions are currently forcing governments to reassess weapons stockpiles and industrial capacity. What is less visible is how the defense manufacturing system itself is responding to that pressure.

There is a persistent mistake in the way most observers attempt to understand the defense industry. They watch the wrong signals.

Public discussion tends to revolve around geopolitical headlines, government budgets, and contract announcements as if those events drive the industrial behavior of the sector. In reality they tend to be the final stage of a much longer process. By the time a procurement contract is announced or a defense company references backlog expansion on an earnings call, the underlying industrial adjustments that make that production possible have already been underway for months.

Weapons production isn’t just policy or political… it is industrial at the core.

We are witnessing the repositioning of the defense industrial base.

Factories must be staffed before production lines can scale.

Companies operating inside the manufacturing pipeline were already preparing their workforce for increased throughput well before the political narrative caught up.

Engineering teams must be assembled before complex systems can be integrated and validated. Testing environments must exist before those systems can transition into operational deployment. The defense economy moves in stages, and the earliest of those stages almost always appears in workforce architecture long before it appears in financial disclosures and it is crucial to understand what is coming.

When companies begin adjusting the composition of their engineering workforce, particularly in highly specialized manufacturing roles, they are straight up telling us where capacity will exist in the future rather than describing what exists today.

Over the past several months I have been examining hiring composition across multiple U.S. defense contractors. Not just the number of engineers being hired, which fluctuates constantly in large technical organizations, but the specific categories of roles those companies are choosing to expand relative to their historical baseline. The workforce is a living, breathing entity and can really tell us cyclical patterns and structural changes. Modern conflict is a contest of industrial capacity.

What emerges from that analysis is a pattern that is difficult to dismiss as routine hiring noise. This is typically the noise you hear surrounding new contracts or just any general posts on social media, especially concerning the overall labor market or even the geopolitical scene. Multiple segments of the defense industrial ecosystem appear to be expanding simultaneously, not just inside one contractor, and not within a single weapons program, but across the upstream manufacturing layer, the midstream integration layer, and the downstream battlefield infrastructure layer.

When those three segments begin adjusting at the same time, the industrial system is usually preparing for something larger than a standard procurement cycle.

Read more

Oracle is about to spend tens of billions on AI data centers and GPUs, and when infrastructure becomes the primary asset of the business, labor gets repriced. The connection between labor and repricing and AI is the #1 question hedge funds are asking me right now.

Bloomberg is now reporting thousands of layoffs and hiring freezes across the cloud division. I flagged those signals again on February 19th.

Being early to these shifts matters. Because the layoffs aren’t the story at all… they’re just the mechanism.

Ok let’s walk it back real quick when I published an analysis Jan 2nd Oracle’s 23 Year Loop arguing that something unusual was happening inside Oracle’s workforce, the signal I was watching wasn’t the headline layoffs everyone focuses on when tech companies restructure. It was the composition of the workforce itself… the percentage shifts in who was being exited out of the system and who was being rebuilt to replace them.

So, when Bloomberg reported that Oracle is preparing to cut thousands of jobs across the company while simultaneously accelerating one of the largest AI data-center expansions in its history, most readers will treat those two developments as separate stories.

They aren’t.

Read more

With tensions in the Middle East rising again and energy markets reacting in real time, the instinct in financial markets is to focus on oil… and that instinct is understandable.

If you want the earliest read on whether this remains a “logistics shock” or becomes a structural repricing, stop staring at crude and start looking at the term structure of op…

Read more

They are asking which model is better, which chip is faster, which company will beat quarterly expectations.

That is not how enterprise spend is routed.

Enterprise spend is routed through authority.

And if you are allocating capital across long-duration assets… across …

Read more

In November 2025, two consecutive months of hiring data indicated a material shift in Palantir’s incremental workforce layering. Hiring intensity declined into the low-teens percentage of its prior expansion velocity, implying that incremental hiring had fallen approximately 85–90% relative to the earlier build phase. This does not appear to be an isola…

Read more

If the largest private employer in the country can grow without adding workers, the benchmark for a “healthy” labor market changes.

Walmart told investors something far more consequential in 2025 than probably anything they could say in their filings… it no longer expects labor to scale with growth.

“When we look out two years, three years, five years… w…

Read more

You know my POV on what is really happening in the workforce. If you are new, don’t worry, you will quickly get up to speed. If something is genuinely transformative, you don’t need to shout about it every earnings call. You don’t need to keep telling people it’s inevitable. It shows up. And structurally, it’s how work actually gets done. It shows up in role clarity, in whether teams stabilize, in whether backfills resume, in whether people feel like they’re building something durable or just bracing.

And right now, when I look inside companies… not at stock prices, not at narratives, not at AI demos… AI does not look like clean productivity. It looks like pressure. It looks like organizations trying to run hotter with fewer people and calling that progress because a model says it should work.

That’s the part that gets me.

What I’m seeing is not companies suddenly operating with more slack or ease. I’m seeing companies compressing the people inside them and then laundering that compression through technical language until it sounds like destiny instead of a choice. And when you start there, the workforce outcomes are predictable… even if the market outcomes lag.

Let me be very explicit about what I actually track, because this matters. I don’t look at headline headcount numbers. Those are blunt, lagging, and easily managed. I look at job architecture. Which roles disappear. Which ones get merged. How spans of control widen. Whether attrition gets backfilled or absorbed. Where contractors appear as pressure valves - because DAMN y’all - it’s such a tell about company health. One entire industry is specifically consistent in their patterns right now, and historically we didn’t see this before.

Across tech right now, the pattern is consistent.

Mid-layer execution roles are really starting to thin out across the board. Managers are seeing twice as many direct reports, and the roles that used to be a bit more distinct are fast becoming hybrid positions with broader scope and less support. Don’t get me started on the ChatGPT style vague job descriptions either… and what I am hearing in the rumor mill is that accountability is harder and harder to trace, which ironically doesn’t help any employee.

Seriously… when productivity is real, you don’t need heroics, you don’t need people doing three jobs. So what I am seeing in the workforce data is consolidation justified by tooling rather than proven output. It’s a weird world we live in right now. AI is being layered on top of work that was already close to capacity, and leadership is assuming the technology will close the gap before the strain shows up.

That assumption is doing an enormous amount of work inside these organizations.

This is where narrative laundering really starts to matter. Narrative laundering is when uncertainty gets cleaned up by a model so no one has to own the downside. It’s not about lying. It’s about taking something operationally messy and reframing it as technically inevitable.

“The model says we have to.”

“This is where the industry is going.”

“We don’t have a choice.”

Those phrases aren’t neutral. You’ve heard CEOs say things just like that recently right? It feels like the new 'cool kids phrase.’ Once these are accepted, responsibility is dunz-o.

AI is being sold as an option-expanding technology. But the way it’s being implemented is option-constricting.

There are some CEOs that shall remain nameless that come across not quite as a villain, but as a signal amplifier. When they goes on stage and talk about the future of compute, markets hear destiny… boards hear inevitability… executives hear cover. Their messaging doesn’t cause layoffs. What it does is synchronize assumptions across companies without coordination. It sets the external clock. It gives internal decision-makers permission to align timing. But what I hear- lies, lies, and more damned lies.

Read more

Snowflake is in one of those moments.

This is not about revenue collapse at all.. frankly it is not even about failure either. It is a story about a company that has reached the phase where growth narratives collide with margin reality, and where the word “innovation”… which was once an asset - just becomes a liability when it no longer translates into operating leverage.

What follows is not a prediction based on rumors. It is an interpretation based on structure.

• Workforce behavior.

• Cost signaling.

• Leadership actions.

If you are looking for a simple claim - “Snowflake will lay people off in February” - you are missing the point. The more important story is why February makes sense, who would be affected first, and how this kind of adjustment actually happens in modern companies that care deeply about optics.

Read more

What’s actually resha…

Read more

Read more

Adam Goldstein missed a CES panel he was scheduled to speak on. The interviewer said he dropped out… but I see it as he avoided a room that would have forced Archer to be priced like every other eVTOL story… timelines, autonomy wish-casting, and certification “roadmaps.”

That panel instead was led by the people who actually live in regulated flight reality including Amazon Prime Air, describing live operations across multiple U.S. metros - Wind River walking through the safety certification stack used in aircraft systems - and a technologist describing how regulators treat “AI” when a human life is on the line. Goldstein’s absence changed the panel’s format, opened up live Q&A, and put constraint front-and-center.

That same day, Archer emphasized a different anchor… the NVIDIA partnership announcement was framed as aviation AI, backed by IGX Thor, safety-capable compute, and a clean story for investors who want a single sentence they can repeat.

If you want my contrarian take: the NVIDIA announcement is less about what Archer can do, and more about how Archer intends to be evaluated. The market is pricing advanced air mobility as a technology race. It’s still clever.

Now I’m gonna unpack the part investors keep missing… like the workforce composition, certification constraints that lock labor demand into short bursts, and why a hiring surge can be a late-stage signal that precedes role consolidation.

If you’re buying eVTOL on “autonomy” and “AI,” you’re modeling the wrong variable.

I like this phrase- Both Archer and Joby are hiring aggressively. That fact alone is meaningless.

Read more

What is actually happening inside the automotive and mobility sector is not exactly a consumer upgrade cycle. It is basically reading like a balance sheet- driven reallocation of capital and labor away from physical iteration and toward software systems that compound cost over time. Per the experts- autonomy, SDVs, and AI copilots… these are not endpoints, but rather you need to think of them as mechanisms.

Capital is no longer chasing moonshots for their own sake.

Conversations at CES made this totally clear, even if unintentionally. Capital is flowing toward architectures that shorten development cycles, reduce recall exposure, and convert one-time vehicle sales into long-lived software platforms. This change matters far more for workforce design than it does for marketing decks.

The defining characteristic of the SDV (software-defined vehicle) is not over-the-air updates or flashy interfaces. It is ownership of the stack.

The interesting game play here is that OEMs that control their software architectures can change behavior without touching the factory line, correct defects without mobilizing dealer networks, and monetize incremental improvements years after delivery. Talk about killer margin stability and operational leverage!!!!!

From a labor perspective, this changes where work happens. Historically, vehicle programs absorbed enormous human effort late in the cycle like physical prototyping, tooling changes, validation loops, and post-sale remediation. Also think along the lines of virtual modeling, simulation, and digital twins. Entire layers of downstream engineering effort are being eliminated or pulled forward into abstraction-heavy phases that require fewer people but higher skill density. That is the key here. Remember when I said specialists over generalists… this is what I am talking about… well one part of it anyways.

Read more