FY 25 Year End: An Audit of Signals, Capital, and Constraints
Theme: Fiscal Year 2025 in Full
Published January 2, 2025
This newsletter is for entertainment purposes only and is not financial advice. 🙂
In This Issue
I. What FY25 Was — and Was Not
II. FY25 in One Framework: Capital, Not Headlines
III. The Core FY25 Insight: Capability Ran Ahead of Reality
IV. Where the Economy Actually Softened
V. Why Markets Didn't Break (Yet)
VI. Why Q3 2026 Matters
VII. What We Did With Capital in FY25
VIII. What Carries Into FY26
I. What FY25 Was — and Was Not
This year-end close-out is not a forecast, a victory lap, or a market call. It is an audit.
FY25 was a year defined less by resolution than by acceleration — particularly in artificial intelligence capabilities — and by the growing realization that physical systems, financing structures, and organizations adapt more slowly than technology itself. The purpose of this letter is to document what we observed, how we interpreted those signals, and how we chose to respond with capital.
It is important to be explicit about what this review is not. It is not an attempt to time markets. It is not a declaration that a downturn is imminent. And it is not an argument that innovation has stalled. Markets can remain resilient for long periods even as underlying pressures build, and innovation can advance meaningfully without immediately translating into earnings or economic output.
Instead, this close-out reflects a disciplined effort to separate capability from deployment, investment from monetization, and narrative momentum from measurable outcomes. Throughout FY25, the dominant story across technology and markets was one of extraordinary promise. Running beneath that story was a quieter, more structural set of constraints — power, financing, labor, integration, and organizational capacity — that do not bend simply because enthusiasm is high.
Our goal throughout the year was not to outguess those narratives, but to observe where they intersected with reality. Where signals were strong and durable, we paid attention. Where evidence remained incomplete, we exercised restraint. And where uncertainty persisted, we preserved optionality rather than forcing conviction.
Editor's Note — This letter looks backward, not forward. It emphasizes process over prediction. And it treats uncertainty not as a flaw in the analysis, but as a condition that demands patience.
II. FY25 in One Framework: Capital, Not Headlines
If you judged FY25 by headlines alone, you'd think we spent the year oscillating between imminent collapse and inevitable techno-utopia. Depending on the day, everything was either about to break — or about to change the world. Reality, as usual, was quieter.
Markets spent much of the year pricing future promise, especially around AI, while the actual economy moved at a far more human pace. Manufacturing cooled. Freight slowed. Hiring became selective. Consumers showed stress in specific places, not everywhere at once. None of this made for great hype, but all of it mattered.
So instead of trying to summarize FY25 with index levels or narrative arcs, it's more useful to look at the year through a simple capital framework: what was created, what was taken off the table, and what was intentionally left flexible.
Capital created is the obvious part — gains realized when positioning and timing aligned. Capital withdrawn is the part most commentary skips over, even though it's just as important. Pulling cash out isn't failure; it's acknowledgment that capital exists to be used, not endlessly optimized on paper. Optionality is the third leg: capital left liquid, uncommitted, and deliberately boring, waiting for clarity instead of chasing momentum.
FY25 wasn't about being fully invested or constantly active. It was about staying aligned with evidence while the bigger questions remained unanswered.
Editor's Note — FY25 didn't resolve the story. It set the conditions for what comes next.
III. The Core FY25 Insight: Capability Ran Ahead of Reality
The defining feature of FY25 wasn't that artificial intelligence failed to deliver. It was that it delivered too quickly — faster than the systems meant to absorb it.
AI capabilities advanced at a remarkable pace. Models improved. Costs per task fell. Use cases multiplied. None of that is in dispute. What became increasingly clear, however, is that progress at the software layer does not automatically translate into progress everywhere else. Power still has to be generated. Data centers still have to be built. Financing still has to clear committees. Enterprises still have to integrate new tools into messy, human organizations.
In other words: capability scaled faster than deployment.
That gap showed up in subtle ways throughout the year. Infrastructure projects moved forward, but more slowly than headlines suggested. Enterprise adoption continued, but often in pilot form rather than at scale. Spending persisted, but with rising scrutiny around return on investment. Meanwhile, the supporting systems — labor, energy, permitting, capital markets — revealed themselves to be less elastic than the narratives assumed.
This isn't a criticism of innovation. It's a reminder of physics and economics. Software can iterate in weeks. Physical infrastructure cannot. Capital allocation has friction. Organizations change slowly, even when the tools are compelling.
The question was never whether the technology worked. It was whether the surrounding systems could keep up. By year's end, the answer was still unresolved.
Editor's Note — As capability surged, constraints became more visible, not less. That tension is the real takeaway from FY25.
IV. Where the Economy Actually Softened
One of the harder parts of FY25 was that the economy never gave a single, dramatic signal that something had broken. Instead, it did something far more common — and far more confusing. It softened unevenly.
Manufacturing was an early example. Orders slowed. New business became harder to come by. Survey data drifted lower, not catastrophically, but persistently. Nothing screamed collapse. What it did suggest was caution — the kind that shows up when companies stop expanding aggressively and start managing what they already have.
Freight and transportation told a similar story. Volumes cooled. Shipping activity flattened. Rail and trucking data pointed to fewer goods moving through the system, even as consumer spending headlines remained upbeat.
Labor followed, but selectively. White-collar hiring tightened first. Layoffs increased in professional, technical, and corporate roles, often framed as "restructuring" or "efficiency." The result was a job market that looked healthy in aggregate while feeling fragile to anyone inside it.
Consumers showed stress in similarly specific places. Auto loan delinquencies rose. Credit growth slowed. Retailers dealt with rising returns and heavier discounting. Food bank usage increased in certain regions.
Economies rarely flip all at once. They fray at the edges first. FY25 wasn't about broad contraction; it was about selective pressure — pressure that didn't force outcomes yet, but made future ones more sensitive.
V. Why Markets Didn't Break (Yet)
Given everything that softened beneath the surface, a reasonable question lingered throughout FY25: why didn't markets react more forcefully? The short answer is that markets weren't blind to the pressure — they just weren't forced to confront it.
Liquidity remained available, but increasingly concentrated. A narrow set of large, highly liquid names absorbed the bulk of capital flows, masking weakness elsewhere. Beneath the surface, participation thinned. Defensive sectors quietly outperformed. Risk didn't disappear; it reorganized.
At the same time, narratives carried more weight than usual. AI investment became a kind of gravitational center for optimism — a long-duration story capable of absorbing short-term inconsistencies. As long as spending continued and expectations pointed forward, markets were willing to tolerate friction in the present.
There was also no forcing mechanism. Credit stress existed, but it stayed compartmentalized. Layoffs rose, but not explosively. Corporate caution showed up as selectivity rather than retreat. Without a clear catalyst — an earnings shock, a financing failure, or a liquidity event — pressure accumulated without release.
Markets don't typically reprice because conditions worsen gradually. They reprice when something breaks alignment. FY25 never crossed that line. The system bent, but it didn't snap.
VI. Why Q3 2026 Matters
FY25 ended without a resolution. That was expected. The thesis was never built around a single year — it was built around a cycle, and cycles don't follow calendar years.
Q3 2026 is the window where the pressure that built throughout FY25 is most likely to force an outcome. The AI infrastructure build cycle will have run long enough to test monetization assumptions. Hyperscaler capex commitments made in 2024 and 2025 will be visible in earnings. Semicap order books will reflect whether demand is real or pull-forward. Energy and grid constraints will be operational, not theoretical.
The May 2026 earnings window — AMAT, LRCX, and the hyperscalers — is the first genuine inflection test. If guidance holds, the timeline extends. If guidance cuts, Phase 2 of the trigger system activates and the correction window opens earlier than the base case.
Either outcome is informative. Neither is a surprise. The thesis was built to survive both.
Editor's Note — Q3 2026 is not a prediction. It is a checkpoint. The data will tell us what it is when it arrives.
VII. What We Did With Capital in FY25
The most deliberate capital decision of FY25 was not a buy. It was a build.
FY25 produced a 25.9% return on a net asset value basis. That number did not arrive by accident. It arrived because the portfolio was positioned early in names the thesis identified — AI infrastructure, semiconductors, and hyperscaler-adjacent holdings — and held through volatility while the narrative caught up to the structure.
But the more important story from FY25 is what happened in Q4.
As the portfolio appreciated through October, November, and December, the decision was made not to redeploy gains into additional risk. Instead, proceeds from position management and incoming cash flows were systematically directed into SWVXX — a money market vehicle that earns yield while preserving full liquidity. The logic was simple: the thesis was working, which meant the correction window was getting closer, which meant liquidity was becoming more valuable, not less.
By December 31, the reserve was started and the portfolio was positioned — appreciated core holdings intact, liquidity building, and the next moves already in view. Additional trimming of appreciated positions is anticipated in Q1 2026, consistent with the same discipline that defined Q4: sell strength, build the reserve, and preserve optionality into the correction window the thesis has identified.
FY25 was the setup. The reserve is the strategy.
VIII. What Carries Into FY26
Three things carry from FY25 into FY26 unchanged.
The thesis carries. AI infrastructure spending will decelerate sooner than markets price. The correction will arrive before consensus expects it. The setup will look obvious in hindsight. None of that changed in FY25. If anything, the year added evidence rather than removing it.
The discipline carries. Sell strength. Build the reserve. Don't chase momentum. Don't average down out of ego. Don't force conviction where evidence is incomplete. Every decision made in FY25 that worked came from applying that framework consistently. Every mistake came from deviating from it.
The patience carries. The thesis is probability-weighted, not time-certain. The correction window is Q2–Q3 2026 into early 2027. That window has not opened yet. Waiting is not indecision. It is the position.
FY26 begins with the reserve started, the thesis intact, and the trigger system monitoring. The next twelve months will tell us whether FY25 was the setup for something significant, or the calm before more calm.
Either way, the framework holds.
Editor's Note — The geese who survive cycles are not the ones who predicted them perfectly. They are the ones who stayed positioned, disciplined, and patient.