Elena VossThirty-six billion dollars. That's the value of the federal government's 10% stake in Intel as of Friday's close, after the company's first-quarter print drove the stock up 23.6% in a single session — Intel's best one-day performance since October 1987. The cost basis was $8.9 billion. The position is up 304% in eight months.
In August 2025, the Trump administration took a 10% equity position in Intel for $8.9 billion. Intel filed an SEC notice warning shareholders of "adverse reactions" — international sales risk, customer concerns, governance distortion. Lawyers warned of legal challenges. Stratechery published a piece titled "The US taking a 10% equity stake in Intel is a terrible idea." Reuters surveyed experts who said the intervention was unlikely to save Intel. The case for it being a terrible idea was the consensus case.
Eight months later, the consensus is the trade.
Compared to What
304% in eight months is a large number. The question is whether anything in the comparison set is in the same range. It is not.
The most successful US government bailout of the post-2008 era, by realized return, was AIG: roughly $182 billion injected, ~$205 billion recovered over multiple years. Net positive, low single-digit return on capital. The auto bailout (Treasury's investments in GM and Chrysler) returned roughly break-even after extended holding periods. The CHIPS Act subsidy regime under Biden — Intel's original $8.5 billion grant in March 2024 — was structured explicitly to deny the government any equity upside; a November 2024 filing stated that the subsidy required the government to take no ownership.
The private comparison is more flattering and still inadequate. Sequoia's best vintage funds compounded at lifetime IRRs in the 60% range. Berkshire Hathaway has averaged roughly 20% annualized over six decades. Norwegian sovereign wealth — the standard for state asset management — runs 5% to 7% annualized. The Trump-Intel position, annualized, is north of 1,400%. It outperforms every benchmark a market participant would reach for.
Two caveats apply. The position is mark-to-market, not realized. The federal government cannot easily exit a 10% public-company stake without signaling effects that compress the price. But the constraint is liquidity, not magnitude. The unrealized gain is $27.1 billion. That is also a number with no comparison.
What the Trade Caught
The mechanism that produced the rally is straightforward and was not in the consensus forecast in August 2025. AI inference demand expanded beyond Nvidia data-center GPUs and into CPUs at a rate that revalued Intel's core business. Intel posted Q1 2026 revenue of $13.58 billion, up 7% year-over-year, beating estimates by more than $1 billion. CEO Lip-Bu Tan, who took over in March 2025, characterized the company as "fundamentally different." The Financial Times observed that Intel's market cap had surpassed its dotcom-bubble high.
None of this is what the August 2025 deal was supposed to fix. The Trump intervention was framed as foundry insurance. The administration priced the insurance in equity rather than subsidy because it wanted upside. What drove the stock from $89 billion to ~$360 billion in market cap was AI inference revaluing CPUs. The administration bought insurance and got asymmetric upside on a tail event nobody had priced.
This is the structure of every trade that returns 304% in eight months. The position has to be taken before the catalyst is visible. By August 2025, the catalyst was not visible in any sell-side model. By April 2026, the catalyst is the only thing visible.
The Criticism Survives
Stratechery's argument that the equity stake was a terrible idea was structural, not predictive. The argument was that government equity in publicly traded firms creates incentive distortions: customer concerns about export controls, foreign-government skittishness, governance ambiguity, and a long-term moral hazard for any other firm seeking similar treatment. Those objections do not depend on Intel's stock price. They are about the second-order effects of policy precedent. The trade being up does not address them.
The Intel SEC filing made the same argument in a different register: the company itself listed "adverse reactions" from international customers and trading partners as risks of the deal, in a document the company was legally required to make accurate. International sales accounted for 76% of Intel's prior-year revenue. The disclosure was a hedge against a real risk, not a hypothetical one.
Both critiques survive the data. They were never refuted by the rally. They simply now have to coexist with $36 billion.
What the Trade Tells You
The piece of information the rally does refute is narrower than the broad critique: it refutes the prediction that the intervention was unlikely to save Intel. Intel was saved, by any reasonable benchmark — operationally by Tan's restructuring, financially by the federal balance-sheet support, and reflexively by the equity stake itself, which gave institutional investors a floor argument they had not previously had. The "save" worked. The dispute is about whether the policy that produced the save is a template or an exception.
The structural reading of an outlier return is that it cannot be repeated by replicating the procedure. The federal government cannot manufacture similar outcomes in defense, biotech, or energy by taking 10% equity stakes — not because the procedure is wrong but because the catalyst (AI inference revaluing CPUs) is not present in those sectors. The trade is a result, not a method. Every successful trade looks like a method in retrospect. Most are not.
Thirty-Six Billion
The opening number is the position's mark on April 24, 2026. It is not a forecast. The stock can fall. The position can be diluted. Future quarters can disappoint, as Intel's have for most of the last decade.
The number is, however, what it currently is. Eight months ago, the consensus among experts, lawyers, the company itself, and the most cited industry analyst was that the federal government had made a strategic mistake. Eight months later, the strategic mistake is up 304% and the federal balance sheet has $27 billion in unrealized gains it did not have. The criticism survives. The trade survives the criticism. By any reasonable benchmark, those two things are now both true.