# The U.S. Government Bought Intel at $20.47. It Just Closed at $113.
> We tracked Washington's Intel trade: $8.9B in at $20.47 in August, $49B by May 6 — and every congressional INTC disclosure before and after the deal.
- Author: Barebone Research, Barebone AI
- Published: 2026-05-06
- Canonical: https://barebone.ai/resources/us-government-intel-stake-best-trade-in-washington
- Publisher: Barebone AI (https://barebone.ai)
---
## The Resignation Demand That Became a Position
On August 7, 2025, the President of the United States posted that Intel's brand-new CEO should resign immediately, citing Lip-Bu Tan's past investment ties to Chinese chip firms. The stock sold off. Days later, Tan was in the White House. Two weeks after that, the United States government owned 9.9% of Intel.
Trump's own description of the meeting: "I said, I think it would be good having the United States as your partner. He agreed, and they've agreed to do it."
We used Barebone to pull the deal's SEC filings, nine months of price history, and every congressional Intel disclosure filed since January 2025. The arithmetic, marked to today's close: **the government paid $20.47 a share in August. Intel closed at $113.01.** An **$8.9 billion** position is now worth roughly **$49 billion** — an unrealized gain of about **$40 billion**, or **+452%**, in eight and a half months.
That's the headline. The findings underneath it are better: members of Congress were accumulating Intel for months before the deal existed — one of them bought twenty-four days before it was announced — and the same disclosure records show them quietly taking profits ever since. We'll show you the receipts.
## Anatomy of the Trade
The deal announced on August 22, 2025 was, mechanically, the strangest stock purchase of the decade. Here are the terms from Intel's 8-K:
| Term | Detail |
|------|--------|
| Shares | **433.3 million** at **$20.47** |
| Cost | **$8.9 billion** |
| Stake | **9.9%** — one of Intel's largest shareholders overnight |
| Cash source | $5.7B in unpaid CHIPS Act grants + $3.2B from the Secure Enclave program |
| The kicker | A 5-year warrant for **240.5 million more shares at $20.00**, exercisable if Intel gives up majority control of its foundry business |
| Governance | Passive — no board seat, no governance rights |
Note what the "cash source" row means: the government didn't write a new check. It converted grants Congress had already awarded Intel — money the company was getting anyway, with no equity attached — into common stock. That's how Trump could post, the day of the deal, "The United States paid nothing for these Shares, and the Shares are now valued at approximately $11 Billion Dollars." And, the following Monday, in all caps: "I PAID ZERO FOR INTEL, IT IS WORTH APPROXIMATELY 11 BILLION DOLLARS."
The $11 billion was accurate — 433.3 million shares were worth about $10.5 billion at that week's prices. "Paid zero" is marketing: the $8.9 billion was real taxpayer money, just money that had previously been a gift. The people who actually paid were Intel's existing shareholders, who absorbed the dilution of 433 million new shares priced below market.
And the warrant turned out to be the quiet masterstroke. At today's close it sits **$93.01 per share in the money** — roughly **$22 billion** of additional paper value, lying dormant unless Intel ever spins off its fabs.
Then the stock did something Intel stock had not done since the summer of 2000.
## Nine Months Up the Wall
The government wasn't alone that August. Three buyers hit the same tape within a month: SoftBank signed for **$2 billion at $23.00** on August 18 (Intel closed at $23.66 that day), the government's $8.9 billion landed on August 22, and on September 18 NVIDIA announced **$5 billion at $23.28** alongside a chip co-development deal. Whatever you think of the politics, the signal was unambiguous — three of the most informed actors in the semiconductor world bought the same beaten-down stock within a month of each other, all between $20 and $24.
From there the tape reads like a staircase: **$33.55** by end of September, **$39.99** by end of October, a wobble back to **$36.90** in December as the hype cooled, then a grind through the mid-$40s as Intel's 18A node — its make-or-break advanced manufacturing process — started showing up in real products.
The detonation came on April 23. Intel's Q1 2026 report: revenue of **$13.6 billion** against a consensus near $12.4 billion, EPS of **$0.29** against a consensus of one cent, data center and AI revenue up **+22%**, and confirmation that 18A had entered high-volume manufacturing with yields tracking roughly two quarters ahead of plan. It was the company's sixth consecutive beat, and the first one nobody could argue with.
The next day the stock rose **+23.6%** to $82.54 — its best session since 1987 — and, in that single move, closed above its August 2000 dot-com record of **$74.88** for the first time. A twenty-six-year round trip, ended in one trading day. The earnings pop alone added about **$6.8 billion** to the government's position before lunch.
By May 1 the stock closed within 38 cents of $100. Today it closed at **$113.01** — up **+206%** in 2026 alone, on top of +84% in 2025.
## The Referee Got the Best Price
Mark all three August–September entries to today's close and an awkward pattern appears:
| Buyer | Date | Entry | Invested | Value at $113.01 | Return |
|-------|------|-------|----------|------------------|--------|
| **U.S. Government** | Aug 22, 2025 | **$20.47** | $8.9B | **~$49.0B** | **+452%** |
| SoftBank | Aug 18, 2025 | $23.00 | $2.0B | ~$9.8B | +391% |
| NVIDIA | Sep 18, 2025 | $23.28 | $5.0B | ~$24.3B | +385% |
SoftBank and NVIDIA negotiated their prices with Intel. The government's counterparty was a company whose CEO it had publicly tried to fire two weeks earlier, whose grants it controlled, and whose export licenses run through Washington. It is not a scandal that the government got the lowest entry — but it is worth saying plainly: the entity that wrote the rules of the negotiation also booked the best price in it.
(NVIDIA's stake is its own story — a hedge on where chips get made, not a trade — but for this scoreboard, the number is the number: $5 billion is worth about $24 billion if it hasn't sold a share.)
## Congress Was Already Long
Here's the part the disclosure records tell better than any press release.
Long before August, members of Congress were buying. STOCK Act filings — the law requiring lawmakers to disclose trades within 45 days — show Rep. Marjorie Taylor Greene bought Intel three separate times in the first half of 2025 (January 8, February 12, May 5; $1,001–$15,000 each), Rep. Robert Bresnahan made multiple purchases in February and March, Rep. Jefferson Shreve bought $15,000–$50,000 worth on March 13, and Rep. Dwight Evans bought in April. All of this with the stock in the teens and low twenties, not long after its September 2024 bottom below $18.
Then the timing gets uncomfortable. On **July 29, 2025**, Rep. Tim Moore of North Carolina disclosed a **$15,001–$50,000** Intel purchase — twenty-four days before the government's stake was announced, and just before reports surfaced in mid-August that the administration was weighing one. By late August, Benzinga's tracking had the stock up more than 25% from Moore's entry. There is no public evidence Moore knew anything. The filing just sits there, dated.
After the announcement, the pattern inverts. From August 22 through today, the filings we pulled show **four congressional purchases and seven sales**:
| Date | Member | Action | Size |
|------|--------|--------|------|
| Sep 11, 2025 | Rep. Michael McCaul | Sell | $15K–$50K |
| Nov 6, 2025 | Rep. Ro Khanna | Sell | $1K–$15K |
| Nov 10, 2025 | Rep. Michael McCaul | Sell ×2 | $15K–$50K + $100K–$250K |
| Dec 15, 2025 | Rep. Gilbert Cisneros | Buy | $1K–$15K |
| Jan 7, 2026 | Sen. Gary Peters | Sell | $15K–$50K |
| Jan 13, 2026 | Rep. Julia Letlow | Buy | $1K–$15K |
| Feb 13, 2026 | Rep. Gilbert Cisneros | Buy | $1K–$15K |
| Mar 9, 2026 | Rep. Ro Khanna | Sell ×2 | $1K–$15K each |
| Apr 7, 2026 | Rep. Ro Khanna | Buy | $1K–$15K |
Accumulate before the catalyst, distribute after it. That's not an indictment — many of these accounts are managed by advisors or held by spouses, the dollar amounts are small, and disclosure ranges are wide. Congressional trades are a noisy signal. But they are a signal, and on Intel the noise pointed one direction before August 22 and the other direction after.
## The Day-Trader Theory Has a Survivorship Problem
The internet's favorite read on all this is that the President is the best day trader alive: in at $20, victory-lap post at $25, and now this, posted at the end of April as Intel cleared $90:
> "I'm very proud of that Company in that I am responsible for making the United States of America over 30 Billion Dollars in the last 90 days on that stock alone."
We checked the math. At the April 29 close of $94.75, the government's total unrealized gain was about **$32 billion** — so "over 30 Billion Dollars" is real. But it's the gain since *August*, not "the last 90 days"; the trailing-90-day gain was closer to **$21 billion**. The right number, attached to the wrong window. When a true number gets exaggerated anyway, you're reading marketing, not accounting.
The bigger problems with the day-trader theory:
**Posts cut both ways.** The same account announced the "Liberation Day" tariffs on April 2, 2025. The S&P 500 fell **-4.84%** and then **-5.97%** on consecutive days — about **$6.6 trillion** erased, the largest two-day dollar loss on record, with the VIX at 45. Anyone "tracking Trump" as a buy signal in 2025 tracked themselves straight into that hole before the recovery came.
**The signal already failed a live test.** On April 10 of this year, Trump praised Palantir's "great war fighting capabilities" in a post that included the ticker — a first for a sitting president. The stock popped about 3% in minutes, faded the same afternoon, and closed lower, still down roughly 24% on the year. A presidential endorsement couldn't beat a software selloff for even one session. Intel rallied because 18A yielded and earnings beat — the posts amplified the move; they didn't cause it.
**And the government can't take the profit.** Selling 433 million shares would crater the stock and the industrial policy the stake exists to serve. The $40 billion is real on paper and nearly unrealizable in practice. This isn't a trade. It's policy with a ticker symbol — which means the usual discipline of a trade (an exit plan, a stop, a thesis review) doesn't apply, and the government's risk profile doesn't translate to a personal one.
One more humbling number: the investors who bought Intel's *last* all-time high, in August 2000, waited twenty-six years to get back to even. A stock that quintuples in nine months has pulled forward a lot of perfect execution.
## What This Means
Strip the politics away and three durable observations remain.
**Equity-for-support is now a repeatable instrument.** Trump said it himself in that April post: "There are others that, likewise, I have been very successful with by taking pieces of the Equity for support." The Intel structure — grants converted to stock, plus a warrant — is a template, and templates get reused. Each reuse creates the same paper trail this one did: an 8-K with the terms, a price, a warrant strike.
**The paper trail was readable in real time.** Every step of this story was a public filing before it was a headline: congressional purchases disclosed under the STOCK Act, the stake terms in an 8-K within days, SoftBank's and NVIDIA's entries in press releases and filings. None of it required access — it required reading documents most people never open.
**Separate the policy floor from the execution bet.** At $20, Intel was priced for irrelevance, and Washington's entry put a floor under the extinction scenario. At $113, the stake is fully priced and the stock now trades on whether 18A yields, foundry customers, and data center share keep compounding. The floor hasn't moved; the building on top of it is forty stories taller.
The government's Intel position will be studied for decades, as industrial policy or as precedent, depending on who's writing. For investors, the lesson is narrower and more useful: the most lucrative information in this entire story — who was buying, at what price, with what terms — was filed publicly, on a schedule, the whole way up. The question is whether you're reading the filings before the move, or the victory-lap posts after it.
---
*Data: Barebone | Sources: Intel 8-K filings (SEC EDGAR), Intel Q1 2026 results, Truth Social posts, congressional STOCK Act disclosures via Benzinga and Quiver Quantitative | Prices as of May 6, 2026 close*