Four companies, four announcements, seven months, and the same sentence in every press release: taxpayers should share in the upside, not just cover the downside. Intel -- $8.9 billion for a 9.9% stake, announced August 22, 2025, structured by converting CHIPS Act grants the company was already awarded, rather than new appropriated money.^[1][2] MP Materials -- $400 million in new Department of Defense convertible preferred equity for a 15% as-converted stake in the only operating rare earth mine in the US, announced July 10, 2025, alongside a separate $150 million processing-upgrade loan.^[3][4] GlobalFoundries -- a $300 million CHIPS R&D grant for silicon photonics work, announced July 29, 2026, with roughly a 1% equity position attached to it.^[5][6] Westinghouse -- part of an $80 billion nuclear-reactor-build partnership announced in October 2025, structured as 20% upside participation, convertible to an equity warrant if the company goes public.^[7][8] Four different mechanics -- a grant re-papered as stock, new money split across equity and debt, a grant with equity as a condition, and profit-sharing convertible on a future IPO -- landing on the same outcome: the federal government becomes a shareholder in a specific private company, not just a funder of a program.

That reads like something new. It isn't. Run the same search against 2008 and you find the government already did this, at a scale that dwarfs the current round. Treasury's Capital Purchase Program, the core of TARP, made $250 billion available starting October 14, 2008, and had actually deployed $172.5 billion into 214 banks by the end of that year alone -- purchasing preferred stock directly, exactly the instrument MP Materials and Westinghouse are using today.^[9][10] And it didn't stop at the preferred stock: Treasury also took warrants to buy common stock worth 15% of each preferred investment, on top of the equity position itself.^[9] Banks that recovered could buy the warrants back; Treasury sold the rest to third-party investors. The mechanism reads like a direct ancestor of Intel's structure seventeen years later -- equity plus warrants, sized to the company's own risk profile, designed so the government's upside scales with the company's recovery instead of stopping at repayment.

Go back further and the warrant-only version shows up too, and it worked. The Chrysler Corporation Loan Guarantee Act of 1979 wasn't an equity stake -- it was $1.5 billion in loan guarantees, of which Chrysler drew $1.2 billion and fully repaid by 1982.^[11] But as compensation for the risk, the government took warrants to buy 14.4 million Chrysler shares at a $13 strike price.^[11] Almost nobody expected those warrants to ever be worth exercising -- for that, Chrysler's stock had to actually recover above $13. It did: at a September 12, 1983 auction, Chrysler itself outbid three Wall Street syndicates and paid $21.602 per warrant -- $311.07 million total for all 14.4 million -- to keep them off the open market.^[12] The government's cost basis on the warrants was $0 (they were granted as compensation, never purchased), so the $311.07 million was pure profit, on top of a loan program that had already been fully repaid the year before and cost taxpayers nothing on its own. Structurally, this is exactly Westinghouse's deal today: no immediate equity, upside participation contingent on the company's own future success.

Then there's the version without any equity at all, and it's the one people actually remember. The Department of Energy's 2009 loan guarantee to Solyndra -- $535 million -- was pure debt.^[13] DOE guaranteed the loan; Solyndra itself had to fund the remaining 27% of project costs through private equity contributions of its own. The government took no ownership position, no warrants, no upside participation of any kind. When Solyndra went bankrupt in September 2011 after burning through $528 million of the guarantee, taxpayers absorbed the loss with nothing to show for it -- no shares to sell, no warrants to exercise, no recovery mechanism at all.^[14] Solyndra became the political shorthand for "government picks losers," and the shorthand isn't wrong about the structure: it's the one case in this set where the government took on real downside risk while deliberately declining any share of the upside.

Read against those three precedents, the current wave reads less like a new tool and more like TARP's own structure being redeployed at a smaller, more targeted scale. Every 2025-2026 deal pairs government capital with some form of upside participation -- equity, warrants, or both -- the exact feature Solyndra's critics said was missing in 2009. The sectors are narrower and more specific this time: not "the financial system" or "one automaker," but semiconductors, rare earth processing, and nuclear construction, each picked for the same stated reason -- reducing dependence on China-controlled supply chains. Nuclear restart financing in the same window runs the Solyndra pattern instead: the DOE's Loan Programs Office issued a loan guarantee (not equity) to restart the Palisades plant, and a separate $1 billion loan guarantee for Three Mile Island Unit 1 -- both real federal capital, both structured as pure debt, in the same nuclear sector where Westinghouse's own deal does carry equity participation.^[22][23] Even within one sector, the government is choosing different instruments company by company.

One more sector is conspicuously absent: oil and gas has no equivalent equity-stake deal in this set. The stated rationale in every case above is reducing dependence on a foreign supplier -- and the US has no comparable foreign-dependency problem in oil and gas; it's been a net exporter for years. The pattern is scoped to where the government has decided a specific supply chain is exposed, not a general preference for owning pieces of American industry.

A real complication sits inside the newest and smallest of these deals. Vulcan Elements, a rare-earth magnet manufacturer, isn't one of the four equity cases above -- its federal package is a $620 million conditional Pentagon loan with warrants plus $50 million from the CHIPS Act, debt rather than direct equity.^[18] What makes it different is the sequence: 1789 Capital, the venture fund where Donald Trump Jr. is a partner, took an undisclosed stake in Vulcan's $65 million Series A in August 2025 -- months before the federal package arrived.^[19] ProPublica and the House Natural Resources Committee (Rep. Jared Huffman) have reported that White House adviser Peter Navarro, a friend of Trump Jr.'s, personally requested the loan for the company; House Democrats forced a vote to subpoena Trump Jr. over the deal, and the Cato Institute separately reported direct White House intervention on Vulcan's behalf.^[19][20][21] None of that is settled -- it's contested, actively-reported journalism, not adjudicated fact -- but it's the one case in this set where the sequence itself (private stake first, federal capital second, with an alleged personal ask in between) is the actual story, independent of whether the underlying rare-earth-magnet investment thesis is sound.

Whether private capital treats a government stake as a signal to follow doesn't have one answer either. At MP Materials, Apple committed $500 million just five days after the DoD's own announcement -- government first, private capital fast behind it.^[15] At Intel, the order runs backward: SoftBank signed its $2 billion investment on August 18, four days before the government's own Intel stake was announced on August 22, and Nvidia's $5 billion arrived a month later, after the government position was already public -- one company, two private investors on opposite sides of the government's own timeline.^[16][17] Vulcan Elements runs backward too, on a longer clock: private money in August, federal money months later. There's no consistent lead-follow relationship across even this small set of four to five deals -- which means "government stake as a market signal" is, so far, a real but inconsistent effect, not a rule.

Why does this matter? The 2008 and 1979 precedents settle one question cleanly: taking equity or warrants alongside a rescue or an industrial-policy bet isn't new, and when the underlying company recovers, it can turn a genuine profit for taxpayers instead of a pure loss. What isn't settled by precedent is the current wave's real distinguishing feature -- these aren't rescues of failing firms, the way TARP and Chrysler were. They're preemptive bets on healthy, strategically-important companies in sectors the government has decided are exposed to a single foreign supplier. That's a different kind of decision than "this firm would otherwise fail" -- it's "this firm, specifically, is where we've chosen to own a piece of the outcome" -- and neither TARP nor Chrysler nor Solyndra is a clean precedent for evaluating that choice, because none of them were made for that reason.