The feds tried and failed to subsidize the chipmaker into success, so now they might just have to buy it.
The stock price of U.S. semiconductor manufacturer Intel jumped Thursday on reports that the Trump administration is negotiating for the U.S. government to take a stake in the company.
Where the constitutional or statutory authority for the president to do this would come from, who can say? But if the government does end up buying a chunk of Intel with taxpayer money — further blurring the distinction between the public and private sectors that underpins a free society — it will be the next stage in a process that began in 2022.
2022 saw the passing of the unfortunately bipartisan CHIPS and Science Act. Defended on both economic and national security grounds, the law envisaged providing private semiconductor companies with up to $52.7 billion to build up chip manufacturing in America. The largest recipient of CHIPS Act funding, by a large margin, was Intel. As of last year, the company had received $8.5 billion in federal grants, plus an additional $11 billion in government loans — supposedly to support a total $100 billion investment by the company in manufacturing facilities.
Cato Institute economist Scott Lincicome wrote in 2024 that, “for many in Washington, the CHIPS Act’s success or failure will hinge on Intel.” The company’s prospects were the law’s single biggest bet. “Should the company recover from its current tailspin, the subsidies will undoubtedly be given much of the credit; should it fail, it’ll be Solyndra on steroids.” That’s especially true because Intel was expected to lead the “bleeding edge” chip manufacturing of the sort of advanced logic chips currently made almost exclusively in Taiwan.
So, how have things at Intel gone over the past three years? Not well. The company’s profits have collapsed since 2022, and in 2024, it reported an annual loss of nearly $19 billion. Intel pinned much of that loss on a one-time, wholesale corporate restructuring — which it had to launch due to sliding competitiveness and mounting financial strain — but the firm is not expected to return to profitability until 2026. Share prices fell in tandem. Lincicome notes that, as of this time last year, Intel’s stock had fallen by 68 percent since its CEO announced the “turnaround plan” in 2021. For the first time in over four decades, the company was trading at below net asset value.
Intel has been dogged by persistent production delays. The enormous manufacturing facility that it’s building in Ohio, which Intel claims will be the world’s largest chipmaking facility, was originally supposed to begin operations in 2025. But that date has been pushed back multiple times, most recently by several years to either 2030 or 2031. Mere months after announcing that newly delayed opening date, Intel said that it would “further slow” construction of the chip factory. Similar production setbacks were a regular occurrence at Intel throughout the 2010s, and the company has a history of making bad bets on the future of the semiconductor industry that spans decades. By lavishing subsidies onto Intel in 2022, Lincicome argues, the federal government effectively rewarded the firm’s record of failure. That Intel’s situation has deteriorated even further should come as no surprise.
Allowing the chipmaker to sink deeper into irrelevance would mean admitting that the government wasted tens of billions of taxpayer dollars on a fantasy, and accepting that the United States may not be able to replicate Taiwan’s chip manufacturing capabilities after all.
That’s not an admission that President Trump appears ready to make. Last week, he publicly declared that the company’s new CEO, Lip-Bu Tan, must resign over his ties to China. That demand, in itself, was an extraordinary use of the president’s bully pulpit. After meeting with the CEO at the White House, however, Trump changed his tune, saying that “his success and rise is an amazing story.” Clearly their conversation went well, as Trump is now keen to invest U.S. tax dollars in Tan’s company. Reports say that the infusion would be to help shore up Intel’s beleaguered Ohio plant, because apparently the $8.5 billion in grants and $11 billion in loans wasn’t enough.
There are some parallels with the fate of that befell Fannie Mae and Freddie Mac, the enterprises that effectively back so much of the American mortgage market, during the 2008 financial crisis. Though privately owned at the time, they received enormous federal support through federal guarantees and a credit line with the U.S. Treasury. Investors believed that the government would never let them fail, and they were right. As the mortgage market imploded and Fannie and Freddie verged on default, the government made its implicit control of the companies explicit by putting both into federal conservatorship, where they remain today.
Intel may be about to meet the same fate: turning eventually from an implicit appendage of government to its explicit property. It is the same process that Friedrich Hayek labeled “the road to serfdom,” in his prescient book of the same name. Hayek described how central planning begets ever greater government control. As problems with the plan inevitably emerge, policymakers must take control of more and more diffuse economic decisions. They ultimately have to take effective ownership of the economy itself, thus reducing its participants to serfs.
Turns out, Intel may have signed up to be a serf of the Trump administration back when it accepted all those enticing federal subsidies. It just didn’t know it yet.