Warsh, a former Fed governor, now has a 47% probability, up from 39% on Sunday and just 11% on Dec. 3. Hassett, director of the National Economic Council, has fallen to 41%, down from 51% on Sunday and 81% on Dec. 3.

A report from CNBC saying Hassett’s candidacy was running into pushback from people close to President Donald Trump seemed to put Warsh on top. The resistance stems from concerns Hassett is too close to Trump.

That followed Trump’s comment late Friday, when he told The Wall Street Journal Warsh was at the top of his list, though he added “the two Kevins are great.”

According to the Journal, Trump met Warsh on Wednesday at the White House and pressed him on whether he could be trusted to back rate cuts. 

The report surprised Wall Street, which had overwhelming odds on Hassett as the favorite, lifting Warsh’s odds from the cellar.

But even prior to the Journal story, there have been rumblings in the finance world Hassett wasn’t their preferred choice to be Fed chair.

At a private conference for asset managers on Thursday, JPMorgan Chase CEO Jamie Dimon signaled support for Warsh and predicted Hassett was likelier to support Trump on more rate cuts, sources told the Financial Times.

And in a separate report earlier this month, the FT said bond investors shared their concerns about Hassett with the Treasury Department in November, saying they’re worried he would cut rates aggressively in order to please Trump.

Trump has said he will nominate a Fed chair in early 2026, with Jerome Powell’s term due to expire in May. 

For his part, Hassett appeared to put some distance between himself and Trump during an appearance on CBS’ Face the Nation on Sunday.

When asked if Trump’s voice would have equal weighting to the voting members on the rate-setting Federal Open Market Committee, Hassett replied, “no, he would have no weight.”

“His opinion matters if it’s good, if it’s based on data,” he explained. “And then if you go to the committee and you say, ‘well the president made this argument, and that’s a really sound argument, I think. What do you think?’ If they reject it, then they’ll vote in a different way.”

But that begs the question: What happens to all those older-generation chips?

The AI stock boom has lost a lot of momentum in recent weeks due, in part, to worries that so-called hyperscalers aren’t correctly accounting for the depreciation in the hoard of chips they’ve purchased to power chatbots.

Michael Burry—the investor of Big Short fame who famously predicted the 2008 housing collapse—sounded the alarm last month when he warned AI-era profits are built on “one of the most common frauds in the modern era,” namely stretching the depreciation schedule. He estimated Big Tech will understate depreciation by $176 billion between 2026 and 2028.

But according to a note last week from Alpine Macro, chip depreciation fears are overstated for three reasons.

First, analysts pointed out software advances that accompany next-generation chips can also level up older-generation processors. For example, software can improve the performance of Nvidia’s five-year-old A100 chip by two to three times compared to its initial version.

Second, Alpine said the need for older chips remains strong amid rising demand for inference, meaning when a chatbot responds to queries. In fact, inference demand will significantly outpace demand for AI training in the coming years.

“For inference, the latest hardware helps but is often not essential, so chip quantity can substitute for cutting-edge quality,” analysts wrote, adding Google is still running seven- to eight-year-old TPUs at full utilization.

Third, China continues to demonstrate “insatiable” demand for AI chips as its supply “lags the U.S. by several generations in quality and severalfold in quantity.” And even though Beijing has banned some U.S. chips, the black market will continue to serve China’s shortfalls.

Meanwhile, not all chips used in AI belong to hyperscalers. Even graphics processors contained in everyday gaming consoles could work.

A note last week from Yardeni Research pointed to “distributed AI,” which draws on unused chips in homes, crypto-mining servers, offices, universities, and data centers to act as global virtual networks.

While distributed AI can be slower than a cluster of chips housed in the same data center, its network architecture can be more resilient if a computer or a group of them fails, Yardeni added.

“Though we are unable to ascertain how many GPUs were being linked in this manner, Distributed AI is certainly an interesting area worth watching, particularly given that billions are being spent to build new, large data centers,” the note said.