Have things fundamentally changed for Intel, or is the stock simply being fueled by vibes?

That’s a hot topic of debate on Wall Street, as some analysts argue that the boom in server central processing units has dramatically altered the financial picture for Intel. But others say a rosy narrative is masking persistent concerns about the company’s fundamentals.

Intel’s stock closed up 23.6% on Friday — its largest one-day gain since Oct. 29, 1987, when it rose 26.4% according to Dow Jones Market Data. Shares are up 124% so far this year, making Intel the fourth-best performer in the S&P 500 this year.

At least one analyst just changed his mind about Intel’s prospects. “It has been easy to not like” Intel, Evercore ISI’s Mark Lipacis wrote in a note to clients. The company had been facing a laundry list of potentially existential issues, like repeated technological missteps and an erosion of market share.

But three things have since changed for Intel, according to Lipacis, who upgraded the stock to outperform from in-line following Thursday afternoon’s earnings report. For one, “the fastest-growing AI workloads need a lot more CPUs,” he said — explaining that whereas AI workloads once required one CPU for every eight graphics processing units, that ratio could completely flip to eight CPUs for every GPU.

Additionally, CEO Lip-Bu Tan “fixed the balance sheet” as Intel’s technology is now more competitive, according to Lipacis. And the company’s manufacturing business is suddenly more attractive as the U.S. government backs domestic chip-making ventures.

Seaport analyst Jay Goldberg said he thinks the company is on its “firmest footing in years.” Intel’s recent announcement of its Terafab semiconductor plant — jointly developed with Tesla, xAI, SpaceX and Super Micro Computer — has been validating for investors, he said, and he expects further announcements later this year.

Goldberg added that while the company still faces constraints on capacity for CPU production, he expects them to ease throughout this year.

Jefferies analyst Blayne Curtis argued that the AI opportunity has made Intel’s other issues fade into the background. “We continue to believe it remains early days for AI’s impact on the CPU market with plenty of room to run,” he wrote in a note to clients. Explosive growth in the usage of accelerated processors “is driving real server-CPU demand,” he added.

That said, Curtis kept a hold rating on Intel’s stock as he waits to see more traction from large customers and for the company’s foundry business.

Bernstein’s Stacy Rasgon stayed on the sidelines as well. While “narrative/headlines may fuel the vibe for now,” he wrote, elements within Intel’s report might resonate with bearish investors.

“Namely, while 18A yields are seemingly running better than expected, they apparently remain underwhelming,” Rasgon said, referring to the portion of usable chips that Intel gets from its manufacturing technology. And the company faces various cost pressures — such that even when Rasgon increases his revenue expectations “fairly decently,” his profit projections barely budge.

Morgan Stanley analyst Joseph Moore was also cautious. He said that the “vast majority” of the company’s CPUs are not in data centers, so AI growth will only apply to a portion of its CPU volumes. He also expressed concerns over the nature of the Terafab project, saying that the structure of the collaboration isn’t yet clear.