Intel (NASDAQ: INTC) Soars on a Rare Sell-to-Buy Upgrade: Buy the AI Comeback or Too Late?

Intel (NASDAQ: INTC) jumped about 10% after Bank of America did something unusual: a “double upgrade”, lifting the stock straight from Underperform (Sell) to Buy and raising its target to US$135 from US$96. The twist is that Intel is already one of 2026’s biggest winners, up around 190% this year and now near US$118. So is this the time to buy the comeback, or has the easy money gone?

Why Bank of America Flipped From Sell to Buy

BofA’s Vivek Arya is a top-ranked analyst, so the U-turn grabbed attention. His bull case rests on three ideas.

First, AI needs more than GPUs. As AI systems do more step-by-step reasoning, they lean harder on central processors (CPUs), Intel’s home turf, where Arya sees server CPU sales reaching about US$40 billion by 2030.

Second, the foundry business (making chips for other companies) is drawing interest, with possible work for Apple and MediaTek, a second income stream beyond selling its own chips.

Third, almost nobody owns it: only about 16% of S&P 500 funds hold Intel, despite its roughly US$590 billion market value, making it one of the least-owned chip names among big funds. Fresh buying alone could lift the price, and Arya sees earning power topping US$6 a share by 2030.

The Catch: The Street Is Far More Cautious

Most analysts are far less bullish: the average 12-month target sits near US$92, below where the stock trades now, and the consensus rating is only “Hold”. BofA is the outlier, not the crowd.

The business still has real problems. Intel’s foundry arm lost about US$2.4 billion in a single quarter, and its key 18A process has not yet proven it can match Taiwan’s TSMC or Samsung on quality, so big foundry deals stay uncertain. After a 190% run and still loss-making on a trailing basis, the shares are priced for a lot to go right.

The Investor’s Takeaway for INTC

So, buy, or is it too late? In our view, the turnaround is becoming real: the AI-CPU angle, a reported Google order for more than 3 million AI chips in 2028, and growing foundry interest are genuine progress. But the price has already raced ahead, and the average analyst still sees downside, so the easy gains look gone.

This suits patient, risk-tolerant investors who back the multi-year foundry story, not momentum chasers. Others may prefer to wait for proof on 18A yields or a calmer entry.

The bottom line: watch the 18A yield news and July’s results. That is what turns the bull case from hope to numbers.