Lip-Bu Tan Flips On Intel's Bleeding-Edge Node in With a Flawless Spec Sheet and Zero Buyers

Intel to Apple be like: “You up?” 

Intel is quite literally on the track of building the most advanced chip line on the planet this week and has, as of this writing, nobody to sell it to. The once mighty Iron Throne holder flipped on production of 18A-P, its bleeding-edge new node, at a chip conference in Honolulu.

Technically the node is in "risk production," an early stage where the chips exist but the customers are still hypothetical. Intel, in other words, is playing just the tip. After years of missing nodes, blowing deadlines, and getting its d*ck kicked in by Taiwan on yield, the 18A family was supposed to be the turnaround that finally made Intel a real foundry, a place other companies pay to print their own chips.

How so? Because 18A-P runs 9% faster or sips 18% less power than plain 18A, shrugs off 20% more heat, and drops straight into the 18A buildouts Intel's been cranking in Arizona since December. (On paper, the chip finally works.) And it’s genuinely competitive. The trouble is the one number Intel has left off the slide. "Yield rate is the number one criteria here," said Counterpoint's Neil Shah, who reckons clearing 90% yield in month one is what finally pulls a few customers off the fence.

The money, at least, has already shown up in some sense. Intel stock has ripped more than 200% this year on top of an 84% run in 2025, juiced by a genuinely wild guest list: Uncle Sam backed up the Brinks truck for a 10% stake in August, and Nvidia chipped in a bigly $5 billion in September.

CEO Lip-Bu Tan says he expects "multiple foundry customers" to commit in the back half of 2026, and the stock already popped 14% in a single month on nothing but a rumor that Apple might be one of them.

That said, there is a catch from the Honolulu roadshow. Intel mostly prints on x86, the architecture it's run since the dinosaurs, while Apple, Google, Amazon, and basically every outfit designing its own silicon builds on Arm, an entirely different instruction set. "Building Arm chips is something they have not done," Shah noted, while market-leading TSMC "has mastered that." Even the bulls are hedging as  analyst Ben Bajarin figures Apple waits for 18A-P specifically before it prints anything, if it ever does.

So you’re telling me there’s a chance? 

Possibly. But it only gets more crowded. TSMC is building a $165 billion mega-campus 50 miles up the road from Intel's Arizona fab, and Intel still has zero “yet to secure a major" outside customers for the node. Its likelier first win is the unsexy part as advanced packaging, the step where chip dies get stitched into bigger systems. (Nobody's ever cut a hype reel about interconnect bridges.) Intel's EMIB tech squares off against TSMC's CoWoS, and TSMC's packaging lines are so jammed that Shah calls Intel's opening there "very low hanging" fruit.

So what do we do with this information? Well, because the stock has already priced in the happy ending… skip the Honolulu confetti and the next Apple rumor, and watch the back half of 2026 for a customer with a real name on a real contract, plus whether that first-month yield clears 90%.

Until then, Intel built the field all right. The only ones who've wandered out of the corn so far are a government holding 10% and a chipmaker holding a $5 billion receipt, and neither one came to buy a chip. Place your bets accordingly, friends. Until next time… 

 

At the time of publishing, Stocks.News holds positions in Apple, Intel, Amazon, and Google as mentioned in the article.