So does Intel still make graphics cards?

There’s something strange going on at Intel. The company is looking to get leaner while at the same time building its chipmaking capabilities. The American chip giant’s fledgling venture into graphics cards has not been appreciated by the PC buying community, especially Battlemage GPUs are at the budget end Like the B850 amid the hype generated by its big Panther Lake announcement, there seemed to be one thing missing: a clear idea of ​​its future plans for GPUs, with or without Nvidia’s help.

(Full disclosure: Intel invited me to its chip factory in Phoenix, Arizona. Intel paid for travel and lodging, but Gizmodo did not guarantee any coverage as a condition of accepting the trip.)

With the introduction of Tiger Lake and Updates to the XeSS upgrade program The new Xe3 graphics microarchitecture comes from Intel, which falls under the Arc B-Series (Battlemage) umbrella. Intel said we should expect better performance at lower wattage than the previous Arrow Lake H lineup and much better frame rates in games with the 12Xe core chip variants. But what about everything else? Intel has detailed that its “next Arc family” will be called Xe3P. No, not XE4. Will it be a discrete GPU, also known as “celestial” or C-line graphics cards?

Tom Petersen, Intel’s head of architecture, graphics and software, told reporters in a roundtable question-and-answer session that Panther Lake will only do what the name suggests. “Xe3P is a significant architectural advance compared to where we are now,” he said. Whether that means it’s a full suite of products doesn’t matter. However, it can still be called “Celestial”, more for continuity than anything else.

“Our naming is not great,” Petersen said. “If we knew what we know now, we would call these things differently.”

Even Intel doesn’t know what to do with Nvidia

Intel has other things on its mind. The company wants you to know that its Fab 52 plant in Chandler, Arizona, generates… The company’s new 18A process Running. So much so that they strapped me and a group of other journalists and analysts into a white bunny suit to inspect the place. Just to step into this temple of silicone, your body is wrapped from head to toe in waterproof layers of Gore-Tex, your eyes and feet are covered, and you begin to blend in with everyone else walking around those floors. What can I tell you? Not much. What is the size of the fancy floor where they make the laminate in square feet? “A lot of bullshit,” or at least that’s what Intel spokesman Thomas Hannaford was allowed to say. I couldn’t take pictures. I couldn’t tell you how big lithography machines are. This would give some competitors an advantage, and perhaps give a signal to the world about how many chips they plan to ship, or so Intel claimed.

As I stared at the flying shuttle robots cruising across the ceiling, they looked like two-slit Recognizer vehicles emerging from… You see film– As they carried the chips back and forth across the bars in the ceiling, I could tell I was the source of Intel’s mission statement. The Fab 52 has been in production since 2021. Since then, Pat Gelsinger, the person who started Intel on this mission to manufacture in the United States, He was pushed out The CEO and the company went into a year-long spiral that reached its peak President Donald Trump Push the federal government to Obtaining a 10% stake in the company. Then Nvidia came along with its products Moneybags the size of Scrooge McDuck ($5 billion to be exact) to pump more fuel into the chip maker’s ovens. Among all the capital that changed hands were the CEOs of Nvidia and Intel The new combo chip is described It would combine Team Blue’s CPU with Team Green’s GPU.

Fabs are just as important as the chips they make them with. And while I could sit here and sing about the company’s Clearwater Forest data center chips, the PC-buying public only cares about what ends up on their desktop or laptop. Companies don’t like to talk about their future, but from what Intel executives said last week, the company itself is still trying to figure out what the partnership means.

“It’s brand new,” Petersen said, referring to the still-unknown chips that Nvidia could make. “We don’t know all the answers to that. You’ll know more about that relatively soon. We’re still in figuring it all out mode.”