Intel gets trapped in Elon’s reality distortion field as it joins in megafab delusions
Space is just the next stop on the AI hype train, right after AGI
In the realm of his other unrealistic plans and potentially broken promises, Elon Musk's Terafab stands out as one of the biggest pipedreams, promising to boost semiconductor production by 50x for the benefit of orbital datacenters. But hey, this idea must have legs, because now Intel has announced it is joining the aspiring Bond villain's initiative.
"If you add up all the fabs on earth combined, they're only about 2% of what we need for the… Terafab project," Musk said.
In case you were wondering, Musk has never built a wafer fab before. Neither has any of his companies. So when the world's richest man revealed a pie-in-the-sky plan to build a factory capable of churning out enough chips to make his dream a reality, folks were understandably skeptical.
Fabs are among the most complex and expensive facilities in the world. It can cost $30 billion and take as long as five years to bring a modestly-sized facility online, and that's if you already know what you're doing.
There is nothing modest about what Musk has proposed.
"We will have all of the equipment necessary to make a chip of any kind [including] logical memory," Musk said of the first Terafab facility, which will purportedly be located in Austin, Texas, which has become the fabulist hecto-billionaire's geographic center of gravity over the last few years. "So in a single building, we can create a lithography mask, make the chip, test the chip, make another mask, and have an incredibly fast recursive loop for improving the chip design."
Intel does bring some much-needed expertise to the endeavor, though it's yet to disclose to what extent it's involved, stating only that it was "proud to join the Terafab project with SpaceX, xAI, and Tesla to help refactor silicon fab technology."
"Our ability to design, fabricate, and package ultra-high-performance chips at scale will help accelerate Terafab's aim to produce 1 TW/year of compute to power future advances in AI and robotics," the company wrote on Musk-owned social network X.
Intel could simply be acting as an advisor, or it could be positioning itself to operate the megafab on Musk's behalf. Having said that, if Intel were involved to that degree, we'd expect to have seen an SEC filing to that effect, as it would be financially material. Nothing yet.
All of this is setting aside the fact that demand for the chips produced by the Terafab facility is predicated entirely on SpaceX's ability to drive down the cost of launches to a level at which orbital datacenters are actually cost-effective.
"As soon as the cost to orbit drops to a low number, it immediately makes extremely compelling sense to put AI in space," Musk asserted. "Increasing power on earth has become harder over time and more expensive over time, but in space, it actually becomes cheaper and easier over time."
If the premise of a space-based AI datacenter sounds like the kind of thing that'd land you in the mental hospital, don't worry, you're not nuts. Space just happens to be the next stop of the AI hype train right after AGI, and Bezos, Google, and Nvidia are all along for the ride.
Not everyone, however, is convinced. In a research note published in February, Gartner analyst Bill Ray described the concept of orbital datacenters as "peak insanity."
"Companies are wasting money by pouring funds into the orbital data center 'bubble,' because the economics do not work," he wrote. "This is due to the prohibitive costs of launching hardware and the immense technical challenges of cooling these orbital datacenters in the vacuum that is space."
Naturally, Musk's Terafab wouldn't just produce chips for orbital datacenter platforms. It would also fab parts for Tesla EVs and humanoid robots. In three to five years. Assuming he ever breaks ground, and the Terafab isn't just another endless kicking-the-can-down-the-road dream like having a Tesla drive itself coast to coast by the end of 2017, putting a million Tesla robotaxis on the road by 2020, landing humans on Mars by 2022, and creating humanoid robots that are not just remote-controlled demoware.
So, perhaps Intel just wants to lock down Tesla's foundry business before Musk gets sidetracked again by something more lucrative than turning humanity into an interplanetary civilization.
Intel declined to comment on the Terafab collaboration, directing us back to its social media posts on X and LinkedIn.