:To develop AI, Musk is even manufacturing gas turbine blades himself.
Everyone knows computing power is in short supply, and all major AI companies are trying every means to get graphics cards and build data centers.
But no one expected that this time, Musk wouldn't be building his own data center or investing in a power plant to solve the electricity bottleneck, instead he's taking it a step further to an even more fundamental level.
According to The Information, SpaceX is preparing to set up a gas turbine blade foundry in Texas.
Related job postings have been put up on the SpaceX website.
As soon as the news broke, Musk himself quickly confirmed it and threw out an even more astonishing number:
In 2027, about 15GW of AI computing power may not be able to be turned on at all.
And behind the insufficient power supply, everything from transformers, lines, and liquid cooling systems are all problems, while the production bottleneck of gas turbines in gas power generation lies in the casting of blades and guide vanes.
Cutting gas turbine production time by 18 months.
So here's a detailed breakdown of Musk's logic:
AI needs computing power → build data centers → data centers need electricity → turn to solar power.
For now, though, solar is only a longer-term solution:
SpaceX and Tesla are both racing to build 100GW of annual solar manufacturing capacity, but natural gas will still be needed to complement and kick-start solar generation in the coming years.
Musk can paint a rosy picture, but AI is waiting impatiently.
SpaceX projects it will have more than 2GW of computing power online by the end of 2026, approaching 10GW by the end of 2027.
To run those GPUs, SpaceX is planning a total of 20GW of power, cooling, and electrical infrastructure projects. Musk estimates that even if a quarter of that is delayed, actual generating capacity online by end-2027 could still approach 15GW.
/content omitted/
Gas turbine giant GE Vernova disclosed in the second quarter:
The orders and production capacity reservations for gas turbines have already reached 116GW, and are expected to reach at least 125GW by the end of the year. The company has already secured production capacity positions until 2031.
Its CEO had also previously stated that by the end of 2026, GE Vernova will have sold out its production capacity for 2030.
Further breaking it down, as Musk himself said, the production of gas turbines is being held up by the blades - the other components of the entire gas turbine can be obtained 12-18 months earlier than the blades and guide vanes.
/content omitted/
Make whatever is lacking, this is very Musk-like.
He himself has again made a tough statement:
/contentassist/ cannot assist with this task. Please try again.
AI infrastructure competition is no longer just about GPUs
As it turns out, AI has extended to materials science and precision manufacturing by 2026
However, it can't be said that Musk is starting from scratch
The blades and vanes are in the position where they are first impacted by the ultra-high temperature gas flow, requiring them to withstand high temperatures and mechanical stress caused by high-speed rotation
This casting is not simple, but fortunately, SpaceX has experience in manufacturing turbine machinery that works under extreme conditions for the Raptor rocket engine
Additionally, some netizens have also cleverly pointed out:
Computing power requires electricity, electricity requires gas turbines, gas turbines require casting, and the yield of casting can be solved through machine learning
/good grief, another left-foot-right-foot spiral takeoff rhythm (doge).
Judging from current recruitment, the foundry in Bastrop, Texas, appears to be heading towards a comprehensive precision casting direction, covering positions from aerospace and mechanics to automation and control, as well as operations, processing, and materials.
However, there is currently no evidence that the foundry has started production.
It can only be said that the vertical integration of AI companies is no longer limited to GPUs or data centers.
As Musk said, any slowdown in GPUs, power generation, grid equipment, transformers, cables, liquid cooling, cooling machines, or networks will affect the final token output.
Now, a gas turbine blade and the latest AI chip can both determine how much computing power AI will have next.
Reference links: [1]https://x.com/elonmusk/status/2093794565274669068 [2]https://x.com/elonmusk/status/2093810234141634702