← All videos

Elon Musk's TERAFAB is INSANE !!

The Space Grid Published Mar 23, 2026 Added 5mo ago 11:25 256 views Open on YouTube ↗

Description

Elon Musk's New TERAFAB is INSANE !!

On March 21, 2026, Elon walked onto a stage at a defunct power plant in Austin, Texas, pointed at the global semiconductor supply chain, a marvel of human engineering that took fifty years and trillions of dollars to build and essentially said, "ofcourse, I can do that myself. Hold my Drink.

He announced something called "Terafab." It is a massive, $20 to $25 billion joint venture between Tesla, SpaceX, and his AI company, xAI. It’s a project so incredibly ambitious, so technologically ridiculous, that It's actually insane to even understand the scale of it.

If you are a space news lover, you already know all about Starship and the race to Mars. But this announcement? This isn't about getting to space. This is about building the brain for the spacecraft, the satellites, and the androids of the future and all of that in space.

terafab,elon musk,spacex,grock,ai,elon,tesla,optimus robots,optimus,latest space news,Elon Musk,Terafab,Tesla,SpaceX,xA

Transcript

Read auto-generated transcript (2066 words)

Kind: captions Language: en In order to understand the universe, you must explore the universe. And that's the motivation to accelerate humanity's future in understanding the universe and extending the light of consciousness to the stars. >> [music] [music] >> Heat. Heat. >> [music] [music] >> important announcement to make which is uh the most uh epic chip building exercised in history by far. [cheering] This is really going to take things to the next level. We aspire to be a galactic organ galactic civilization. Ladies and gentlemen, welcome to Terrafab, the most ambitious, most expensive, most potentially insane project in human history. Or you can say the beginning of a galactic civilization. Strap in because this one's going to blow your mind. What is Terraab? [music] Let's start from the top because this thing is genuinely hard to wrap your head around. Terraab is a joint chip fabrication facility being built by Tesla, SpaceX, and XAI. Three companies, [music] one man. The facility is planned for the north campus of Giga Texas in Austin and is designed to consolidate every single stage of semiconductor production under one roof chip [music] design, lithography, fabrication, memory production, advanced packaging and testing. Think of it like this. Right now, when Tesla needs [music] a chip, they call TSMC in Taiwan. TSMC builds a chip, [music] ships it over, Tesla puts it in a car or a robot. That's the normal way. That's how literally [music] every major tech company operates. What Elon is proposing is why depend on anyone when you can do everything yourself inside a [music] single building in Texas. Musk's goal is to produce more than a terowatt. That's 1 trillion watts of AI computing power per year. [music] 1 terowatt. For reference, the current output of AI compute across [music] the entire planet is roughly 20 gawatt per year. So Elon wants to produce [music] by himself 50 times what the entire world currently produces in AI compute from one facility in Austin. That's not ambitious. That's astronomical. That is a man pointing at Mount [music] Everest and saying, "You know what? I'm going to build a second one but taller and I'll own it." And Musk had a very musky way of explaining the name. Terrafab is a next step towards harnessing the power of the sun and creating a galactic civilization. A galactic civilization. Not a slightly better civilization. Not a hey, we fixed the pothole civilization, but a galactic civilization. Why is Elon Musk doing this? Okay, but why? Why now? Because Elon [music] Musk is running into a wall, specifically a silicon wall. At Tesla's Q4 2025 earnings call, Musk warned investors that external chip capacity [music] from TSMC, Samsung, and Micron would hit a limit within 3 to four years. Think about what [music] Musk needs chips for right now. Tesla's self-driving cars need chips. The Optimus humanoid [music] robots need chips. The Cyber Cab robot axis need chips. The orbital satellite constellation SpaceX is trying to launch also needs chips. very specifically designed chips that [music] can survive in space. Terrafab will produce two categories of chips. Inference chips [music] for Tesla vehicles and Optimus robots such as the current AI4 and upcoming [music] AI5 and D3 chips custom design for orbital AI satellites. [music] And here's the part that honestly makes a whole thing make sense, even if you think Musk is overreaching. [music] Musk's projections for Optimus require chip volumes that no existing external [music] supplier can commit to on Tesla's timeline. If you want to build millions of humanoid robots, you can't just call TSMC and say, "Hey, give me all your chips." They have AMD, Apple, Nvidia, and about 400 other clients in the queue. You're not special. Musk himself acknowledged this at the event. [music] He said, and I love this quote, we either build a terapab or we don't have the chips. And we [music] need the chips, so we build a terraab. That's it. That's the whole pitch. A $25 billion. We don't have a choice pitch. The scale of the thing. Now, let's talk about how absolutely enormous this project is because the numbers here are genuinely staggering. The cost estimate is $20 billion to $25 [music] billion for the Terafab facility itself. And that's on top of Tesla's existing [music] 2026 capital expenditure of more than $20 billion. Tesla's CFO actually admitted that the full cost of Terapab is not [music] yet included in that 2026 budget. So when they say $20 billion plan for the year, Terra Fab is extra like we forgot to mention the >> [music] >> $25 billion side project. My bad. Tesla targets two nanometer process technology. [music] the most advanced node currently entering commercial production. For context, two nanometers chips are [music] the cutting edge, the bleeding edge, the edge so sharp it can cut itself. TSMC spent [music] $165 billion over years to build six fabs in Arizona. [music] And those fabs won't even reach 2 nmters production until 2029. Tesla wants to do it faster with no experience from [music] scratch. The Austin Fab will house equipment for logic, memory, packaging, testing, and lithography mass production in a single building. That is not a chip factory. That is a chip civilization. And if you're wondering what that looks like in the night sky, bright beams [music] of light projected from the old SEO power plant into the sky during the announcement, prompting confused questions from Austinites on social media. Imagine texting your [music] friend, "Hey, what are those lights?" And the answer is, oh, Elon is announcing a $25 [music] billion chip factory that will build the future robotics and space exploration. Just another Saturday in Austin, the Carterf scale thing. Now, here's where the event got truly must-pill. He didn't just talk about chips. He talked about the Cardartesef scale. If you don't know what that is, the Cartesef scale is a theoretical framework that ranks [music] civilizations by how much energy they can harness. A type 1 civilization harnesses all the energy of [music] its planet. Type two harnesses its star. Type three harnesses its entire galaxy. Elon's ultimate aim is presented at the Terrafab event is to transition humanity into a galactic civilization and [music] a multilanet species. The primary goal in his words is to make science fiction into science fact. So we went from chip fab in Austin to [music] we are becoming a multilanetary galactic species in about 15 minutes of presentation. This is normal misbehavior. This is just how he speaks. He cannot announce [music] a product without tying it to the fate of human consciousness itself. And honestly [music] part of you has to respect the audacity. Most CEOs announce a new [music] product and say we expect 12% revenue growth in Q3. Elon announces a chip factory and says, [music] "And this will help us colonize a galaxy." The governor of Texas was in the crowd nodding along. Greg Abbott reportedly posted afterward, "Thanks, Elon. Extraordinary event tonight. Your vision is powerful [music] and we are proud of all you do in Texas. Texas just loves Elon." They gave him the power plant [music] in the night sky and presumably someone brought chips, the snack kind. What happens next? So, where does this go from here? Small batch production of the AI5 chip is expected in 2026 with volume production [music] projected for 2027. Though Tesla already delayed the AI5 to mid 2027 before this announcement and the AI6 chip has been delayed roughly 6 months due to Samsung's 2nm production slipping. Multiple job openings under Terrafab [music] on Tesla's website shows roles based in Austin and Palo Alto, [music] California. Construction work was already visible near the Giga Texas site in the days before the announcement. So, the project is real. The construction appears to have begun. The jobs are being posted. Texas is thrilled. The governor is tweeting. Competitors like Whimo and GM's crews remain dependent on third party [music] silicon, leaving them exposed to the same supply chain vulnerabilities, which Tesla is now working to eliminate entirely. If terra fab works [music] even partially, Tesla gains an enormous strategic advantage. It would own its AI stack from software to silicon. That's something no western automotive or robotics company [music] has ever done. But if it fails or delays, it will go down as the most expensive [music] tech announcement in history that produced nothing. The honest answer is nobody knows [music] yet. Musk gave no specific timeline for when Terraab would begin producing chips or reach its target output. No dates, no [music] commitments, just vibes, light shows, and the Cardishef scale. So, let's [music] bring it home. Terra Fab is either the most visionary infrastructure project of the 21st century or [snorts] it's the world's most expensive [music] press release. It's a $25 billion bet that one man who has no background in semiconductor [music] manufacturing, can outbuild TSMC, which has been doing this for 40 years with a team he's still [music] hiring, targeting a technology node that the entire global industry won't reach until 2029. [music] And he wants to use it to power robots, self-driving taxis, and orbital satellites to build a galactic civilization. Maybe must pulls it off. He's done the impossible before. SpaceX was supposed to be a joke. Tesla was [music] supposed to go bankrupt. The guy has a habit of making skeptics eat their words. But he also has a habit of announcing timelines [music] that make absolutely no contact with reality. The truth is somewhere in between. Terra Fab will most probably get built. It probably won't [music] hit one ter by whenever he vaguely implied. It will probably [music] change the semiconductor landscape in ways nobody fully understands yet. It will almost certainly be announced in a [music] decommissioned power plant with lasers shooting into the sky because that's just how Elon does things. [music] If you're new here and you enjoyed this breakdown, make sure you subscribe. I cover the biggest stories in tech in [music] space without the PR filter and without the sensationalism, just the real story told straight. Stay curious and I will see you in the next >> [music]

Related coverage