正在加载视频...

视频加载失败

加载此视频时出现问题。这可能是由于临时网络问题，或视频可能不可用。

TSMC is exploring a role in Terafab, Elon's Texas chip project. Elon's reply was "Just discussions, but something may come of it." Jeff Lutz 🔋 calls it the battery playbook again. Build at a scale nobody else will, and suppliers can't afford to sit out. Intel's already in, Samsung's... show more

Herbert Ong

106,675 subscribers

224,783 次观看 • 1 天前 •via X (Twitter)

新闻政治财经科学技术

Anya Rossi• Live Now

Private livecam show

5 条评论

Elon Musk 的头像

Elon Musk16 小时前

@thejefflutz No, we will build and run the fab. Let there be ZERO doubt about that. Maybe TSMC subleases part of the Terafab if they want, but nothing more than that.

Amy 的头像

Amy21 小时前

@thejefflutz Great discussion and I agree with Jeff’s analysis.

Relax_to_Rich 的头像

Relax_to_Rich21 小时前

@thejefflutz Scale brings negotiating leverage with suppliers. Fab construction takes years with many unknowns, talks don’t equal finished deals.

Jamie 的头像

Jamie1 天前

I could be wrong, but I really don't recall a product where Musk gives away control of a critical asset to another company. EVs, FSD, energy, satellites, datacenters, space launch....horizontal and vertical supply chains have always been seen as critical to success for his companies. TSMC owning and running THE fab that makes SpaceXAI and Tesla dependent on someone else in my view would be a no go. Of course, fabs aren't actually rocket science😎

Vikram Kumar 的头像

Vikram Kumar1 天前

@thejefflutz You and @Nainahf are my go-to for chip news, and the battery playbook framing fits here.

相关视频

🚨 NEWS: Tesla Picks Texas for Terafab Tesla’s massive Terafab will be built in Grimes County, Texas, with a goal of becoming the largest chip manufacturing facility ever built. With Tesla and SpaceX expecting chip demand to outgrow global supply, Terafab is targeting over 1 terawatt of compute per year. The research fab already underway at Giga Texas will serve as its precursor. $TSLA $SPCX

🚨 NEWS: Tesla Picks Texas for Terafab Tesla’s massive Terafab will be built in Grimes County, Texas, with a goal of becoming the largest chip manufacturing facility ever built. With Tesla and SpaceX expecting chip demand to outgrow global supply, Terafab is targeting over 1 terawatt of compute per year. The research fab already underway at Giga Texas will serve as its precursor. $TSLA $SPCX

Herbert Ong

44,302 次观看 • 2 个月前

Terafab is starting to sound like a lot more than just a chip factory! Tesla and SpaceX are reportedly putting $16.8B into phase one, with plans to handle everything from chip design and fabrication to memory, packaging and testing. The idea is basically to bring pieces of what TSMC, Micron and NVIDIA do under one roof. And the chips could eventually power everything from Tesla cars and Optimus to SpaceX data centers and AI compute in space. $TSLA $SPCX

Terafab is starting to sound like a lot more than just a chip factory! Tesla and SpaceX are reportedly putting $16.8B into phase one, with plans to handle everything from chip design and fabrication to memory, packaging and testing. The idea is basically to bring pieces of what TSMC, Micron and NVIDIA do under one roof. And the chips could eventually power everything from Tesla cars and Optimus to SpaceX data centers and AI compute in space. $TSLA $SPCX

Herbert Ong

31,110 次观看 • 1 个月前

Elon Musk says every new chip fab takes five years from start to volume. Every AI company on Earth needed more chips by 2030. OpenAI. Anthropic. Google. Meta. Each one had built the budget. The chips didn't exist yet. Capacity through 2030 had already been booked at every major fab. "I ask TSMC or Samsung, okay, what's the timeframe to get to volume production?" Break ground. Climb the yield curve. Reach volume. Each stage took at least a year, and the curve refused to flatten. Yield meant percent of wafers that ship. Yield meant survival. Then Musk dropped the actual number. "That, from start to finish, is a five-year period." Musk gave the constraint a name: **the five-year fab clock**. Five years. No shortcuts. Musk, who had prepaid TSMC and Samsung through 2030, knew the clock. His own AI5 chip targets production in 2026's second quarter. A new fab needed to break ground, install ASML tools, train tens of thousands of operators, run pilot wafers, debug each process step, and ramp from one wafer per day to fifty thousand. TSMC Taiwan. Samsung Korea. TSMC Arizona. Samsung Texas. All booked. Every shovel of dirt accounted for. After Musk drew the clock, every 2027 AI promise became unfulfillable. Compute would have to wait for cement to dry. Musk, on the suppliers who were trying: "They're going balls to the wall, as fast as they can. It's still not fast enough." What ceiling on your roadmap is bound by physical constraints nobody is naming? P.S. If you want to stop overthinking, control chaos, and navigate any decision with the clarity... I made a full playbook breaking down the timeless decision-making mental models used by history's greatest thinkers. Comment "models" and follow GeniusThinking so I can DM you a copy. If you're new here, GeniusThinking is a gallery for the greatest minds in economics, psychology, and history. Follow along for more similar content. — Elon Musk ( Elon Musk ), CEO of Tesla and SpaceX, on Dwarkesh Patel's ( Dwarkesh Patel ) podcast

Elon Musk says every new chip fab takes five years from start to volume. Every AI company on Earth needed more chips by 2030. OpenAI. Anthropic. Google. Meta. Each one had built the budget. The chips didn't exist yet. Capacity through 2030 had already been booked at every major fab. "I ask TSMC or Samsung, okay, what's the timeframe to get to volume production?" Break ground. Climb the yield curve. Reach volume. Each stage took at least a year, and the curve refused to flatten. Yield meant percent of wafers that ship. Yield meant survival. Then Musk dropped the actual number. "That, from start to finish, is a five-year period." Musk gave the constraint a name: the five-year fab clock. Five years. No shortcuts. Musk, who had prepaid TSMC and Samsung through 2030, knew the clock. His own AI5 chip targets production in 2026's second quarter. A new fab needed to break ground, install ASML tools, train tens of thousands of operators, run pilot wafers, debug each process step, and ramp from one wafer per day to fifty thousand. TSMC Taiwan. Samsung Korea. TSMC Arizona. Samsung Texas. All booked. Every shovel of dirt accounted for. After Musk drew the clock, every 2027 AI promise became unfulfillable. Compute would have to wait for cement to dry. Musk, on the suppliers who were trying: "They're going balls to the wall, as fast as they can. It's still not fast enough." What ceiling on your roadmap is bound by physical constraints nobody is naming? P.S. If you want to stop overthinking, control chaos, and navigate any decision with the clarity... I made a full playbook breaking down the timeless decision-making mental models used by history's greatest thinkers. Comment "models" and follow GeniusThinking so I can DM you a copy. If you're new here, GeniusThinking is a gallery for the greatest minds in economics, psychology, and history. Follow along for more similar content. — Elon Musk ( Elon Musk ), CEO of Tesla and SpaceX, on Dwarkesh Patel's ( Dwarkesh Patel ) podcast

GeniusThinking

93,319 次观看 • 4 个月前

Many still don’t understand why Elon is building Terafab Terafab is an extension to all the chip makers in the world It’s not about replacement, not a rivalry and absolutely not competing It’s being built to fulfill the massive chip orders that Tesla, SpaceX and xAI actually need TSMC’s most advanced 2nm capacity is totally booked through 2028 Tesla signed a massive $16.5 billion deal with Samsung back in July 2025 to produce AI6 chips at their Taylor, Texas factory and Samsung is building a Tesla Exclusive chip manufacturing plant to full fill this orders When Elon announced Terafab on March 21, 2026...he made it clear: “That rate is much less than we’d like. We either build the Terafab or we don’t have the chips, and we need the chips, so we build the Terafab” He basically told the chip makers: “Produce as much as you comfortably can. We will take them all. Actually we want even more” Even today Elon said: "SpaceX/Tesla will be always be major customers of TSMC and not competitors in the normal sense of the word" Current production rates are much less than they need....That’s why Terafab exists Terafab is an extension to every chip maker… not competition, not rivalry, absolutely not Even Intel has joined as a partner Even if chip supply improves, massive bottlenecks still exist with memory and advanced packaging You simply can't risk those supply chain breaks at this scale That’s why Terafab is being built to vertically integrate everything - chips, memory, advanced packaging all under one roof, targeting 1 terawatt of AI compute capacity per year This is a ludicrous amount of chips that no chipmaker currently produces at this scale. I don't think even TSMC and Samsung truly understand these numbers yet It’s about building the capacity the future actually demands

Many still don’t understand why Elon is building Terafab Terafab is an extension to all the chip makers in the world It’s not about replacement, not a rivalry and absolutely not competing It’s being built to fulfill the massive chip orders that Tesla, SpaceX and xAI actually need TSMC’s most advanced 2nm capacity is totally booked through 2028 Tesla signed a massive $16.5 billion deal with Samsung back in July 2025 to produce AI6 chips at their Taylor, Texas factory and Samsung is building a Tesla Exclusive chip manufacturing plant to full fill this orders When Elon announced Terafab on March 21, 2026...he made it clear: “That rate is much less than we’d like. We either build the Terafab or we don’t have the chips, and we need the chips, so we build the Terafab” He basically told the chip makers: “Produce as much as you comfortably can. We will take them all. Actually we want even more” Even today Elon said: "SpaceX/Tesla will be always be major customers of TSMC and not competitors in the normal sense of the word" Current production rates are much less than they need....That’s why Terafab exists Terafab is an extension to every chip maker… not competition, not rivalry, absolutely not Even Intel has joined as a partner Even if chip supply improves, massive bottlenecks still exist with memory and advanced packaging You simply can't risk those supply chain breaks at this scale That’s why Terafab is being built to vertically integrate everything - chips, memory, advanced packaging all under one roof, targeting 1 terawatt of AI compute capacity per year This is a ludicrous amount of chips that no chipmaker currently produces at this scale. I don't think even TSMC and Samsung truly understand these numbers yet It’s about building the capacity the future actually demands

X Freeze

53,291 次观看 • 5 个月前

Ben on Intel and TSMC: "A lot of tech companies didn't fully appreciate the extent to which TSMC offloaded risk onto the big tech companies. The risk TSMC is worried about is overcapacity. If we build a fab, we expect that fab to run for 30 years. So they are very biased toward being much more conservative. Risk doesn't disappear. It just moves. Today, when we say there's not enough compute, it's not like all the money that companies are putting in today manifests in compute tomorrow. It manifests in compute in 2028 and 2029. But this is where TSMC in some respects made the same mistake as the memory makers. Because they didn't invest, the shortages are going to be so acute, big tech companies that were foregoing so much revenue and so many profits will go through the pain of getting Intel up to speed, of getting Samsung up to speed. It never made rational sense for anyone to go work with Intel. In an unchanging world, TSMC would just win forever. The scarcity is what ultimately saved Intel. It was ultimately TSMC brought it on themselves."

Ben on Intel and TSMC: "A lot of tech companies didn't fully appreciate the extent to which TSMC offloaded risk onto the big tech companies. The risk TSMC is worried about is overcapacity. If we build a fab, we expect that fab to run for 30 years. So they are very biased toward being much more conservative. Risk doesn't disappear. It just moves. Today, when we say there's not enough compute, it's not like all the money that companies are putting in today manifests in compute tomorrow. It manifests in compute in 2028 and 2029. But this is where TSMC in some respects made the same mistake as the memory makers. Because they didn't invest, the shortages are going to be so acute, big tech companies that were foregoing so much revenue and so many profits will go through the pain of getting Intel up to speed, of getting Samsung up to speed. It never made rational sense for anyone to go work with Intel. In an unchanging world, TSMC would just win forever. The scarcity is what ultimately saved Intel. It was ultimately TSMC brought it on themselves."

Patrick OShaughnessy

101,529 次观看 • 1 个月前

ELON: TESLA'S PEDAL TO THE METAL ON AI5 & AI6 "Tesla's pedal to the metal max production is to get Tesla AI5 chip design into production and then reaching scale, that'll probably happen around the second quarter of next year. AI6 would hopefully follow less than a year later. We've secured all the chip fab production that we can using TSMC Taiwan, Samsung Korea, TSMC Arizona, Samsung Texas. The limiting factor once you can get to space is chips, but the limiting factor before you can get to space will be power." Source: Cheeky Pint Podcast, Elon Musk

ELON: TESLA'S PEDAL TO THE METAL ON AI5 & AI6 "Tesla's pedal to the metal max production is to get Tesla AI5 chip design into production and then reaching scale, that'll probably happen around the second quarter of next year. AI6 would hopefully follow less than a year later. We've secured all the chip fab production that we can using TSMC Taiwan, Samsung Korea, TSMC Arizona, Samsung Texas. The limiting factor once you can get to space is chips, but the limiting factor before you can get to space will be power." Source: Cheeky Pint Podcast, Elon Musk

Mario Nawfal

72,185 次观看 • 8 个月前

During a presser with U.S. President Donald J. Trump, U.S. Secretary of Commerce Howard Lutnick, TSMC CEO C. C. Wei, and “White House Crypto and AI Czar” David Sachs, the president announced that TSMC will be investing at least $100 billion in new capital towards building out domestic chip manufacturing in the United States. Commerce Secretary Howard Lutnick said the move is one of many to “bring production to America” and to avoid tariffs.

During a presser with U.S. President Donald J. Trump, U.S. Secretary of Commerce Howard Lutnick, TSMC CEO C. C. Wei, and “White House Crypto and AI Czar” David Sachs, the president announced that TSMC will be investing at least $100 billion in new capital towards building out domestic chip manufacturing in the United States. Commerce Secretary Howard Lutnick said the move is one of many to “bring production to America” and to avoid tariffs.

OSINTdefender

431,843 次观看 • 1 年前

Elon Musk told Ron Baron he's building a chip that will be 2 to 3 times better than NVIDIA - at 10% of the cost "I have the entire physical design of the chip laid out in memory - I can visualize the whole thing" when TSMC told him a new fab takes 5 years he said: "five years to me is an eternity - my timelines go one year, two year, and at year three it goes to infinity" so he's building his own Tesla's self-driving has logged 10 billion miles - 4 times safer than a human driver - with the new chip it will be 10x "I don't own any vacation homes - just one house in Austin and a tiny house at Starbase - friends come to visit and they think I'm kidding" watch the full conversation ↓

Elon Musk told Ron Baron he's building a chip that will be 2 to 3 times better than NVIDIA - at 10% of the cost "I have the entire physical design of the chip laid out in memory - I can visualize the whole thing" when TSMC told him a new fab takes 5 years he said: "five years to me is an eternity - my timelines go one year, two year, and at year three it goes to infinity" so he's building his own Tesla's self-driving has logged 10 billion miles - 4 times safer than a human driver - with the new chip it will be 10x "I don't own any vacation homes - just one house in Austin and a tiny house at Starbase - friends come to visit and they think I'm kidding" watch the full conversation ↓

Сarm1ne

2,651,516 次观看 • 4 个月前

$TSLA says Terafab will be built in Grimes County Texas as both it and $SPCX prepare for chip demand beyond global supply. The company plans to build “the largest chip manufacturing facility ever” with a goal of producing more than 1 terawatt of compute annually.

$TSLA says Terafab will be built in Grimes County Texas as both it and $SPCX prepare for chip demand beyond global supply. The company plans to build “the largest chip manufacturing facility ever” with a goal of producing more than 1 terawatt of compute annually.

Shay Boloor

154,769 次观看 • 2 个月前

Elon Musk told Ron Baron he's building a chip 2 to 3 times better than Nvidia, at 10% of the cost. I have the entire physical design of the chip laid out in memory. I can visualize the whole thing. No CAD software, no design team walking him through it, just memory. When TSMC told him a new fab takes 5 years he said: "five years to me is an eternity. My timelines go one year, two year, and at year three it goes to infinity." So he's building his own Same mindset runs the factory floor. Tesla's cycle time per car is already down to 35 seconds, with a target of 10. Tesla's self-driving has logged 10 billion miles, 4 times safer than a human driver. With the new chip it'll be 10x. I don't own any vacation homes. Just one house in Austin and a tiny house at Starbase. Friends come to visit and they think I'm kidding. Watch the full conversation ⤵

Elon Musk told Ron Baron he's building a chip 2 to 3 times better than Nvidia, at 10% of the cost. I have the entire physical design of the chip laid out in memory. I can visualize the whole thing. No CAD software, no design team walking him through it, just memory. When TSMC told him a new fab takes 5 years he said: "five years to me is an eternity. My timelines go one year, two year, and at year three it goes to infinity." So he's building his own Same mindset runs the factory floor. Tesla's cycle time per car is already down to 35 seconds, with a target of 10. Tesla's self-driving has logged 10 billion miles, 4 times safer than a human driver. With the new chip it'll be 10x. I don't own any vacation homes. Just one house in Austin and a tiny house at Starbase. Friends come to visit and they think I'm kidding. Watch the full conversation ⤵

Pixel

114,339 次观看 • 20 天前

Elon Musk: Most engineering is done with people who don’t have PhDs. You don’t need a PhD for that — but you do need competent personnel. Right now, Tesla is pedal to the metal, going as fast as possible to get AI5 chip design into production, reach full scale, and secure every chip fab we possibly can. If everything goes at maximum speed, AI5 will reach production around the second quarter of next year, and then AI6 would hopefully follow less than a year later. We are using TSMC Taiwan, Samsung Korea, TSMC Arizona, and Samsung Texas — and even after securing all of that, we still don’t have enough capacity. When you ask, “What’s the timeframe to reach volume production?” the answer is: You’ve got to build the fab, start production, climb the yield curve, and reach high-yield volume production — and that takes about 5 years.

Elon Musk: Most engineering is done with people who don’t have PhDs. You don’t need a PhD for that — but you do need competent personnel. Right now, Tesla is pedal to the metal, going as fast as possible to get AI5 chip design into production, reach full scale, and secure every chip fab we possibly can. If everything goes at maximum speed, AI5 will reach production around the second quarter of next year, and then AI6 would hopefully follow less than a year later. We are using TSMC Taiwan, Samsung Korea, TSMC Arizona, and Samsung Texas — and even after securing all of that, we still don’t have enough capacity. When you ask, “What’s the timeframe to reach volume production?” the answer is: You’ve got to build the fab, start production, climb the yield curve, and reach high-yield volume production — and that takes about 5 years.

Ian Miles Cheong

92,034 次观看 • 8 个月前

Everyone is analyzing Terafab as a chip factory. I think that's the wrong frame entirely. It isn't a fab. It isn't a memory plant. It isn't Tesla's answer to TSMC. It's a city. Semiconductor manufacturing is the most fragmented supply chain on earth. Lithography, fabrication, memory, packaging, test, each step sits with a different company, in a different country, with handoffs measured in weeks. Nobody owns the whole stack because nobody can. So if you want to collapse that, you don't build a better fab. You build a place where every specialist can physically move in, bring their core competency, and have it connect directly to everyone else's. Infrastructure designed around the seams rather than around one process step. The tell is who's already involved. Intel is in. Samsung is in. Nvidia is clearly interested. If Micron or SK Hynix show up next, or ASML, the pattern confirms itself. One site, one stack, no handoffs. $TSLA $SPCX

Everyone is analyzing Terafab as a chip factory. I think that's the wrong frame entirely. It isn't a fab. It isn't a memory plant. It isn't Tesla's answer to TSMC. It's a city. Semiconductor manufacturing is the most fragmented supply chain on earth. Lithography, fabrication, memory, packaging, test, each step sits with a different company, in a different country, with handoffs measured in weeks. Nobody owns the whole stack because nobody can. So if you want to collapse that, you don't build a better fab. You build a place where every specialist can physically move in, bring their core competency, and have it connect directly to everyone else's. Infrastructure designed around the seams rather than around one process step. The tell is who's already involved. Intel is in. Samsung is in. Nvidia is clearly interested. If Micron or SK Hynix show up next, or ASML, the pattern confirms itself. One site, one stack, no handoffs. $TSLA $SPCX

Jo Bhakdi

10,955 次观看 • 20 天前

Jeff Lutz was working at Motorola when Apple launched the first iPhone. He says nobody there was laughing. The story everyone repeats is that the phone companies mocked it. Lutz was in the building and remembers something else. The two companies had already shipped a project together, the first phone to carry iTunes. Apple executives were still carrying Motorola Razrs right up to the launch. What the iPhone did was force companies to write off their existing business and turn straight into smartphones. "I can tell you this at Motorola, we were not laughing." - They had shipped a project together - Apple executives carried Razrs themselves - Deep mutual respect between both companies Lutz calls the version everyone repeats bad information. — Jeff Lutz (Jeff Lutz 🔋) on Brighter with Herbert

Jeff Lutz was working at Motorola when Apple launched the first iPhone. He says nobody there was laughing. The story everyone repeats is that the phone companies mocked it. Lutz was in the building and remembers something else. The two companies had already shipped a project together, the first phone to carry iTunes. Apple executives were still carrying Motorola Razrs right up to the launch. What the iPhone did was force companies to write off their existing business and turn straight into smartphones. "I can tell you this at Motorola, we were not laughing." - They had shipped a project together - Apple executives carried Razrs themselves - Deep mutual respect between both companies Lutz calls the version everyone repeats bad information. — Jeff Lutz (Jeff Lutz 🔋) on Brighter with Herbert

GeniusThinking

29,253 次观看 • 26 天前

ONE DAY BEFORE THE BIGGEST IPO IN HISTORY, MUSK IS PITCHING A $119 BILLION CHIP FACTORY TO THE ONLY COMPANY THAT CAN BUILD IT. 🇺🇸 This week Musk speaks at ASML's private conference, the firm whose machines print every advanced chip on earth. No ASML, no Nvidia, no TSMC. The topic isn't a secret. It's Terafab, a SpaceX-Tesla-xAI plan for a 2nm fab in Texas. First phase $55 billion, total up to $119 billion. The timing isn't random. Days before SpaceX goes public, Musk is handing IPO investors a growth story beyond rockets: Custom silicon for orbital AI data centers. This is how you justify a $1,750,000,000,000 valuation.

ONE DAY BEFORE THE BIGGEST IPO IN HISTORY, MUSK IS PITCHING A $119 BILLION CHIP FACTORY TO THE ONLY COMPANY THAT CAN BUILD IT. 🇺🇸 This week Musk speaks at ASML's private conference, the firm whose machines print every advanced chip on earth. No ASML, no Nvidia, no TSMC. The topic isn't a secret. It's Terafab, a SpaceX-Tesla-xAI plan for a 2nm fab in Texas. First phase $55 billion, total up to $119 billion. The timing isn't random. Days before SpaceX goes public, Musk is handing IPO investors a growth story beyond rockets: Custom silicon for orbital AI data centers. This is how you justify a $1,750,000,000,000 valuation.

Merlijn The Trader

246,413 次观看 • 4 个月前

Japan Quietly Upgraded Its Military — With Taiwan’s Help. This TSMC Move Just Made a Taiwan Invasion Much Harder. China keeps talking about invading Taiwan. But while everyone was watching the headlines, something far more important happened. TSMC—the world’s most advanced and profitable chipmaker—decided to bring 3-nanometer production to Japan. Not legacy chips. Not trailing nodes. The crown jewels. In this video, I break down how Taiwan and TSMC just helped Japan upgrade its deterrence capabilities at the silicon level—why this move quietly reshapes the military balance in East Asia, raises the cost of a Taiwan invasion, and leaves China structurally locked out of the future.

Japan Quietly Upgraded Its Military — With Taiwan’s Help. This TSMC Move Just Made a Taiwan Invasion Much Harder. China keeps talking about invading Taiwan. But while everyone was watching the headlines, something far more important happened. TSMC—the world’s most advanced and profitable chipmaker—decided to bring 3-nanometer production to Japan. Not legacy chips. Not trailing nodes. The crown jewels. In this video, I break down how Taiwan and TSMC just helped Japan upgrade its deterrence capabilities at the silicon level—why this move quietly reshapes the military balance in East Asia, raises the cost of a Taiwan invasion, and leaves China structurally locked out of the future.

Ken Cao-The China Crash Chronicle

82,836 次观看 • 8 个月前

Elon Musk explains why Terafab must be built and gives an update on the Tesla and SpaceX R&D chip fab in Austin: “For some reason, at some point, chips may not continue coming here from Taiwan. That would really make things difficult if we didn’t have any chips. That’s an important reason to have Terafab. There needs to be certainty of future supply of chips, even if things become challenging geopolitically.” “Long term, there’s a scaling challenge. If you want to really scale AI, both at the servers and for edge compute, humanoid robotics and cars, you kind of run out of capacity with the existing fabs. All the fabs are running at max capacity. It’s either build Terafab or fail to scale. Those are the two options.” “There’s an R&D fab that we’re building in Austin. That’s a collaboration between Tesla and SpaceX. We have all the equipment on order. I think we’ll probably be able to make something useful by the end of next year. It really is crawl, walk, run. The first step is, can we make anything? Then make useful chips at scale. And then run would be, now let’s make them at massive scale.”

Elon Musk explains why Terafab must be built and gives an update on the Tesla and SpaceX R&D chip fab in Austin: “For some reason, at some point, chips may not continue coming here from Taiwan. That would really make things difficult if we didn’t have any chips. That’s an important reason to have Terafab. There needs to be certainty of future supply of chips, even if things become challenging geopolitically.” “Long term, there’s a scaling challenge. If you want to really scale AI, both at the servers and for edge compute, humanoid robotics and cars, you kind of run out of capacity with the existing fabs. All the fabs are running at max capacity. It’s either build Terafab or fail to scale. Those are the two options.” “There’s an R&D fab that we’re building in Austin. That’s a collaboration between Tesla and SpaceX. We have all the equipment on order. I think we’ll probably be able to make something useful by the end of next year. It really is crawl, walk, run. The first step is, can we make anything? Then make useful chips at scale. And then run would be, now let’s make them at massive scale.”

DogeDesigner

127,258 次观看 • 22 天前

Elon Musk said Tesla will soon reveal the location and full plans for Terafab and the announcement will get its own dedicated spotlight: “Terafab is going to be an amazing initiative and a necessary one Without it, we will be constrained in our ability to scale Optimus production because we simply will not have enough AI chips We have to solve memory, logic and packaging in order to scale Optimus” Tesla has already placed equipment orders for a development fab in Austin that will bring: • Lithography mask production • Logic chips • Memory • Advanced packaging • Chip testing All under one roof This would allow Tesla to design a new AI chip, manufacture it, package it, test it and quickly iterate on the next version from the same facility A building like this does not exist anywhere else on Earth Terafab is not just another factory It is going to become the chip-development engine Tesla needs to scale Optimus and pursue high-risk, high-reward AI hardware breakthroughs at unprecedented speed

Elon Musk said Tesla will soon reveal the location and full plans for Terafab and the announcement will get its own dedicated spotlight: “Terafab is going to be an amazing initiative and a necessary one Without it, we will be constrained in our ability to scale Optimus production because we simply will not have enough AI chips We have to solve memory, logic and packaging in order to scale Optimus” Tesla has already placed equipment orders for a development fab in Austin that will bring: • Lithography mask production • Logic chips • Memory • Advanced packaging • Chip testing All under one roof This would allow Tesla to design a new AI chip, manufacture it, package it, test it and quickly iterate on the next version from the same facility A building like this does not exist anywhere else on Earth Terafab is not just another factory It is going to become the chip-development engine Tesla needs to scale Optimus and pursue high-risk, high-reward AI hardware breakthroughs at unprecedented speed

X Freeze

18,749 次观看 • 2 个月前

ELON MUSK: All the fabs on Earth combined are only 2% of what Terafab needs. “If you add up all the fabs on Earth combined, they’re only about 2% of what we need for the Terafab project. We’re very grateful to Samsung, TSMC and others. We would like them to expand as quickly as they can, and we will buy all of their chips. There’s a maximum rate at which they’re comfortable expanding, but that rate is much less than we would like. So we’re going to build Terafab. The key missing ingredient is a terawatt of compute. That’s why we are building Terafab.”

ELON MUSK: All the fabs on Earth combined are only 2% of what Terafab needs. “If you add up all the fabs on Earth combined, they’re only about 2% of what we need for the Terafab project. We’re very grateful to Samsung, TSMC and others. We would like them to expand as quickly as they can, and we will buy all of their chips. There’s a maximum rate at which they’re comfortable expanding, but that rate is much less than we would like. So we’re going to build Terafab. The key missing ingredient is a terawatt of compute. That’s why we are building Terafab.”

DogeDesigner

6,178,869 次观看 • 2 个月前