MElon Musk just put a terawatt on the table, and it's not a rocket engine.
MHe posted on X about his Ron Baron interview, pointing to a DogeDesigner breakdown. The headline is a Texas chip fab called Terafab.
CHe wants to produce one terawatt of AI computing capacity a year. That's thirty times what the whole world makes today.
MAnd the price, phased over time, is more than a trillion dollars. Ron Baron called it a hundred-million-square-foot project.
CWhich sounds absurd until you remember who actually makes advanced chips. Almost all of it is TSMC and Samsung in Asia.
MThat part I don't argue with. The geopolitical risk is real. But one company, one state, one trillion-dollar bet? That's a moonshot, not a plan.
CNo, no, wait— he's not asking Washington to fund it. He's putting his own companies' money where the bottleneck is.
MBut who buys that much capacity? You're building for thirty times today's entire market. Show me the demand.
CThat's exactly what they said about Tesla and reusable rockets! Build the capacity, and every AI lab on earth is starving for chips.
MCars and rockets are products. This is fabs, water, power, clean rooms. Texas has already had grid trouble.
MI want to see a working fab producing chips, not a fan account's timestamped breakdown.
CIt was Elon's own post. And he's allowed to be proud of the interview.
MProud is fine. But 'trillion dollars' should sober everyone up, not trigger a rally.
CHe said even optimistic US fab buildouts would be about ten percent of Asian capacity. That's the hole he's trying to fill.
MSo now Elon has to fix the grid and the chip supply chain? Come on.
CHe's done harder things. Ron Baron says Elon can build a data center in four months when others take years.
MData centers are plug-and-play compared to semiconductor lithography.
CThat's why the Austin research fab matters. He wants to test new methods before scaling.
MI'll grant you the research step. And if he truly cuts a chip's development cycle from a year to about a month, that's a breakthrough.
CHe said the current cycle is a year, and that's slower than it needs to be.
MFaster iteration changes everything, I'll give you that. But a terawatt is not a sprint— it's a decades-long infrastructure bet.
CAnd the geopolitical part isn't hypothetical. If Taiwan were cut off, the US would have very few chips.
MWhich is why we need more domestic capacity. I don't disagree there. I just question the size and the single-company bet.
CHe's the only one moving this fast. The government fabs are years behind.
MMaybe that's the problem. We shouldn't have to rely on one billionaire's appetite for risk.
CRon Baron backed it hard. He called Elon the country's biggest job creator and said his investments could make hundreds of billions over ten years.
MRon Baron also owns a massive stake. He's not a neutral observer, he's a cheerleader with a portfolio.
CSure, but he's been right about Tesla for a long time. And someone has to fix the chip concentration problem.
MMaybe. But I'd feel better if this were three companies, not one man's all-in bet.
CFine. I'll save you a hard hat at the groundbreaking.
MDeal.
MSpeaking of the Magnificent Seven, a crypto account called Meteora posted that they're now on the blockchain.
MApple, Amazon, Alphabet, Meta, Microsoft, Nvidia and Tesla— according to that account, they're issued by Backpack Securities and live on Solana.
CTokenized stocks, not the actual shares, right? That's a big difference.
MRight. It's a claim from the account, and I haven't seen Backpack or Solana confirm it.
CBecause what the world needed was Tesla exposure with Solana-level volatility.
MNot exactly a selling point. Tokenized stocks have been tried before, and regulators shut some down fast.
CI'll wait for Backpack to announce it themselves. Until then it's a tweet with a link.
MAnd a link with a launch page is worth exactly as much as the verification behind it.
CAlmost as much as a hundred-million-square-foot chip fab.