A look at The Stargate Project: unanswered questions, PR wins, SoftBank only has ~$30B cash, many projects are old news, and Elon Musk trolled the announcement
$500B in CapEx for AI in the US feels like a good thing for US security and for AI's continued progress in capabilities. Google's entire capex in 2024 was $50B. This is a hard amount of money to actually spend. Maybe just 1% of that can go to Gabriel / @gbrl_dick : important to temper your expectations re: stargate imo. historically masa-san has aimed for a real ROI roughly between negative 2 per cent and positive 1 per cent on investments greater than one billion dollars. Sunny Madra / @sundeep : My guess is Stargate will mostly be debt funded... (think Coreweave and most datacenters at this point) At 10x leverage, you only need $10b in equity to fund the first $100b commitment... Pedro Domingos / @pmddomingos : Stargate would look a lot better if DeepSeek hadn't just matched o1 at a tiny fraction of the cost. Peter Yang / @petergyang : In light of Masa's Stargate investment - I present to you his truly visionary slides about AI from a few years ago [image] @basedbeffjezos : Did... OpenAI just raise half a trillion?? @thejeffdo : $500 billion dollars. It's a little bit shy of $7 trillion @sama was looking for before. But it's a good start. 🤯 Anshu Sharma / @anshublog : My company announces $5B in funding... ...using the new formula for announcements. You announce $500B, then commit to $100B for “this year” and then everyone finds out only ~$10B is actually secured. Deedy / @deedydas :Masayoshi Son'sSoftbank investment in the $500B Stargate is far from the first crazy investment he's done. Masa invested $20M for 30% of Jack Ma'sAlibaba in 2000 in a 10 minute meeting at his house. That stake was worth $200B at peak, a 10,000x return. Packy McCormick / @packym : This is so cool until you realize the government could have installed ~500 EV charging stations with this money. @eyuplovely : This feels like something that is going to make the world immeasurably worse in many ways @boltzmannbooty : Effectively doubling the budget of the US intelligence community! Congratulations everybody Ludwig / @ludwigabap : what's the name when animals in danger start posturing and puffing their feathers etc to make themselves appear bigger/more dangerous when they're losing? Glenn / @glennluk : Keep in mind chips & servers are short-lived assets that depreciate quickly; there's a reason accountants calculate useful lives in the 3-5 year range. (This is fundamentally different from long-lived assets like rail & power infrastructure w/ 50+ year useful lives) @basedbeffjezos : He wasn't kidding... 😳 [image] Stephen McAleer / @mcaleerstephen : Compute is not a bottleneck Mark Nelson / @energybants : PROJECT STARGATE ENERGY ANALYSIS Project Stargate is going to need an enormous amount of power: each of the five Stargates needs about 5 gigawatts (about five million people's worth) of constant electricity for each ~$100b supercomputer project. Let's break it down. Each Andreas Bigger / @andreaslbigger : Softbank has $40B in cash. Where is $500B coming from? [image] @pseudolad : AGI for the low low price of just 16 manhattan projects [image] Tom Ough / @tomough : Worth noting the use of the term “Stargate” in 2001: A Space Odyssey. It's a portal via which an astronaut takes what seems to be the next evolutionary step: transcending his corporeal form to become a celestial being of energy. It's a singularity of sorts. Buckle up! @mreiffy : Ironically, 100k jobs created to ultimately end millions @oracle : We are proud to be working with OpenAI, SoftBank, and MGX on the Stargate Project to secure American leadership in AI, create hundreds of thousands of American jobs, and generate massive economic benefit for the entire world. David Ulevitch / @davidu : We need nuclear. Now. A lot. Large reactors. Small reactors. Micro reactors. Pradyumna / @pradyuprasad : > will develop superintelligence > create thousands of jobs ???? Bone / @bonegpt : Where to put your money now that Stargate was announced? A thread on what companies stand to benefit from $500b in AI infrastructure spending 🧵 Luis Garicano / @lugaricano : Europe needs to wake up NOW: The US races ahead with regulatory flexibility and a huge ($500bn!) investment in AI infrastructure. AI is a horizontal technology that will serve as the basis of all future R&D. Without AI, Europe will fall behind in innovation in every sector. Eli Dourado / @elidourado : Time to update the article, lol [image] Henry / @arithmoquine : idk, i feel like people just don't quite get it. this is among the most expensive undertakings in human history. it's a scale mankind has never operated at before. there is literally no parallel Samantha A / @samantha1989tv : anything? [image] @owl_posting : trump: i dont get it sam altman: what dont you get? trump: if you build agi, doesnt that get rid of all jobs? sam altman: no honey...thats not how it works trump: oh. okay. sam altman: trump: sam? sam altman: yes? trump: how does it work? sam altman: well. think Sophie / @netcapgirl : [image] Aidan McLau / @aidan_mclau : fun fact: if you open apple news and scroll aimlessly past ~1000 articles, you'll find zero (0) mentions of the stargate project Justin Amash / @justinamash : The Stargate Project sounds like the stuff of dystopian nightmares—a U.S. government-announced partnership of megacorporations “to protect the national security of America and its allies” and harness AGI “for the benefit of all of humanity.” Let's maybe take a beat here. @tsarnick : Larry Ellison says the Stargate Project will construct the largest computer ever built which will enable AI to create cancer vaccines, personalized medicine and pandemic prevention [video] Sam Stein / @samstein : This is definitely a first for me. A SENIOR official in the White House publicly declaring that the president's big announcement from the day prior is kind of a ruse [image] Gavin Baker / @gavinsbaker : Stargate is a great name but the $500b is a ridiculous number and no one should take it seriously unless SoftBank is going to sell all of their BABA and ARM. SoftBank has $38b in cash, $142b in debt and generates $3ish billion in FCF per BBG. They own $143b in ARM and $18b in Noam Brown / @polynoamial : This is on the scale of the Apollo Program and Manhattan Project when measured as a fraction of GDP. This kind of investment only happens when the science is carefully vetted and people believe it will succeed and be completely transformative. I agree it's the right time. Sam Altman / @sama : @SmokeAwayyy absolutely not! very important and huge partnership, for a long time to come. we just need moar compute. Tanay Jaipuria / @tanayj : Wow! Stargate Project will invest $500B over the next 4 years - that's ~0.4% of US GDP over that period. For comparison, the inflation-adjusted dollars spent on other large undertakings: • Interstate Highway System: ~$650B • Apollo Program: ~$280B • Manhattan Project: ~$35B @kakashiii111 : So, basically, “Stargate” is just a recycled marketing name for a move that was already announced back on June 12, 2024? It's amazing how desperate they are for more catalysts, recycling them over and over. [image] Sam Altman / @sama : build monuments in the desert @smokeawayyy : OpenAI and Microsoft are done.
SpyglassM.G. Siegler
Context & Ripple Effects
Stargate's announcement put a $500 billion US AI-infrastructure commitment at the center of OpenAI's compute strategy, but the coverage highlights the gap between a headline commitment and identifiable capital, power, and build plans. That gap matters because the project is framed around strategic US AI capacity rather than a conventional single-company expansion.
SoftBank, OpenAI, Oracle, and MGX face immediate pressure to clarify which portions of the stated commitment are funded, newly committed, and executable, rather than broad or overlapping plans.
The project's credibility becomes tied to securing financing and power alongside hardware; the reported cash constraints and questions over re-announced projects make the headline total an insufficient measure of near-term capacity.
Second-order effects
Data-center financing and contracted compute become central to execution: partners and lenders will focus on credit structure, utilization, and power availability rather than the aggregate commitment alone.
If major AI buildouts are increasingly assembled through debt, long-term capacity contracts, and multiple partners, AI infrastructure could become more financialized and less comparable to straightforward corporate capex.
The project also tests whether strategic AI ambitions can translate into delivered power and data-center capacity; if not, execution constraints—not announced investment totals—will set the pace of expansion.
The trend: AI competition is shifting from model announcements toward the financing, power procurement, and delivery discipline needed to turn strategic compute commitments into operating capacity.
America just won geopolitics for the next 50 years with Project Stargate. Let me explain why. This is where things get interesting - actually, scratch that, this is where things get revolutionary. We're about to witness something that happens maybe once in a generation: the
Stargate is probably a grift. Not in the sense that $500B will be stolen or anything but just in that it's a commitment to spend money on capex that these corporations have already earmarked, but now presented as some proactive Trumpist initiative. They're racing anyway.
Worth noting that Stargate has been under development for over a year and a spend of $100B/yr was already publicly reported back in April This is a big deal but it isn't new - it was already happening
Under Biden, when companies directly got aid from the IRA or Bipartisan Infrastructure Bill, they often went out of their way to say how they didn't need it, government didn't help them, etc. Elon constantly talked about how the $465 million bailout he received under Obama had
Stargate Project was announced — $500B in CapEx for AI in the US feels like a good thing for US security and for AI's continued progress in capabilities. Google's entire capex in 2024 was $50B. This is a hard amount of money to actually spend. Maybe just 1% of that can go to
important to temper your expectations re: stargate imo. historically masa-san has aimed for a real ROI roughly between negative 2 per cent and positive 1 per cent on investments greater than one billion dollars.
My guess is Stargate will mostly be debt funded... (think Coreweave and most datacenters at this point) At 10x leverage, you only need $10b in equity to fund the first $100b commitment...
My company announces $5B in funding... ...using the new formula for announcements. You announce $500B, then commit to $100B for “this year” and then everyone finds out only ~$10B is actually secured.
Masayoshi Son's Softbank investment in the $500B Stargate is far from the first crazy investment he's done. Masa invested $20M for 30% of Jack Ma's Alibaba in 2000 in a 10 minute meeting at his house. That stake was worth $200B at peak, a 10,000x return.
what's the name when animals in danger start posturing and puffing their feathers etc to make themselves appear bigger/more dangerous when they're losing?
Keep in mind chips & servers are short-lived assets that depreciate quickly; there's a reason accountants calculate useful lives in the 3-5 year range. (This is fundamentally different from long-lived assets like rail & power infrastructure w/ 50+ year useful lives)
PROJECT STARGATE ENERGY ANALYSIS Project Stargate is going to need an enormous amount of power: each of the five Stargates needs about 5 gigawatts (about five million people's worth) of constant electricity for each ~$100b supercomputer project. Let's break it down. Each
Worth noting the use of the term “Stargate” in 2001: A Space Odyssey. It's a portal via which an astronaut takes what seems to be the next evolutionary step: transcending his corporeal form to become a celestial being of energy. It's a singularity of sorts. Buckle up!
We are proud to be working with OpenAI, SoftBank, and MGX on the Stargate Project to secure American leadership in AI, create hundreds of thousands of American jobs, and generate massive economic benefit for the entire world.
Europe needs to wake up NOW: The US races ahead with regulatory flexibility and a huge ($500bn!) investment in AI infrastructure. AI is a horizontal technology that will serve as the basis of all future R&D. Without AI, Europe will fall behind in innovation in every sector.
idk, i feel like people just don't quite get it. this is among the most expensive undertakings in human history. it's a scale mankind has never operated at before. there is literally no parallel
trump: i dont get it sam altman: what dont you get? trump: if you build agi, doesnt that get rid of all jobs? sam altman: no honey...thats not how it works trump: oh. okay. sam altman: trump: sam? sam altman: yes? trump: how does it work? sam altman: well. think
The Stargate Project sounds like the stuff of dystopian nightmares—a U.S. government-announced partnership of megacorporations “to protect the national security of America and its allies” and harness AGI “for the benefit of all of humanity.” Let's maybe take a beat here.
Larry Ellison says the Stargate Project will construct the largest computer ever built which will enable AI to create cancer vaccines, personalized medicine and pandemic prevention [video]
This is definitely a first for me. A SENIOR official in the White House publicly declaring that the president's big announcement from the day prior is kind of a ruse [image]
Stargate is a great name but the $500b is a ridiculous number and no one should take it seriously unless SoftBank is going to sell all of their BABA and ARM. SoftBank has $38b in cash, $142b in debt and generates $3ish billion in FCF per BBG. They own $143b in ARM and $18b in
This is on the scale of the Apollo Program and Manhattan Project when measured as a fraction of GDP. This kind of investment only happens when the science is carefully vetted and people believe it will succeed and be completely transformative. I agree it's the right time.
Wow! Stargate Project will invest $500B over the next 4 years - that's ~0.4% of US GDP over that period. For comparison, the inflation-adjusted dollars spent on other large undertakings: • Interstate Highway System: ~$650B • Apollo Program: ~$280B • Manhattan Project: ~$35B
So, basically, “Stargate” is just a recycled marketing name for a move that was already announced back on June 12, 2024? It's amazing how desperate they are for more catalysts, recycling them over and over. [image]