https://www.ft.com/content/ae9bb47d-bd1d-473c-b4c5-abae0420cc12?syn-25a6b1a6=1 Elon Musk, Sam Altman and Dario Amodei once sparred over how OpenAI could catch Google in AI.Now, the trio are preparing blockbuster flotations that could mint trillion-dollar companies, revive moribund markets for US initial public offerings and test the ability of public investors to absorb a tidal wave of tech listings.The rivalry — sharpened by Musk and Amodei’s acrimonious departures from OpenAI in 2018 and 2020 — has also set up a contest over which company can command the deepest pool of capital.Amodei’s Anthropic, Musk’s SpaceX and Altman’s OpenAI could make 2026 the biggest year for US IPOs. Share sales from all three companies would push such fundraising well beyond the record $156bn raised in 2021, ending four lean years for venture investors and Wall Street bankers.Investors, bankers and advisers involved in the offerings said executives were weighing how much capital they could raise across the three deals without overstretching the market.
Weak debuts could also undermine the AI optimism that has helped propel US equities higher despite inflation and geopolitical turmoil.All three private companies moved closer to listing this week, as leaks about OpenAI’s accelerated IPO timetable and Anthropic’s first quarterly profit competed for attention with SpaceX’s long-awaited S-1 filing.The leaks signal to public market investors that the AI labs could follow SpaceX in short order.Rob Hilmer, founder of Goanna Capital, which has a position in all three companies, said it was no coincidence they were all preparing to list at the same time.“In a strong risk-on environment, which we’ve been in for a while, how many large-scale tech IPOs would we expect after five years of not many?” Backers believe the companies can ride a wave of AI enthusiasm among institutional and retail investors alike. SpaceX is aiming to raise about $75bn at a $1.75tn valuation, while OpenAI was recently valued at $852bn and Anthropic is close to sealing a $30bn funding round at a $900bn price tag.“There is almost $8tn in money market funds today,” said another investor in all three companies. “Absorbing SpaceX’s [expected $75bn] float is just 1 per cent of that. There’s cash on the sidelines ready to put to work.”Public investors have spent years “trying to get exposure to AI in derivative ways, particularly via semis [semiconductor stocks such as Nvidia]. As soon as they can get access to the labs they will own directly. It won’t be them that has a challenge, it will permeate to the rest of the market,” they added.
Peter Hébert, co-founder of venture capital firm Lux Capital, was also confident public markets could absorb the massive floats. “A $75bn mega raise by SpaceX would not even be the largest primary capital raise announced since the start of 2026, that record is held by OpenAI,” he said, referring to the ChatGPT maker’s $122bn funding round earlier this year. “In a world looking for growth, there is a lack of those assets in public markets today...I think they’ll be extremely well received,” he added.Backers believe the companies can ride a wave of AI enthusiasm among institutional and retail investors alike. SpaceX is aiming to raise about $75bn at a $1.75tn valuation, while OpenAI was recently valued at $852bn and Anthropic is close to sealing a $30bn funding round at a $900bn price tag. But the trio remain heavily lossmaking and public investors may prove less tolerant of vast cash burn and unfunded commitments than private backers have been.At the $1.75tn valuation bankers are targeting, SpaceX would trade at 91 times its $19bn of revenues over the past year, a multiple that eclipses the priciest of its Big Tech peers. Nvidia, the most expensive of the Magnificent 7 stocks on a revenue basis, trades at 21 times trailing sales and is massively profitable.To pull off the SpaceX flotation, Musk needs investors to buy into his vision as well as the numbers. The prospectus reads as much like a manifesto as a financial disclosure: 14 of its first 16 pages are dominated by images of rockets, satellites and planets.
But the trio remain heavily lossmaking and public investors may prove less tolerant of vast cash burn and unfunded commitments than private backers have been.At the $1.75tn valuation bankers are targeting, SpaceX would trade at 91 times its $19bn of revenues over the past year, a multiple that eclipses the priciest of its Big Tech peers. Nvidia, the most expensive of the Magnificent 7 stocks on a revenue basis, trades at 21 times trailing sales and is massively profitable.To pull off the SpaceX flotation, Musk needs investors to buy into his vision as well as the numbers. The prospectus reads as much like a manifesto as a financial disclosure: 14 of its first 16 pages are dominated by images of rockets, satellites and planets. He noted the fully reusable rockets SpaceX was building would enable the whole business, from satellite-beamed internet and cell service on Earth to data centres in space.Anthropic’s profitability may prove fleeting as spending increases. The group this month signed on to become SpaceX’s biggest customer, agreeing to spend $15bn a year on data centre capacity and computing power after also making commitments worth hundreds of billions more in deals with Google and Amazon.OpenAI’s outlay is more ambitious. The company booked almost $6bn in revenue last quarter, driven by ChatGPT and growing use of its coding tool Codex, said a person with knowledge of the matter.But it has told investors it expects to burn through about $600bn before turning profitable in 2030. It has raised more than any start-up in history, churning through Big Tech and sovereign backers, and is looking to public investors to extend its runway.
Part of OpenAI’s pitch is that it will be first to achieve artificial general intelligence — the vaguely defined point at which AI surpasses human capability — and that the rewards will dwarf today’s spending.Grand visions that play well in private markets do not always survive contact with public ones.WeWork’s mission to ‘elevate the world’s consciousness’ by renting offices with ping-pong tables and beer taps was lapped up by private investors including SoftBank’s Masayoshi Son, but given short shrift by public funds, leading the company to pull its $47bn listing in 2019.That episode haunts some private investors.But others are more sanguine: comparison between the doomed office company and today’s AI darlings “conflates a big company that was a terrible business with three of the best-quality companies ever. These are well run, high-growth businesses,” said Goanna’s Hilmer.“The last thing I’m concerned with is whether there is capital to invest in them.”
>>1516732>without overstretching the marketlmao
>>1516736>data centres in space.But why though? What advantages does a space data center have?
>>1516732Is Google really ahead? It doesn’t feel that way aside from maybe the resilience of their servers.
>>1516794I think some analysts count them as "ahead" only because of how google has integrated gemini into their search engine, which they can then claim gets used more often than the others.
>>1516792Cooling. Space is cold.Power. Solar panels are even stronger without the atmosphere in the way.Security. Us proles can't attack them as the dystopia becomes more and more clear.
>>1516802>Cooling. Space is cold.It’s harder to radiate heat
>>1516802How long would that last before the GPUs become obsolete and the station keeping thrusters run out of fuel? 5 years? It still doesn't seem economically viable, especially when any X-class solar flare could knock it offline and/or out of orbit.
>>1516807I guess?>>1516808>It still doesn't seem economically viableI don't think so either, I'm just explaining the benefits.
>>1516740It's going to be crash the market because the Nasdaq and S&P are changing their rules to make a special exception for these three companies. This is basically going to be a wall street AI bailout.
>>1516802>prolesThe elites don't call you a prole, they call you goyim, as proven by the Epstein files
Hope U.S bros didn't expect their index fund to be worth anything by the end of 2026.
>>1516802Exactly the opposite. AI data centres put out enough heat to visibly raise the local temperature for hundreds of metres when they have the benefit of massive water cooling systems and an atmosphere to radiate into.In space? Nada. Thermoses use vacuum layers to insulate their contents for a reason. That heat takes a long-ass time to leave, there's simply not enough free mass for there to be sufficient coolant, and on top of that direct exposure to the sun WILL raise surface temperatures by hundreds of degrees. Any sort of data centre in space would need a radiator farm the size of a small city to adequately cool the cunt.And for what? For Joe Retard to futz his paperwork again?
>>1516918>AI data centres put out enough heat to visibly raise the local temperature for hundreds of metres when they have the benefit of massive water cooling systems and an atmosphere to radiate into.We should probably stop building these things.
>>1516924>My city has been in a 6 year legal battle just to tear down an abandoned race track and build condos over environment concerns>2.5 mile data center near the lake got approved last year in private and the city will be subsidizing water and electricity for it>Was approved by 6 people and no one was allowed to object or debate it, which is illegal
