Can Anthropic Beat SpaceX’s IPO Valuation?

In June, the SpaceX [SPCX] market debut raised the vast sum of $1.77tn, making it the largest IPO in history. But that record could be broken by artificial intelligence (AI) lab Anthropic, which is reportedly aiming to go public some time towards the end of November.

Anthropic’s IPO prospectus leaked at the end of September. Aside from revealing that the company believes its technology poses “existential risks to humanity”, the document contains some very striking numbers.

For instance, it notes that in 2025 the firm racked up $12.65bn in operating expenses, more than half of which went to compute. All in, it logged a net loss of $42bn last year, although some $34bn of that was a non-cash accounting charge.

That’s an impressive total, but it pales in comparison to the $518bn Anthropic says it has committed to infrastructure obligations over the coming years.

At least $111.1bn will go to Alphabet’s [GOOGL] Google, $110bn to Amazon [AMZN] and $31.4bn to Microsoft [MSFT], under long-term infrastructure service obligations. In addition, there are some $161.2bn of equipment lease obligations to Broadcom [AVGO]. Of that total, 80% is non-cancellable or requires payment regardless of usage.

The firm saw its revenue multiply twelvefold in 2025, climbing to $4.6bn. Nearly a quarter of that came from just two clients, which the prospectus did not name. The Information recently reported that both Meta [META] and Microsoft are scaling back their use of Anthropic’s Claude Code in favour of in-house systems.

So, can Anthropic best SpaceX’s $1.77tn?

It’s a definite possibility.

The AI lab is reportedly targeting a valuation of $1.8tn to $2tn, which would indeed top the satellite firm’s IPO. And there is no shortage of hype. Investors are excited by the prospect of finally getting a pure-play stake in one of the companies actually building the AI models behind the current boom.

At present, investors can gain exposure through AI suppliers, but chipmakers and data-centre operators offer more indirect exposure, while the big hyperscalers are sufficiently diversified that AI represents only one part of their investment case. This is why supposedly AI-focused ETFs can’t do much more than put together baskets of the same mega-cap technology stocks.

In that sense, Anthropic’s IPO would offer something the market has lacked: direct exposure to a leading frontier AI developer.

However, investors would also be making an enormous bet on the technology’s ability to justify the spending required to build it. Anthropic’s infrastructure commitments alone dwarf its current revenues, while the company remains loss-making. The bull case therefore rests on AI becoming sufficiently valuable – and Anthropic capturing enough of that value – to generate returns on a level that can ultimately support such a valuation. And that is not a straightforward proposition.

Further, SpaceX was able to command an extraordinary valuation partly because it offered exposure to several powerful themes at once, from satellite communications to space infrastructure and AI. Anthropic’s proposition is narrower, and its capital requirements are considerably more daunting.

Leaving aside the inherent value of Anthropic itself, there is a broader macro angle that needs to be taken into account.

The IPO market seems to be souring

After a strong start to the year for US IPOs, rising bond yields, rate hike fears, geopolitical tensions and growing concerns over the sustainability of AI spending have all begun clouding sentiment.

At the end of September, smart ring maker Oura postponed its US IPO. It had planned to sell 50m shares at $40 to $44 each, potentially raising $2.2bn and giving it a fully diluted valuation of $15.62bn. Its offering had been expected to provide a key test of appetite for high-growth technology companies following the summer quiet period.

Oura is not alone. Nuclear services company Holtec suspended its planned IPO earlier in September, while Bamboo Insurance also delayed its listing. Most relevantly, Anthropic’s major rival, OpenAI, recently announced it was also postponing its own IPO, purportedly over safety concerns.

A cooling IPO market does not necessarily kill the deal. AI remains one of the strongest investment themes in global markets. But Anthropic may have to convince investors that its growth prospects justify accepting an unusually large amount of financial and execution risk.

Ultimately, the success of the IPO will depend on how excited investors are about it. There’s a lot that can happen before end-November. And there are a number of concerning developments in the offing.

For instance, the Financial Times recently noted that insurers are currently prepping for multimillion-dollar claims related to AI agents going rogue and hacking companies and organisations. There are fears that CEOs such as Anthropic’s Dario Amodei (pictured) or OpenAI’s Sam Altman could be held responsible.

Elsewhere, on 5 October, as part of the largest chip-financing deal ever, Wall Street banks began offloading a $60bn debt package to fund Anthropic’s lease of Google semiconductors. If appetite for this debt is limited, that might have an impact on the firm’s eventual IPO price.

Is there a sense that Anthropic is rushing to IPO before these and other worries truly manifest?

In short, Anthropic’s IPO could be the biggest test yet of whether investors remain willing to finance the extraordinary capital requirements of frontier AI, particularly as the broader IPO market becomes less accommodating. A $2tn valuation would make history, but Anthropic will need to convince investors that AI’s future economic value can justify the price.

CMC Aureon’s proprietary theme relevance system maps the world’s biggest investing megatrends. For in-depth analyses of stocks with high growth potential, subscribe to CMC Aureon Foresight.