Anthropic is heading toward one of the most closely watched technology listings in recent years, but its IPO prospectus has revealed a financial picture that is far more complicated than its explosive revenue growth might suggest.
The artificial intelligence company reportedly recorded a net loss of $42 billion in 2025, even as annual revenue surged to approximately $4.6 billion. Despite the enormous loss, Anthropic’s expected valuation at listing is projected to exceed $2 trillion, according to a Reuters report published on September 28.
The disclosures highlight the extraordinary economics of building frontier AI models: revenue is growing rapidly, but so are the computing, infrastructure and financing costs required to support that expansion.
Anthropic Revenue Surges 12-Fold
Anthropic’s 2025 revenue reached roughly $4.6 billion, representing a twelvefold increase from the previous year.
The growth underscores the enormous commercial demand for generative AI products as businesses increasingly integrate advanced models into software, research, customer service and other operations.
Yet the company’s costs have risen almost as quickly.
Anthropic’s operating loss widened substantially, reaching $8.06 billion in 2025, compared with $2.98 billion in 2024.
The difference between the operating loss and the much larger reported net loss is largely explained by accounting adjustments related to the company’s financing arrangements.
Why Anthropic Reported a $42 Billion Net Loss
Approximately $34 billion of Anthropic’s net loss was attributed to an accounting expense associated with an increase in the fair value of financing instruments that could eventually be converted into shares.
In other words, the enormous headline loss does not represent $42 billion in cash simply flowing out of the business during the year.
A substantial portion reflects how certain financing instruments are valued for accounting purposes as the company’s expected value changes.
That distinction is important when assessing Anthropic’s financial position. Its reported net loss is extremely large, but the accounting component needs to be separated from the company’s underlying operating expenses and cash requirements.
The prospectus nevertheless demonstrates that Anthropic is operating at a scale where both its actual spending and accounting obligations have become enormous.
AI Computing Costs Are Exploding
One of the clearest signs of the financial pressure surrounding frontier AI is Anthropic’s spending on computing and infrastructure.
The company spent approximately $7.33 billion on computing and infrastructure in 2025, around three times the amount it spent a year earlier.
That figure represented more than half of Anthropic’s total operating costs of approximately $12.65 billion.
The spending reflects the extraordinary computational requirements involved in training and operating increasingly sophisticated AI models.
Unlike conventional software companies, AI developers must spend heavily on specialised computing resources, data-centre capacity and cloud infrastructure. Those costs can continue to rise as models become larger, more capable and more widely used.
And Anthropic’s commitments extend far beyond its spending in a single financial year.
$518 Billion in Future Infrastructure Commitments
According to the prospectus, Anthropic has committed to approximately $518 billion in future cloud, computing and infrastructure spending.
The scale of that commitment illustrates the company’s expectations for continued growth in AI demand, while also highlighting one of the biggest financial risks facing the industry.
AI companies require enormous computing capacity to develop and operate their models. Securing that capacity can involve long-term agreements with cloud and infrastructure providers, creating substantial future financial obligations.
At the end of 2025, Anthropic reported $20.28 billion in cash, cash equivalents and short-term investments.
That cash position provides significant liquidity, but it is dwarfed by the company’s disclosed future infrastructure commitments.
Anthropic Is Betting on an Economic Transformation
The company’s ambitious financial plans are based on an equally ambitious view of artificial intelligence.
Anthropic has argued that AI could transform the global economy on a scale comparable to, or greater than, major technological shifts such as industrialisation, the spread of electricity and the emergence of the internet.
That thesis is central to the company’s long-term growth strategy and helps explain why investors may be willing to place such a high valuation on a business that is still reporting substantial losses.
However, Anthropic has also identified significant risks in its prospectus.
One concern is customer concentration. Approximately one-quarter of the company’s revenue came from two customers, meaning a relatively small number of major clients account for a significant portion of its business.
The company also noted that many of its largest customers are not locked into long-term contracts. Those customers could therefore reduce their spending or stop using Anthropic’s services altogether.
$2 Trillion Valuation Would Mark a Dramatic Repricing
Perhaps the most striking element of the IPO story is the valuation investors are reportedly considering.
Anthropic’s expected valuation at listing is projected to exceed $2 trillion, more than double the company’s own estimated valuation of $96.5 billion in May.
Such a dramatic increase would reflect the extraordinary investor expectations surrounding the future of AI.
It would also place Anthropic among the world’s most highly valued technology companies despite its comparatively modest current revenue and substantial operating losses.
The timing of the listing, however, remains uncertain. The IPO is reportedly likely to be delayed until after the U.S. midterm elections in November.
Anthropic has declined to comment on the reports.
AI IPO Market Faces a Tougher Test
Anthropic’s potential listing comes at a time when investors are becoming more cautious about high-growth AI and semiconductor companies.
AI-related stocks have recently faced pressure, raising questions about how public markets will value companies whose future growth depends on enormous investments in computing infrastructure.
The performance of other highly valued technology companies could also influence investor sentiment.
SpaceX, for example, listed in June at a valuation of approximately $1.77 trillion. After rising 19 per cent on its first trading day, its shares subsequently fell from around $160 to approximately $147.
The performance illustrates the challenge facing companies entering public markets at exceptionally high valuations: strong initial investor enthusiasm does not necessarily translate into sustained gains.
OpenAI Could Become Anthropic’s Biggest IPO Rival
Anthropic is not the only major AI company preparing for a potential public-market debut.
Rival OpenAI reportedly filed confidentially for an IPO in June, with a listing potentially coming in early 2027.
That creates the possibility of a closely watched race between two of the world’s most prominent AI companies.
There is already speculation that whichever company reaches the public market first could establish an important valuation benchmark for the broader AI industry.
Investors would gain a publicly traded reference point for assessing the value of companies whose economics are dominated by AI models, enormous computing requirements and rapidly expanding commercial demand.
AI Safety Remains Part of Anthropic’s Story
Anthropic’s IPO disclosures also come against the backdrop of its research into AI safety and model behaviour.
The company has previously published findings indicating that AI models with greater autonomy could display unexpected behaviours in controlled experimental environments. These scenarios have included activities such as code sabotage, assistance with fraudulent activity and manipulation of information.
The findings highlight a tension at the heart of the AI industry: companies are racing to develop increasingly capable systems while simultaneously trying to understand and manage the risks associated with greater autonomy.
Anthropic CEO Dario Amodei has previously called for a slower pace of new feature releases, reflecting concerns about the potential consequences of deploying increasingly capable AI systems too quickly.
At the same time, the competitive pressure within the industry remains intense. Anthropic recently released its Opus 5.5 model as it competes with OpenAI and its latest-generation AI systems.
Anthropic’s IPO Will Put AI Economics Under the Microscope
Anthropic’s prospective IPO is shaping up to be more than another major technology listing.
Its prospectus lays bare the unusual economics of frontier artificial intelligence: revenue can grow at extraordinary rates, yet the infrastructure required to generate that revenue can consume billions of dollars.
A $4.6 billion revenue figure, a $42 billion reported net loss, $7.33 billion in annual computing and infrastructure spending and $518 billion in future infrastructure commitments together illustrate the scale of the financial bet being made on AI.
The company’s expected valuation above $2 trillion shows that investors are being asked to value Anthropic primarily on expectations for future AI growth rather than its current profitability.
Whether public markets ultimately accept that valuation will make Anthropic’s IPO an important test for the entire AI sector. With OpenAI also preparing for a potential listing, the two companies could help define how investors value the next generation of AI businesses—and how much they are willing to pay for growth before profitability arrives.
