By Manya Saini and Echo Wang
Oct 26 (Reuters) – Anthropic disclosed steep losses and red-hot revenue growth in its US IPO filing on Monday, as the Claude-parent laid out its ambition to build AI systems capable of reshaping the global economy despite the high costs of developing the technology.
The prospectus highlights the AI company’s rapid growth over the past year, fueled by Big Tech’s breakneck spending that has driven a rally in AI stocks while raising concerns about lofty valuations and rising debt.
The company also flagged growing risks from AI technology, underscoring concerns over its potential to cause harm which had led CEO Dario Amodei to call for a slowdown in development of advanced models.
Anthropic’s net loss in the first half of 2026 widened sharply to $80.7 billion, largely due to a rise in the fair value of bonds convertible into company’s shares. Revenue in this period stood at about $16.3 billion.
The company’s revenue grew a staggering 1,368% in the three months ended June 30 from a year ago, but warned in the prospectus that revenue growth rate could decline over time and fluctuate from time to time.
The IPO, which is likely to be as early as later this month, is expected to establish Anthropic, the AI lab behind the viral Claude models, as a benchmark for how Wall Street values leading AI companies, including rival OpenAI.
The IPO push comes despite a once-busy fall listing calendar thinning out as choppy markets prompt companies to delay their offerings. Analysts said that would give Anthropic access to a deep pool of capital to fund its costly AI ambitions.
Growing investor uneasy with the debt-fueled AI investment boom has led to a global selloff in AI and tech stocks as they start to doubt if the trillions spent on data centers, chips and computing capacity will generate fast returns to support lofty valuations. Higher borrowing costs and rising bond yields have added to their caution.
However, tech companies are still hungry for cash. Meta is aiming to raise up to $95 billion in its largest-ever equity fundraising, while SpaceX has inked a $40 billion deal with Apollo to lease AI chips and server hardware.
VALUATION EXPECTATIONS
Some current and prospective investors are valuing Anthropic at about $1.5 trillion, down from roughly $2 trillion a week ago, after the AI startup disclosed some financial details at a recent investor event, Reuters reported.
“While Anthropic is well-capitalized, an IPO would give the company some valuable financial flexibility. No company wants to spend too long on the cusp of going public, since that can be a distraction,” said Matt Kennedy, senior strategist at Renaissance Capital, a provider of IPO-focused research and ETFs
“Maybe they could have gotten a higher valuation next year, but they’ve decided to “rip the band-aid off”.
The filing confirmed a series of Reuters exclusives that offered an early peek at its contents before Anthropic made it public, including details of its financial performance, growth trajectory and IPO plans.
BALANCING AI’S PROMISES AND PERILS
Anthropic said more than half of Fortune 500 companies and over 30% of Global 2000 firms were customers generating at least $100,000 each in revenue for the company over the 12 months ended September 30.
The IPO comes as Anthropic’s own research raises concerns about advanced AI systems, which in controlled tests have shown behaviors including code sabotage, fraud assistance and information manipulation.
The findings have fueled concerns about how companies can keep powerful AI systems under control as they race to deploy them commercially.
The prospectus also highlighted risks associated with its AI models, which it said could exhibit “self-preserving behaviors”, including behavior “resembling blackmail”.
“Our increasing use of AI systems to accelerate our own research and development may cause the capabilities of our models to advance more rapidly than our ability to develop, validate, and implement corresponding safety, alignment, and security measures,” it said.
The company said it relies on a limited number of cloud services providers such as Amazon, Google and Microsoft and suppliers such as Broadcom for compute and critical parts of its infrastructure These relationships can create “dependencies and potential vulnerabilities”, it said.
Anthropic will trade on the Nasdaq after the IPO closes under the ticker symbol “ANT”.
The filing did not reveal the terms of its offering, including the number of shares it plans to sell and the proposed price range.
RED HOT GROWTH
Anthropic, founded in 2021 by a team of researchers, was valued at $965 billion in May in a funding round that drew some of the biggest names in Silicon Valley and Wall Street.
Amazon and Google were among Anthropic’s early strategic partners, investing billions of dollars in the startup while providing the cloud infrastructure needed to train and deploy Claude models.
“We believe this abundance of AI will rapidly transform every sector of the global economy in ways greater than the impact of industrialization, electricity, semiconductors and the internet,” the prospectus said.
The company said it expects compute and infrastructure expenses to increase in absolute dollars and remain its largest operating expense for the foreseeable future.
Anthropic remains locked in a fierce battle with OpenAI to lead the race to build the most advanced AI systems. The ChatGPT maker confidentially filed for an IPO in June and is expected to list by early 2027.
Anthropic’s IPO follows that of Elon Musk’s SpaceX, which rewrote the IPO record books in June, when it priced a $75 billion offering that valued the company at $1.75 trillion.
Goldman Sachs and Morgan Stanley were named as the lead banks for the IPO alongside 20 others as underwriters. Morgan Stanley will act as a stabilization agent, overseeing the stock’s early trading.
The roles mark another major win for Goldman and Morgan Stanley as they had led SpaceX’s IPO too.
(Reporting by Manya Saini in Bengaluru and Echo Wang in New York; Additional reporting by Pritam Biswas and Utkarsh Shetti; Editing by Saumyadeb Chakrabarty, Sayantani Ghosh and Arun Koyyur)
Comments