The world of artificial intelligence has its newest public titan. Anthropic's IPO filing on October 26, 2026, pulled back the curtain on a company experiencing truly explosive growth and spending on a scale that is hard to comprehend.
An Unprecedented Revenue Surge
The numbers revealed
in Anthropic's S-1 filing are staggering. The company reported revenue of $16.3 billion for the first half of 2026 alone, a massive leap from its $4.6 billion total for all of 2025. An amended filing later showed third-quarter revenue hitting nearly $17 billion. This incredible growth is driven primarily by enterprise customers adopting its Claude family of AI models. The company's revenue model is a mix; for the nine months ending September 30, 2026, about 79% of revenue came from consumption-based usage fees, with the remaining 21% from subscriptions. This 'pay-for-intelligence' approach, where customers pay for what they use, has found fertile ground. The company noted that as of September 2026, over 11,900 customers were spending at least $100,000 annually, a more than twelve-fold increase from just 950 a year prior.
The Astronomical Cost of Intelligence
This phenomenal growth comes at a jaw-dropping cost. The AI business is fundamentally a game of massive capital expenditure. Training and running advanced AI models requires immense computational power, and Anthropic's filing lays this bare. The company disclosed contractual commitments for computing and AI infrastructure totaling nearly $900 billion. These obligations include massive deals with cloud providers and hardware companies like Google, Amazon, and Broadcom. In the first half of 2026, compute and infrastructure expenses reached $8.9 billion, representing 55% of revenue. While that's an improvement from 2025, where the same costs were 160% of revenue, the sheer scale of future spending commitments is daunting. These costs are a core feature of the AI industry, where the ongoing expense of running models for users (inference) can far exceed the initial cost of training them.
The Rocky Path to Profitability
For a brief moment, Anthropic proved profitability was possible. The company posted its first-ever operating profit of $974 million in the second quarter of 2026. However, it swung back to an operating loss of $567 million in the third quarter. The reason? The company had to buy expensive, short-term computing capacity to meet overwhelming customer demand, a move that prioritized growth over immediate profit. This highlights the central tension for investors: can the company's hyper-growth outpace its colossal spending? The S-1 filing shows that Anthropic can be profitable at scale, but maintaining it requires a delicate balance between managing compute costs and satisfying a rapidly expanding customer base. The company's ability to turn its long-term, lower-cost compute contracts into consistent profit will be a key metric to watch.
What Investors Will Be Watching
With its IPO expected in November under the ticker 'ANT' on the Nasdaq, investors will be scrutinizing every metric. A primary focus will be on the efficiency of its spending. Revenue per employee has grown substantially, a sign that the company is scaling effectively without a proportional increase in headcount. Another key indicator is net revenue retention, which was an astounding 900%, showing that existing customers are dramatically increasing their spending. However, the market environment is not without risk. A recent tech selloff, spurred by fears of increased regulation around AI and data centers, shows how sensitive the sector is to external pressures. Investors will weigh Anthropic's incredible growth trajectory and strong enterprise footing against the immense financial commitments and potential market volatility. The story is less about whether AI is useful, and more about whether it can be a sustainably profitable business at this scale.
