Anthropic’s IPO filing highlights the commercial growth behind AI infrastructure

Anthropic’s proposed stock-market listing offers a detailed view of the commercial growth behind the AI economy, and the computing infrastructure required to support it.

The company behind Claude filed its public IPO registration statement on 26 October, applying to list on Nasdaq under the symbol ANT. First-half revenue reached US$16.3bn, compared with US$1.2bn a year earlier. Its compute and infrastructure expenses were US$8.9bn, exceeding the US$7.3bn recorded across the whole of 2025.

Behind that expansion are processors, data centres, power, cooling and the engineering needed to deliver AI services reliably. As adoption grows, the infrastructure opportunity encompasses both developing increasingly capable models and running them for customers every day.

Growth across the major cloud platforms

Anthropic’s expansion sits within a broader pattern of cloud growth.

In the quarter ended June 2026, Microsoft’s Azure and other cloud services revenue increased 43%. Azure also surpassed US$100bn in annual revenue for the first time. Management said customer demand continued to exceed available capacity, with additional capacity delivered during the quarter quickly generating revenue.

Over the same period, Amazon Web Services revenue grew 37% to US$42.2bn, while Google Cloud revenue increased 82% to US$24.8bn. Alphabet identified demand for enterprise AI infrastructure and solutions among the drivers of its growth.

These cloud businesses support workloads beyond AI, but their results provide important context: substantial revenue growth is occurring alongside the continuing expansion of computing capacity.

AI is strengthening established businesses

Meta illustrates another route through which AI creates commercial value.

For the quarter ended September 2026, Meta reported US$64.6bn in revenue, up 26%. Management said AI was delivering substantial additional advertising revenue through stronger engagement, better recommendations and more effective monetisation. Meta does not separately quantify that AI contribution, but it explicitly links the technology to improvements in its core business.

The company invested US$34.9bn in capital expenditure during the quarter and forecast US$135bn–US$145bn for the full year. Its investment plans support both established services and emerging AI initiatives.

This broadens the infrastructure story. Demand is supported by companies selling AI products and by businesses using AI to improve services that already generate significant revenue.

Computing capacity is becoming a substantial revenue stream

SpaceX’s results provide a further example of the commercial scale emerging around AI infrastructure.

Its Super Intelligence segment generated US$7.56bn in revenue in the September quarter, representing approximately 52% of the group’s US$14.6bn total revenue. Within that segment, US$7.02bn came from solutions and infrastructure, with the remainder from advertising. Cloud services were a principal driver of growth, alongside subscriptions and the contribution from the acquired Cursor business.

The segment combines infrastructure, software and other services. Its results nevertheless demonstrate that supplying computing capacity is already contributing materially to revenue, alongside the development and sale of AI applications.

An opportunity extending beyond the hyperscalers

Anthropic’s filing also illustrates the role of specialist providers. It discloses approximately US$41.3bn in commitments to Lambda, over terms extending into 2034. Fees accrue as capacity is delivered and accepted, subject to contractual conditions.

For specialist infrastructure businesses, the commercial challenge is to turn demand into dependable capacity at competitive cost. Access to processors is one part of that task. Power availability, suitable locations, deployment speed, cooling and operational expertise are also essential.

Understanding the IPO’s financial picture

Anthropic’s first-half operating loss narrowed to US$693m, from approximately US$2.1bn. Its US$80.7bn net loss included a substantial US$72.4bn fair-value charge relating to financing instruments, so the headline loss is not equivalent to cash consumed.

Media reports have suggested a potential IPO valuation of US$1.5tn–US$2tn. The initial prospectus leaves the offering price and share count unspecified.

The Bolt perspective

For Bolt, these developments strengthen the commercial rationale for focusing on the infrastructure behind AI.

The evidence spans several business models: frontier laboratories selling intelligent software, cloud providers supplying computing capacity, and established platforms using AI to improve customer engagement and revenue generation.

CUDO operates within this wider ecosystem, with a focus on the sustainable computing infrastructure needed to train and run AI. Our interest centres on the practical requirements behind deployment: power, location, performance, reliability and operating economics.

Growing demand creates opportunity. Building lasting value requires disciplined execution and infrastructure that meets customers’ needs efficiently. Anthropic’s filing, viewed alongside the performance of the wider sector, gives that opportunity a more substantial commercial foundation.