Anthropic aims to list at a valuation above $2 trillion, exposing Gulf sovereign backing and a gap between its revenue of $4.6 billion and $7.33 billion in costs

Anthropic is preparing to list its shares on public markets at a valuation that could exceed two trillion dollars, in a move that shifts the race to develop foundational AI models from closed private-funding rooms to direct financial scrutiny by investors. This shift not only puts the company’s rapid commercial growth rate under the microscope, but also highlights the financing base backed by sovereign Gulf capital that has accompanied its major expansion phases, alongside the costly operating and infrastructure expenses imposed by competition standards at the technology frontier.

The company’s financial statements reveal a sharp widening of the gap between revenue and computing costs; Anthropic’s revenue in 2025 jumped roughly twelvefold to about $4.6 billion, compared with the prior year. In contrast,Computing and infrastructure costs rose to $7.33 billion.This, equivalent to three times the previous year’s spending, represents more than half of total operating expenses of $12.65 billion. This expansion pushed the operating loss to $8.06 billion, compared with roughly $2.98 billion in 2024, while the recorded net loss reached about $42 billion, $34 billion of which stemmed from accounting adjustments linked to changes in the value of financing instruments convertible into equity.

The forward-looking figures in the prospectus show cloud infrastructure and computing commitments of $518 billion over the coming years, with roughly 80 percent of those commitments classified as non-cancellable or payable contracts regardless of actual usage. These massive financial burdens are accompanied by a clear revenue concentration risk, as nearly a quarter of the company’s 2025 income came from just two customers, alongside official warnings that some major clients may not honor long-term contracts, and internal research that observed unexpected behaviours of autonomous systems in experimental test environments.

This listing represents a pivotal moment for Gulf sovereign investors who have been a core pillar in private-funding rounds; Abu Dhabi’s MGI and Qatar Investment Authority have bolstered their presence in successive financing rounds that accompanied the private valuation’s rise toward the trillion-dollar threshold. Qatar Investment Authority began its investment in September 2025, then participated in the “G”-class round in February 2026, and the “H”-class round in May 2026 that raised $65 billion and valued the company at roughly $965 billion. MGI followed a parallel path, leading the “G”-class round with $30 billion at a $380 billion valuation, then joining the May financing round, linking sovereign capital to the upper layers of model architecture, alongside complementary infrastructure deals such as MGI’s acquisition of Aligned Data Centers for about $40 billion.

This shift directly affects technology officers and institutions across the Gulf and the Arab East. Anthropic’s move to public trading will subject the pricing of inference services and model access to quarterly profitability pressures and open-market calculations, after a period of support through venture-capital rounds. Technology leaders, banks and government bodies in the region, which build their digital infrastructure on the company’s models, will need to prepare for possible changes in pricing structures and reduced contractual flexibility as the company seeks to meet computing commitments exceeding half a trillion dollars. Public-market pricing will also give regional sovereign funds a clear daily benchmark for assessing their stakes in the advanced AI race.

The upcoming public offering will be a decisive test of how markets value the next phase of AI developmentThis comes as competition accelerates with OpenAI and AI operations at Google, Meta and XAI, while Amazon and Google continue to provide capital and cloud infrastructure. The challenge is no longer limited to the technical superiority of models, but to how developer companies can convert growing demand into operating profits that cover the massive computing-infrastructure cost.