HSBC Says Anthropic Deal Could Add Azure Revenue

HSBC estimates that Microsoft's partnership with Anthropic could generate up to $43 billion in additional annual Azure revenue by 2030. Seeking Alpha first reported the projection, attributing it to HSBC analyst Stephen Bersey. The figure is a sell-side estimate; neither Microsoft nor Anthropic has endorsed it.

Bersey's model is a chain of assumptions rather than a disclosed plan. It assumes Anthropic grows to roughly $241 billion in revenue by 2030, up from less than $5 billion in 2025; that compute spending reaches about 60% of revenue, implying a cloud opportunity near $144 billion across all providers; and that Microsoft Azure captures about 30% of that spend, in line with HSBC's broader Azure share assumptions. Multiply those together and the result is the headline ~$43 billion annual figure. The estimate is anchored to a concrete deal: in November 2025 Microsoft committed to invest up to $5 billion in Anthropic, and Anthropic agreed to purchase $30 billion of Azure compute capacity.

The projection is useful for market-sizing discussions because it makes the bull case for AI-driven cloud revenue explicit and quantified. But its sensitivity is the point: small changes in any single assumption, Anthropic's revenue trajectory, the compute-to-revenue ratio, or Azure's share, swing the dollar outcome by tens of billions. Comparable third-party AI cloud forecasts over the past two years have shown wide variance for exactly this reason.

Headline revenue projections ultimately resolve into infrastructure questions. Whether usage converts into platform revenue depends on the unglamorous levers practitioners control: model-serving volume, accelerator capacity, managed-service pricing, and the cost controls and observability that keep large inference workloads economical. Those are the variables that determine if a $241-billion-revenue scenario is even physically and financially serviceable.

• •Disclosed Anthropic-on-Azure consumption or enterprise adoption metrics, which would replace assumptions with data.
• •Pricing and revenue-share terms that determine how much of Anthropic's compute spend lands as Azure gross bookings.
• •Capacity signals from chipmakers and cloud infrastructure teams indicating whether supply can meet the projected growth.