What happened

Anthropic, the company behind Claude, is reportedly struggling to secure enough cloud capacity. TechCrunch says a leaked S-1 draft reviewed by The Wall Street Journal described trouble getting capacity from AWS, Google, and Microsoft. Those companies supply computing resources, but they also build competing AI products. Anthropic did not immediately answer TechCrunch’s request for comment. TechCrunch

Why Anthropic needs so much capacity

Cloud capacity means rented computing power. It lets a company use large server systems without owning every machine. Anthropic is also leasing data centers and working with outside operators. Yet its reported filing worries that some outside operators have limited operating histories. They may be less dependable than the largest cloud providers.

The report says Anthropic could spend more than $750 billion on future cloud contracts. Some deals are still being negotiated. So the number describes a possible commitment, not a final bill. It also shows how much the company expects its AI services to grow.

Why the supplier problem matters

Normally, a customer wants a supplier to provide more capacity. Here, the supplier may also want to win the same AI customers. That creates a more complicated relationship. Anthropic needs reliable capacity for training and running Claude. AWS, Google, and Microsoft must decide how much capacity to reserve for Anthropic while expanding their own AI businesses.

This does not prove that any provider is refusing Anthropic. It shows a reported risk: a fast-growing AI company depends on companies that have reasons to compete with it. If capacity arrives late, Anthropic could face pressure when demand is rising. If contracts are expensive, growth could require still more money.

What the reports say

The reported draft lists $16.3 billion in revenue for the first half of 2026. It also lists an $80.7 billion net loss. Two customers made up about one-fifth of sales. TechCrunch also reported that third-quarter revenue reached $16.8 billion. That was 45% higher than the previous quarter, based on what Anthropic told potential investors. The company is reportedly aiming for a November IPO. TechCrunch

These figures point in two directions. Demand appears to be growing quickly. The infrastructure needed to support that growth is also becoming a central financial question. The loss and revenue figures do not show the final cost of each cloud deal. They do show that scale and cost are rising together.

What remains uncertain

The S-1 was leaked, so it may change before a public filing. The exact amount already signed is unclear. The final timing and capacity of each AWS, Google, and Microsoft arrangement are also unclear. We do not know whether outside data-center operators can provide the reliability Anthropic wants. Anthropic’s November IPO remains a reported plan, not a completed event.

What to watch next

The next useful evidence will be a formal filing and confirmed contract terms. Watch whether the $750 billion figure shrinks, grows, or separates signed deals from negotiations. Watch whether Anthropic adds dependable capacity outside the three large providers. Investors will also compare revenue growth with infrastructure commitments. The larger lesson is simple: advanced AI depends not only on better models. It also depends on who controls the computers those models need.