80x. That's how much Anthropic grew in the first quarter of 2026 on an annualized basis, according to CEO Dario Amodei speaking at the company's developer conference in San Francisco.
The number explains something Claude users have been noticing: capacity problems. Amodei said Anthropic had planned for roughly 10x growth heading into Q1 - already an aggressive forecast - but actual demand came in eight times higher than that projection. The result was a company that couldn't buy and deploy enough compute (the GPU hardware that runs AI models) fast enough to keep pace. As CNBC reported, Amodei was direct about the mismatch.
The admission is unusually candid. Most AI companies talk around capacity problems with vague language about "high demand." Naming the exact multiple gives users a clearer picture of why Claude Code and other Claude products have sometimes felt unreliable - it's not a model problem, it's a physical hardware bottleneck.
GPU supply chains are constrained across the industry, and being an 80x-growth company in a hardware-limited environment is a genuine operational problem. Chips take months to order, configure, and bring online. Forecasting at 10x and getting 80x means that planning gap shows up as rate limits, slower responses, and degraded availability for paying customers. Claude Code users running long agent sessions - which require sustained, uninterrupted compute - have felt this most acutely.
The bigger question now is whether Anthropic's infrastructure investments can catch up to actual demand, or whether the gap between supply and growth keeps widening through the rest of 2026.