Dario Amodei: Anthropic's Actual Growth Far Exceeds Plans, Q1 Revenue and Usage Annualized Growth of 80 Times
Anthropic CEO Dario Amodei stated that the company originally planned resources for a 10x annual growth, but in the first quarter of 2026, revenue and usage achieved an explosive annualized growth of 80 times.
This unexpected growth rate has directly led to a severe shortage of computing power, which is the core reason for the current tightness in the company's computing resources.
Source: Public Information
ABAB AI Insight
Dario Amodei has previously stated that Anthropic planned infrastructure based on a "10x/year" long-term growth model. The Q1 80x annualized growth significantly exceeds expectations, continuing the explosive demand for the Claude series in enterprise deployment and API calls. The company has already responded urgently by signing large long-term computing power contracts with Google, Amazon, and others.
In terms of capital strategy, Anthropic is investing all available funds and revenue cash flow into securing computing power, rapidly expanding capacity through significant prepayments to multiple cloud providers (Google 5GW, Amazon $100 billion, etc.). This high growth also validates the penetration speed of Claude in the enterprise market, creating a positive feedback loop of "demand superlinear explosion → computing power crisis → massive capital prepositioning."
Similar to the computing power bottleneck experienced by OpenAI during the GPT-4/o1 period and the rapid deployment of xAI Colossus, leading AI labs are currently in a phase of superlinear expansion from "planned growth" to "demand completely out of control."
Essentially, this represents capital concentration: Anthropic's 80x growth will highly concentrate global computing resources into a few leading model companies, reconstructing capital from dispersed cloud consumption into exclusive infrastructure for specific AI labs through massive prepayment contracts. Mechanically, this forces the industry to complete computing power deployment at a pace far exceeding plans due to the unexpectedly high commercial adoption rate, accelerating the structural transition of AI from an experimental tool to a large-scale productivity platform.
ABAB News · Cognitive Law
When actual growth is 8 times the plan, the computing power shortage is not a problem but a proof of success. The best crisis is one where growth is so fast that it breaks all budgets and plans. Truly leading companies are always paying heavily for "computing power that is never enough."