About to become the first profitable AI lab: Anthropic‘s growth story
Key Takeaways
- Anthropic is projected to achieve an operating profit of $559 million in Q2, marking a significant turnaround from previous expectations.
- The main drivers for this profitability are the rapid adoption of Anthropic’s coding tools and the utilization of its AI assistant Claude in agentic tasks.
- However, the growth has also led to increased costs due to the need for more computational resources, prompting Anthropic to sign new data center deals with entities like SpaceX.
Anthropic is on track to become the first AI lab to achieve profitability. The company reported a revenue of $10.9 billion in Q2, marking a 130% increase from the previous quarter. This growth trajectory places Anthropic ahead of notable tech companies like Zoom during the pandemic and Google/Facebook before their respective IPOs.
The primary driver behind this profitability is the widespread adoption of Anthropic’s coding tools by businesses worldwide. Additionally, “agentic” tasks involving Claude have been utilized more extensively, leading to higher demand that sometimes outstripped available compute resources. This necessitated measures such as throttling access and signing new data center deals.
AI usage is getting pricier, and Anthropic benefits
The increase in costs associated with agentic AI tasks has been driven by a new tokenizer that splits text into more tokens. For example, an 80-round session of Claude interaction could see cost increases ranging from 20 to 30%. The situation is similar for Anthropic’s flagship model, Opus 4.7, which now costs the same per token as its predecessor but at a higher rate due to increased complexity.
OpenAI has also implemented aggressive pricing strategies: GPT-5.5’s list price doubled compared to GPT-5.4, with input tokens now priced at $5 and output tokens at $30 per million. This move led to significant real-world cost increases for users, ranging from 12% to 27% for prompts over 2,000 tokens.
Despite these challenges, Anthropic’s efficiency is improving. In Q1, the company spent 71 cents per dollar of revenue on compute. By Q2, this figure is expected to drop to 56 cents. CEO Dario Amodei humorously noted that revenue growth had become a challenge due to the need for more computational resources.
AI News Without the Hype – Curated by Humans
Subscribe to THE DECODER for ad-free reading, a weekly AI newsletter, our exclusive “AI Radar” frontier report six times a year, full archive access, and access to our comment section.
Originally published at the-decoder.com. Curated by AI Maestro.
Stay ahead of AI. Get the most important stories delivered to your inbox — no spam, no noise.
