Just now, a $2.6 trillion AI unicorn was born. NVIDIA and Microsoft are making bets, and Elon Musk is getting anxious.

On February 13th, Zhidx reported that early this morning, AI unicorn Anthropic announced the completion of its Series G financing of $30 billion (approximately RMB 207.019 billion). After the investment, its valuation reached $380 billion (approximately RMB 26.2 trillion). The company's annual recurring revenue has reached $14 billion (approximately RMB 69.909 billion), with an average annual growth of over 10 times in the past three years.

▲ Anthropic's announcement (Source: X)

According to Anthropic's announcement, the funds from this financing will support cutting - edge research, product development, and infrastructure expansion. Currently, Claude is the only AI model available on the three major global cloud platforms: Amazon Web Services Bedrock, Google Cloud Vertex AI, and Microsoft Azure Foundry.

As soon as the news was released, Elon Musk immediately came out to question that Anthropic's Claude model involves issues of racial and gender discrimination.

▲ Elon Musk expressed his doubts (Source: X)

Some netizens commented below, saying, "Please, Elon, focus on Grok." Musk replied, "4.2 will be better."

▲ Musk's reply to a netizen's comment (Source: X)

The announcement stated that this round of financing of Anthropic was co - led by the Ontario Teachers' Pension Plan and Coatue. The co - leading investors also include D. E. Shaw Ventures, Dragoneer, Founders Fund, ICONIQ, and MGX. Major investors in this round include Blackstone Group, the growth equity division of Goldman Sachs Alternative Investments, JPMorgan Chase, Sequoia Capital, Temasek, etc.

This round of financing also includes the investments from NVIDIA and Microsoft announced in November 2025. That is, NVIDIA and Microsoft promised to invest $10 billion (approximately RMB 69.006 billion) and $5 billion (approximately RMB 34.503 billion) in Anthropic respectively. (Anthropic Receives Hundreds of Billions in Investments from Microsoft and NVIDIA! Valuation Soars to 2.5 Trillion)

01.

Over 500 Customers Spend Millions of Dollars Annually

The Top 10 Fortune Companies Rush to Use It

The announcement revealed that in terms of annual recurring revenue, the number of Claude customers spending over $100,000 (approximately RMB 690,100) annually increased 7 times in the past year. Enterprises that started using Claude from single - use - case APIs, Claude Code, or Claude for Work are expanding its integration across the entire organization.

Two years ago, only a dozen or so Claude customers spent over a million dollars annually. Now, this number has exceeded 500, and 8 of the top 10 Fortune companies are now Claude customers.

Since Claude Code was fully launched in May 2025, its annual recurring revenue has now exceeded $2.5 billion (approximately RMB 17.252 billion), doubling its growth since the beginning of 2026.

The number of weekly active users of Claude Code has also doubled since January 1st, 2026. Recent analysis shows that 4% of the code in global public GitHub submissions is generated by Claude Code, which is 2 times the proportion a month ago.

The number of enterprise subscriptions for Claude Code has also increased 4 times since the beginning of 2026, and the contribution from enterprise use has exceeded half of the total revenue.

02.

Opus 4.6 Launches with a One - Million - Token Context Window

Claude Code Attracts Developers

In the past year, Anthropic has achieved widely recognized results in both model and product aspects.

Anthropic's Claude series of models have been continuously iterated, upgrading from Version 3.7 to the latest Version 4.5. Its programming ability and Agent ability have been significantly improved, enabling it to complete complex tasks more end - to - end and over a long period in productivity scenarios.

On February 6th, 2026, Anthropic released its flagship model, Claude Opus 4.6, which is Anthropic's first flagship - level model to launch the test function of a one - million - token context window. This model has improved in long - context query, reasoning, planning, and other capabilities. (Claude Opus 4.6 Dominates Programming Competitions! Discovers 500 0 - day Vulnerabilities, Generates K - line Volume Distributions, and Even Outputs PPTs Directly)

▲ Tweet about the release of Claude Opus 4.6 (Source: X)

In its specialized programming field, although Claude is not always ranked first on the list, it has a good reputation among developers.

On the product side, Anthropic has created a hit product, Claude Code. In February 2025, Claude Code was released together with Claude 3.7 Sonnet and quickly became a hit among developers.

As an AI programming assistant that can automatically execute engineering tasks, read and write code libraries, and run tests, Claude Code uses a command - line interface familiar to developers. It has also continuously improved in context persistence and Agent ability, with its user experience constantly optimized.

▲ Claude Code (Source: Anthropic)

In January this year, Anthropic also launched a graphical UI version of Claude Code - Cowork. Now, more non - technical personnel can also use the Agent ability of Claude Code to complete non - programming tasks such as organizing files and scraping web pages.

In addition, Anthropic is also trying to define the open ecosystem of the Agent track. In October 2025, Anthropic introduced the concept of Agent Skills and officially released it as an open standard in December of the same year, aiming to enable agents to have modular capabilities similar to "skill packages". These skills can be encapsulated, reused, shared, and combined.

Agent Skills are essentially a folder - like structure composed of instructions, resources, and optional scripts. Claude will automatically identify and load relevant skills as needed according to the current task, thus achieving repeatable and combinable professional task execution capabilities, rather than being driven by prompt words every time.

In addition, Anthropic is preparing for an IPO, which may be carried out as early as 2026. Anthropic has hired a law firm to carry out preliminary preparations and has had preliminary contacts with several investment banks regarding the listing.

03.

Conclusion: Anthropic's Commercialization Process

Hits the Accelerator

Different from OpenAI's To C approach, Anthropic focuses more on the To B market.

Anthropic has achieved an annual revenue of $14 billion, with an average annual growth of over 10 times in three years. Its product, Claude Code, has contributed over $2.5 billion in annual revenue. These achievements indicate that Anthropic is advancing its commercialization process at an amazing speed, transforming its technological leadership into a market advantage.

Anthropic's diversified layout at the infrastructure level spans AWS Trainium, Google TPU, and NVIDIA GPUs, providing enterprise customers with infrastructure support that combines performance and resilience. This is the moat that Anthropic has built in the fierce competition.

This article is from the WeChat official account “Zhidx” (ID: zhidxcom). Author: Wang Han, Editor: Bing Qian. Republished by 36Kr with authorization.