Google Invests $40 Billion in an “Enemy”: Allies of OpenAI Have Started Renewing Subscriptions for Backup Options

On April 25th, Anthropic announced that Google would invest up to $40 billion in it - an initial injection of $10 billion in cash, with a valuation of $380 billion. An additional $30 billion would be added if performance milestones were reached. Google has its own Gemini, which is in direct competition with Claude in the global market. They are fighting on one hand and investing on the other.

It seems contradictory. But in today's AI arms race, some analysts believe that this is one of Google's most strategically valuable decisions at present.

The question is, what does this move mean for OpenAI?

An Anthropic financing list clarifies one thing

If you list Anthropic's financing records from the past six months, you'll find an intriguing reality:

▸ Amazon: $5 billion in cash, with an upper limit of $25 billion, plus 5 gigawatts of Trainium computing power (Source: Sina Finance, April 2026)

▸ Google: $10 billion in cash, with an upper limit of $40 billion, plus 5 gigawatts of TPU computing power, to be launched starting in 2027 (Source: Sina Finance, April 25, 2026)

▸ NVIDIA: An upper limit of $10 billion and 1 gigawatt of GPU supply (Source: Wall Street Insights, April 2026)

▸ Microsoft: An upper limit of $5 billion, and Anthropic will purchase $30 billion worth of computing power from Azure (Source: 36Kr, April 2026)

▸ Anthropic's annual recurring revenue (ARR): Over $30 billion (only $9 billion at the end of 2025) (Source: Sina Finance, April 25, 2026)

▸ Cumulative computing power commitment: Over 11 gigawatts, equivalent to the power generation of about 10 nuclear power plants (Source: Wall Street Insights, April 2026)

Four top players - Amazon, Google, NVIDIA, and Microsoft - all appear on Anthropic's shareholder roster. Among them, Microsoft is also OpenAI's biggest ally.

Many industry analysts judge that the landscape has evolved from the "Big Three" (OpenAI, Google, and Anthropic standing in a tripartite confrontation) to a two - power standoff between the Anthropic camp and OpenAI. The term "Big Three" is becoming history.

Why does Google want to nurture a competitor?

▍Logic 1: If computing power isn't sold, it becomes the most expensive inventory

Google's capital expenditure plan for 2026 is as high as $185 billion, with a large amount of funds being poured into data centers and TPU production capacity. The problem is that if there are no major customers to absorb the TPUs, they will become extremely expensive idle assets. Anthropic is the most ideal "ballast stone" for Google's TPU production capacity - a single agreement locks in a long - term demand of 5 gigawatts, which means Google has found a major buyer willing to sign a multi - year contract for its chips.

▍Logic 2: The investment is a defense to prevent Anthropic from fully aligning with Amazon

Amazon started to layout Anthropic earlier and with greater intensity. If Google doesn't follow up, the Claude ecosystem will be deeply bound to AWS, and the competitive space for Google Cloud will be significantly narrowed. Some analysts view this $40 - billion investment as a "strategic defense fee" - in return, it gets Anthropic's long - term procurement obligation for TPUs and the right to collaborative development in the Google Cloud ecosystem.

▍Logic 3: Instead of suppressing, it's better to bind with capital

Google knows internally that in the general large - model race, no single company will dominate in the short term. Instead of spending energy on suppressing Anthropic, it's better to bind it with capital and let the other party's computing power consumption become its own revenue. Competing on one hand and benefiting from the other party's growth on the other - this is the unique business logic in the AI landscape of this era.

OpenAI's current situation is more difficult than it seems

OpenAI's core computing power comes from the Stargate project, which is deeply bound to NVIDIA. The target scale is $500 billion, and full production is expected to be achieved around 2029. However, the physical progress of the first Texas data center is still slow.

The Anthropic camp has committed more than 11 gigawatts of computing power, and a large part comes from the ASIC route (Google TPU + Amazon Trainium), rather than NVIDIA GPUs. Two completely different technological routes are being advanced simultaneously. Whichever forms a scale first will gain the pricing power - some believe that this industry game will be decided in the next 3 to 5 years.

In the view of some industry observers, the biggest challenge for OpenAI is not that Anthropic surpasses GPT - 4o in a certain benchmark, but that its allies are quietly renewing their support for alternatives. When the "loyalty" of the ecosystem begins to waver, some believe that the connotation of the original competitive barriers is undergoing profound changes.

What does this mean for the A - share market?

From the perspective of the A - share market, the most notable aspect of the Google - Anthropic agreement is the impact of the "divergence of two technological routes" on the domestic computing power industry chain.

The core of the ASIC route (TPU/Trainium) lies in customized chip design, while the core of the NVIDIA GPU route lies in the ecological software stack. The domestic computing power industry chain currently leans towards the NVIDIA GPU ecosystem. However, if the ASIC route is verified on a large scale under the impetus of Anthropic, the valuation logic of the customized AI chip track will be affected - this has tracking and reference significance for relevant A - share sectors. Anthropic may launch an IPO as early as October this year, which will be another important observation window. This article does not constitute investment advice. Please refer to research reports from professional institutions for specific investment targets.

The Benchmark score doesn't determine the outcome of the AI industry. The three - line game of cash, computing power, and ecosystem is the real arena for this reshuffle.

At the moment when Google invested $40 billion, the outline of the first round of the evolution of the current AI landscape has become clear. But this is not the end, but the starting point for the next round of the game.

【Disclaimer】This article is only for information sharing and industry analysis and does not constitute any investment advice, investment analysis opinion, or trading invitation. The market is risky, and investment should be made with caution. Any investment decision made based on the content of this article will result in risks and profits or losses being borne by the investor himself. The author and the publishing platform will not assume any legal liability.

Information Sources

1. Sina Finance: "Google Plans to Invest Up to $40 Billion in Anthropic" (April 25, 2026)

2. Wall Street Insights: "Google Spends $40 Billion to Lock in Anthropic, Ending the Era of the Big Three" (April 25, 2026)

3. 36Kr: "Google Invests $4 Billion in Anthropic, and Computing Power Becomes the Weight in the Capital Game of AI Giants" (April 25, 2026)

4. Cailian Press: "Google Allegedly Plans to Invest Up to $40 Billion to Deepen Cooperation with Anthropic" (April 25, 2026)

This article is from the WeChat official account "BT Finance" (ID: btcjv1), author: Shuyan. It is published by 36Kr with authorization.