Google and Amazon Simultaneously Invest in Rivals: AI Business Logic Reversed

Introduction: Big Tech's 'Fund Your Enemy' Game

In the traditional business world, no company would willingly hand billions of dollars to a competitor. But in the AI era, this is happening in real time — and more than one company is doing it.

Google has invested a cumulative $2 billion-plus in Anthropic, while Amazon went even further, committing $4 billion to become Anthropic's largest external investor. Meanwhile, Microsoft's $13 billion deep partnership with OpenAI is already well known. The three cloud computing giants have almost simultaneously chosen the seemingly absurd path of paying to nurture their competitors.

This is neither charity nor a blunder — it is the AI era rewriting the fundamental rules of business competition.

The Core: Why Giants Willingly Bankroll Their Rivals

To understand this phenomenon, one must first grasp the value chain structure of the AI industry. Today's AI industry is roughly divided into three layers: the bottom layer is computing infrastructure (chips, data centers, cloud services); the middle layer is foundation models (R&D and training of large-scale models); and the top layer is applications (AI products for end users).

The core revenue source for Google, Amazon, and Microsoft is cloud computing services. For them, AI model companies are "super customers" — training and deploying large models requires massive computing power, and that computing power is purchased through cloud platforms. Amazon's investment agreement with Anthropic explicitly requires Anthropic to use AWS as its primary cloud service provider. Google's investment comes with similar cloud service lock-in provisions.

In other words, a significant portion of the money these giants invest flows back into their own pockets in the form of cloud computing bills. It is a carefully designed capital loop: investing equals customer acquisition; subsidizing equals lock-in.

The deeper logic lies in ecosystem control. In the AI era, whichever cloud platform runs the most and strongest foundation models attracts the most developers and enterprise customers. Prosperity at the model layer directly drives growth at the infrastructure layer. The giants don't need to personally win every model race — they just need to ensure that whoever wins is running on their cloud.

Analysis: Diverging Evolution of U.S. and Chinese AI Infrastructure

This "invest in your competitor" model is essentially a unique product of the American tech ecosystem. In the United States, AI infrastructure is forming a symbiotic structure of "cloud platforms + independent model companies." Cloud giants provide the computing foundation, independent AI companies focus on model innovation, and the two are deeply bound through capital ties and commercial agreements, creating a complex relationship of simultaneous competition and interdependence.

Notably, U.S. and Chinese AI infrastructure are exhibiting different evolutionary tendencies.

In China, AI infrastructure development tends toward "vertical integration." Companies like Baidu, Alibaba, and Huawei prefer to build out the entire chain — from chip adaptation and cloud services to foundation models and end-user applications. The advantage of this model lies in high coordination efficiency and strong self-reliance, especially against the backdrop of restricted external chip supply, where vertical integration has become almost an inevitable choice.

The U.S. "horizontal specialization" model, by contrast, emphasizes specialization and capital efficiency. Companies like Anthropic and OpenAI can own zero physical infrastructure yet gain access to world-class computing resources through fundraising. This model has spawned more startups and competing technical approaches, but it also raises concerns about "giants steering innovation through capital control."

Each model has its pros and cons. The Chinese model demonstrates greater resilience under geopolitical pressure but may sacrifice a degree of innovation diversity. The American model is more dynamic in sparking breakthrough innovation, but the ecosystem's stability is highly dependent on capital market confidence and the delicate balance among giants.

Reflection: Risks and Paradoxes of the 'Fund Your Enemy' Logic

This "fund your enemy" game is not without risks. First, there is the pressure of antitrust scrutiny. The U.S. Federal Trade Commission (FTC) has already launched investigations into the investment relationships between Microsoft and OpenAI, and between Amazon and Anthropic, examining whether these deals constitute de facto merger control.

Second, there is the uncertainty of commercial returns. Large model companies are currently operating at massive losses, with both Anthropic and OpenAI incurring annual operating costs measured in the billions of dollars. If AI commercialization progresses slower than expected, these investments could face significant write-downs.

The more fundamental paradox is this: when your competitor becomes powerful enough, will it still be content to remain your "tenant"? OpenAI is already exploring the possibility of building its own data centers, and Anthropic is seeking more diversified cloud service partnerships. The moats that giants have built with capital may not be as solid as they imagine.

Outlook: The Next Turning Point in the AI Race

From the vantage point of mid-2025, the competitive landscape of the AI industry is entering a critical inflection period. In the short term, the "invest for lock-in" model will persist and may even intensify — more model startups will become targets in the giants' tug-of-war.

In the medium to long term, however, two variables could change the rules of the game. First is the rise of open-source models: Meta's Llama series and open-source forces like DeepSeek are eroding the bargaining power of closed-source model companies. Second is the diversification of AI-specific chips: as more chip alternatives emerge, the scarcity of computing power may gradually ease, thereby reducing cloud platforms' leverage.

Regardless, the fact that Google and Amazon are simultaneously nurturing competitors has already become one of the most iconic business phenomena of the AI era. It reminds us that in waves of technological transformation, the boundaries between competition and cooperation are never fixed. The true winners are often those who can redefine the rules of the game.