Google Bets $40B on Anthropic: Is Google Cloud the New Growth Engine as Meta Challenges Ad Dominance?

Anthropic's need for Google's computing power stems from a compute shortage that has already begun to hinder its growth. Due to these constraints, its Claude service has recently suffered from persistent outages, rate limiting, and performance issues, even prompting some customers to switch to other AI platforms and creating business pressure for Anthropic.

According to reports on Monday, Anthropic has committed to Amazon over the next decade (AMZN) Web Services (AWS) to purchase more than $100 billion in computing resources and will utilize up to 5 gigawatts of Amazon AI chip capacity; Amazon is also investing an additional $5 billion in Anthropic, a move seen as a response to the compute crisis.

This latest procurement of computing power from Google further underscores the severity of Anthropic's compute shortage. Google's TPUs are among the most competitive rivals to Nvidia's (NVDA) chips and feature lower costs than Nvidia's, making them the ideal choice for Anthropic.

For Google, why spend $40 billion to invest directly in its large model competitor Anthropic, when it already owns DeepMind, an AI-specialized subsidiary, the high-performing Gemini large model, and the world's largest TPU cluster?

On one hand, OpenAI, Google, and Anthropic have formed a tripartite rivalry in the large model space. This massive investment essentially brings Anthropic under Google's wing, hedging its bets on the future of large models: if Anthropic eventually emerges as the winner among the three, Google will still benefit.

Another competitor, OpenAI, has already formed an alliance with various tech giants, establishing a Microsoft (MSFT) -NVIDIA-Oracle (ORCL) -OpenAI symbiotic chain. However, Google struggles to compete on its own; its partnership with Anthropic aims to leverage external strength to dominate the large model sector.

In more specialized segments, Gemini’s developer share in the enterprise AI market has consistently lagged behind Claude, which remains superior in coding logic. The Claude Code agent, launched in 2025, is a command-line interface (CLI) tool designed specifically for programmers and is considered the highest-quality code generation tool currently available.

Furthermore, even setting aside the intense competition in large models, Anthropic becoming a TPU customer not only helps bolster its chip shipments but also provides a third-party endorsement for its hardware, thereby enhancing Google’s standing in the chip market.

Currently, Google's core revenue source remains advertising, but its dominance is gradually eroding. The latest forecast from research firm eMarketer suggests that Meta (META) is poised to overtake Google in digital advertising revenue for the first time in 2026, becoming the world's largest digital advertising company. Meta's full-year net advertising revenue is projected to reach $243.46 billion, accounting for 26.8% of global advertising spend; Google is projected at $239.54 billion, or 26.4%. Although it is only a slight edge, it signifies that Google's status as the global advertising hegemon is slipping.

In recent years, Google has been betting heavily on its cloud services as a second growth engine. By the end of 2025, Google Cloud's market share in the global cloud infrastructure services market had risen to 14%, trailing Amazon AWS's 28% and Microsoft Azure's 21%. Despite being the fastest-growing player—with Q4 2025 revenue increasing by a staggering 48% year-over-year, far outstripping its competitors—its original base is smaller, and it will still take time to catch up to AWS and Azure.

Given the expectation that advertising revenue could be surpassed at any time, Google Cloud urgently needs a customer like Anthropic, whose massive and stable demand can bolster its revenue.

This content was translated using AI and reviewed for clarity. It is for informational purposes only.