Warren Buffett's Berkshire Hathaway has made a bold move, investing $5.1 billion in Google (Alphabet), signaling a significant shift in the AI chip landscape. This investment comes after Google launched Ironwood, a chip that rivals Nvidia's GPUs at a fraction of the cost, and Gemini 3, the most advanced AI model trained without Nvidia chips.
Google's Ironwood chip is a game-changer, offering computational power at 1/5 the cost of Nvidia's GPUs. This development has major implications for AI companies, which currently pay $3-4 billion to train frontier models. Google's cost advantage, paying only $600-750 million, could disrupt the market.
The Rise of Google's AI Chip
Google's custom-built Tensor Processing Units (TPUs) are designed for demanding AI workloads, including large-scale model training and high-volume inference. The company's seventh-generation Ironwood TPU is now available on Google Cloud, offering lower costs and improved performance .
Nvidia's Dominance Challenged
Nvidia's $3.5 trillion empire is facing a significant challenge. With Google's Ironwood chip and TPUs, AI labs are switching to more cost-effective options. Anthropic has already ordered 1 million Google TPUs, and OpenAI is rumored to be following suit .
Buffett's Signal
Warren Buffett's investment in Google is a vote of confidence in the company's AI strategy. Buffett, known for his value investing approach, sees Google's control over its chip design and cloud hardware as a significant advantage.
The Future of AI Chips
The AI chip market is becoming increasingly competitive, with Google emerging as a strong contender. As Google Cloud grows, Nvidia's pricing power may collapse. The future of AI belongs to whoever owns the chips, and Google is making a strong bid for dominance.
