When Berkshire Hathaway revealed a new multibillion-dollar stake in Alphabet, most observers initially viewed it as another major addition to Warren Buffett’s long-term portfolio. Yet the timing and strategic context suggest the investment could be more than a routine allocation. The world’s most influential value investor has chosen to back Alphabet at a moment when the company is strengthening its control over the hardware that underpins advanced AI systems.
Google’s AI chips may be shifting the balance of power
For nearly a decade, Alphabet has been developing its own artificial intelligence chips, known as Tensor Processing Units (TPUs). In November 2025, the company announced general availability of its seventh-generation Ironwood TPU, describing it as designed for demanding AI workloads, including large-scale model training and high-volume inference. This announcement was made publicly on Google Cloud’s official blog, where the company highlighted lower costs and improved performance for AI developers.Most AI labs still rely heavily on Nvidia’s high-performance GPUs. These GPUs remain the industry’s gold standard, but they are extremely expensive and often in short supply. This has created bottlenecks for AI companies that depend entirely on Nvidia for compute capacity.Google is in a unique position. By designing its own chips, running its own data centres and integrating its hardware into products such as Gemini and Google Cloud, Alphabet gains levels of efficiency that many competitors cannot match without years of hardware investment. Analysts and industry researchers have suggested that TPU-based training may offer meaningful cost advantages compared to traditional GPU-heavy infrastructure, although exact ratios vary and are not always publicly disclosed.Buffett’s investment therefore can be understood not as a claim that Alphabet has overtaken Nvidia, but as confidence in Alphabet’s vertically integrated approach to AI. By controlling more of the stack, Alphabet potentially insulates itself from the price swings, supply constraints and dependence that characterise the GPU market.
A shifting competitive dynamic, but not a dethroning
Nvidia remains the clear leader in AI chips and maintains a robust ecosystem of developers, tools and software support. Google’s TPUs represent a serious alternative in some workloads, but not a full replacement for GPUs across all use cases.The publicly available evidence shows:
- Google trains and serves Gemini models on TPU infrastructure.
- Google Cloud is offering Ironwood TPUs to customers as part of its cloud platform.
- Companies such as Anthropic have announced plans to use up to one million Google TPUs, citing performance and cost considerations.
These developments suggest that the AI compute market may become more competitive. However, they do not confirm that Nvidia’s dominance is definitively weakened. Instead, they indicate that Alphabet is emerging as a strong hardware contender in a market previously dominated by a single supplier.
What Buffett’s investment likely signals
Berkshire Hathaway’s disclosed stake in Alphabet was valued at approximately US$4.3 billion at the time of the filing. The filing does not state Berkshire’s motives. However, investors and analysts widely agree that Buffett’s approval of such a large position indicates confidence in Alphabet’s long-term strategy, including its growing emphasis on efficient AI infrastructure.Buffett has consistently favoured companies with durable advantages or “moats.” Alphabet’s traditional moats have included Search, YouTube and Android. In the AI era, Google’s control over its own chip design and cloud hardware may represent an additional moat that differentiates it from rivals.This does not imply that Google is attempting to dismantle Nvidia’s leadership. Instead, it suggests that Alphabet is building the capability to operate more independently in an AI-driven future, and that this capability may strengthen its competitiveness in cloud computing and AI services.
Implications for the AI chip race
The next phase of AI is likely to be shaped by which companies can train and deploy models most efficiently. If Google Cloud attracts more third-party training due to cost, performance and availability, it could gradually shift cloud market dynamics. Nvidia will remain central to the ecosystem, but Google’s TPU ecosystem adds a new dimension to the competition.Buffett’s investment amplifies the narrative that companies with deeper control over hardware, software and data will hold significant advantages in the years ahead. Alphabet is signalling that it intends to be such a company.The takeaway is clear: this is not the end of Nvidia’s dominance, but it may be the beginning of a more competitive era in AI hardware.