The AI company Anthropic is taking the expansion of its infrastructure to a new level. In an extended partnership with Google and Broadcom, several gigawatts of computing power are to be provided in future. The aim is to operate the next generation of Claude models and meet the rapidly increasing demand for AI services. The expansion is to start gradually from 2027 and marks one of the biggest infrastructure steps in the AI sector.
- Several gigawatts of TPU computing power planned for AI models
- Focus on the next generation of Claude AI
- Expansion of infrastructure mainly in the USA
- Partnership strengthens collaboration with Google Cloud and Broadcom
Anthropic partnership with Google and Broadcom drives AI scaling forward
The new Anthropic partnership with Google and Broadcom clearly shows how rapidly the demand for AI computing power is developing. Instead of small, incremental expansions, Anthropic is focusing on a massive expansion on an industrial scale. The planned capacities in the gigawatt range are a clear signal that modern AI models have long been among the largest energy and computing consumers in the tech industry.
Google’s specialized TPU chips, which are optimized for the training and operation of AI models, are at the heart of the collaboration. Broadcom is contributing its expertise in chip development and infrastructure. Together, they are creating a platform that is specifically tailored to the requirements of modern AI systems.
The strategic classification is also interesting: Anthropic is deliberately pursuing a multi-cloud approach. In addition to Google Cloud, the company also continues to use Amazon Web Services and platforms such as Microsoft Azure. This diversification ensures greater flexibility and reliability – a decisive factor for companies that use AI in critical applications.
The scale of the project is considerable. A large part of the new infrastructure is to be built in the USA and follows on from earlier investment plans. Anthropic had already announced in 2025 that it would invest billions in the expansion of its data centers. The new partnership continues this course consistently and takes it to a new level.
Exploding demand for AI drives expansion of data centers
The background to this massive expansion lies in the rapidly growing demand for AI services. Anthropic’s Claude models are increasingly being used in companies – from automation and data analysis to complex decision-making processes. Within a short space of time, the number of large business customers has increased significantly, many of whom invest millions in AI solutions every year.
This development is also reflected in sales growth, which has multiplied within a year. For providers such as Anthropic, this means one thing above all: the existing infrastructure is quickly reaching its limits. Without massive investment in new computing capacities, further growth would hardly be possible.
Another important aspect is technical diversity. Anthropic does not rely exclusively on one hardware platform, but combines different technologies – including TPUs, GPUs and specialized AI chips. This strategy makes it possible to optimally distribute different workloads and maximize efficiency.
At the same time, it is clear that competition in the AI market is increasingly being decided by infrastructure. Those who have access to the most powerful systems can develop more complex models and react more quickly to new requirements. The partnership with Google and Broadcom gives Anthropic a decisive advantage here.
Conclusion
The expanded Anthropic partnership with Google and Broadcom marks a milestone in the expansion of global AI infrastructure. With several gigawatts of planned computing power, the company is setting new standards and responding to the enormous increase in demand. Prices or specific availability play a subordinate role in this context, but the launch from 2027 shows how long-term such projects are designed to be. This step is a clear signal for the future of AI: scaling is becoming a decisive success factor.
