Google has announced the need for a massive increase in its artificial intelligence capabilities to meet the rapid growth in demand for related services. According to Google’s management, the AI infrastructure must double every six months, which implies a thousandfold increase over the next 4-5 years.
This is reported by Business • Media
Scaling Challenges: Chip and Energy Shortages
Google Cloud Vice President Amin Vahdat reported at an employee meeting that the company is facing serious constraints in chip supply, energy resources, and data storage network capabilities. He stated that scaling the infrastructure must occur without a significant increase in costs and energy consumption. Vahdat emphasized that Google aims to create an infrastructure that is “more reliable, productive, and scalable,” while also avoiding a sharp rise in the operational costs of data centers.
“The company needs to scale another 1000 times in 4-5 years,” emphasized Amin Vahdat.
With the increasing demand for AI technologies, Google, like other market players, faces the question of whether this growth is driven by genuine user activity or is a result of integrating AI into the company’s existing services.
Focus on In-House Development and Optimization
OpenAI is experiencing similar challenges, investing over $400 billion in building six new data centers in the U.S. as part of the Stargate project. Another significant hurdle remains the global shortage of Nvidia graphics processors, which are essential for training and deploying AI models. According to Nvidia, the demand for their chips exceeds supply, which is already impacting the implementation of Google’s latest solutions.
CEO Sundar Pichai cites the example of the Veo tool, which could not be scaled for a large number of users due to technical limitations. To reduce dependence on Nvidia, Google is actively developing its own technologies: in November, the company introduced seventh-generation TPU processors that are nearly 30 times more efficient than previous models.
Additionally, Google plans to optimize the architecture of its AI models and expand the physical infrastructure of its data centers. Pichai warns that 2026 will be particularly challenging due to competition and the need to meet demand for cloud services, while discussions about the “overheating” of the AI market remain relevant.