Google is betting big on the future of artificial intelligence (AI) as it plans to increase its computing capabilities by 1,000 times within the next 4-5 years.
This ambition was revealed in a report by Consumer News and Business Channel (CNBC), taking place in the context of both Silicon Valley and Wall Street debating whether the current AI explosion is showing signs of a new bubble.
Amin Vahdat, VP of Google Cloud, said at an internal meeting in early November that the company needs to double its computing capabilities every six months to meet the huge demand for AI products.
Mr. Amin Vahdat emphasized that Google must be able to provide 1,000 times more computing, networking and storage capabilities than it currently has while keeping costs and energy at the same level.
According to Mr. Vahdat, the AI infrastructure competition is becoming the most expensive but decisive stage in the entire AI race.
At this meeting, Alphabet CEO Sundar Pichai and CFO Anat Ashkenazi were also present, emphasizing the strategy of building a more reliable AI infrastructure that can expand better than any competitor.
Instead of increasing huge spending like Microsoft or Amazon, Google aims to optimize architecture, co- design hardware and software for long-term efficiency.
However, growth ambitions come at a time when the market is concerned about the possibility of an AI bubble forming.
The four technology giants Google, Microsoft, Amazon and Meta are expected to spend more than 380 billion USD on AI infrastructure in 2024 alone.
Alphabet alone has raised its spending forecast for the year to $9193 billion twice and continues to increase significantly in 2026.
However, according to Pichai's analysis, strong investment is still a must if you want to avoid falling behind as the demand for AI skyrockets.
Mr. Sundar Pichai gave an example of the Veo video creation tool, a product that caused a fever when it was launched but could not be widely deployed due to lack of computing capabilities.
Answering the question about the possibility of an AI bubble breaking, the CEO said that the risk is real, but not investing enough will be even more dangerous.
Along with expanding its infrastructure, Google also focuses on improving performance by optimizing models and developing specialized silicon chips.
Last week, the company announced that the 7th generation TPU, Ironwood, could save more than 30 times more electricity than the first TPU in 2018.
Meanwhile, the latest model Gemini 3 is designed to handle more complex questions, increasing direct competition with GPT-4 and other advanced models.
With these strategies, Google shows that it is willing to invest long-term to maintain its position in the AI race, even when the market faces many unpredictable fluctuations.
