Google must double AI compute every 6 months to meet demand, AI infrastructure boss tells employees

During a recent all-hands meeting, Amin Vahdat, Google Cloud's AI infrastructure chief, conveyed a clear message to employees: the company must double its computing capacity every six months to keep pace with the surging demand for AI services. This presentation, which was viewed by CNBC, highlighted the urgent need for advancements in AI infrastructure, with Vahdat stating, "Now we must double every 6 months.... the next 1000x in 4-5 years." Vahdat emphasized that the competition surrounding AI infrastructure is not only critical but also one of the most financially demanding aspects of the AI landscape. His comments came shortly after Alphabet reported strong third-quarter results, prompting an upward revision of its capital expenditure forecasts to a staggering $91 billion to $93 billion for the year, along with a significant increase anticipated in 2026. The tech giants, including Microsoft, Amazon, and Meta, have also raised their capital expenditure expectations, with the four companies projected to spend over $380 billion collectively this year. Vahdat clarified that while Google is committed to building robust infrastructure, the goal isn't merely to outspend competitors. Instead, he stressed the importance of developing infrastructure that is "more reliable, more performant, and more scalable than what’s available anywhere else." To bolster its capacity, Google is not only focusing on infrastructure enhancements but also leveraging more efficient AI models and custom silicon. Recently, the company unveiled its seventh-generation Tensor Processing Unit, named Ironwood, which is reported to be nearly 30 times more power-efficient than its first Cloud TPU launched in 2018. Vahdat noted that Google benefits significantly from its collaboration with DeepMind, which is exploring the future of AI model development. Looking ahead, Vahdat asserted that Google must deliver a thousandfold increase in capability while maintaining the same cost and energy consumption. He acknowledged the challenges but expressed confidence that through collaboration and innovative design, this goal is attainable. Alphabet CEO Sundar Pichai echoed the sentiment, noting that 2026 will be a pivotal year for the company amidst fierce AI competition and increasing demand for cloud services. He addressed concerns regarding a potential bubble in the AI market, a topic that has gained traction among investors skeptical of the anticipated massive spending in the coming years. Pichai reassured employees that Google is strategically positioned, citing impressive quarterly growth in its cloud business and a substantial backlog of $155 billion. Despite the positive outlook, Pichai recognized that the road ahead would be filled with challenges, urging employees not to become complacent. He stated, "It's a very competitive moment so, you can't rest on your laurels. We have a lot of hard work ahead but again, I think we are well positioned through this moment." As discussions about a potential AI bubble intensify, Nvidia's CEO dismissed the notion during a recent earnings call, highlighting the strong demand in the sector. Google, meanwhile, is racing to enhance its AI offerings, launching its latest model, Gemini 3, which aims to provide superior answers to complex queries. However, Pichai pointed out that the primary bottleneck remains the availability of computing power, restricting the rollout of advanced tools like the video generation tool Veo. In response to questions regarding capital expenditures and free cash flow, Google’s finance chief, Anat Ashkenazi, emphasized the company's prospects for growth, indicating significant opportunities as more customers transition from physical data centers to the cloud. The message from Google leadership is clear: the tech giant is determined to not only keep pace with the AI revolution but to lead it, despite the challenges ahead.