To meet the huge demand for AI, Google wants to double the overall size of its servers every six months. This is a growth rate that will generate 1000 times more capacity over the next four to five years.
The statement was reportedly made by Amin Vahdat, Google’s head of AI infrastructure, during an all-hands meeting on November 6th. CNBC. Google’s parent company Alphabet is certainly doing well, so such a requirement may be within the company’s financial capabilities. The company reported strong third-quarter numbers at the end of October, raising its capital spending forecast from $91 billion to $93 billion.
In response to employee questions about the company’s future amid talk of an “AI bubble,” Vahdat reiterated the risks of not investing aggressively enough. In cloud operations, these infrastructure investments are paying off. “The risk of underinvestment is very high (…) With more computing, the cloud numbers would have been even better.”
Google’s cloud business continues to grow at about 33% annually, creating a revenue stream that allows the company to be “more resilient to failure than other companies,” he said.
We believe that with infrastructure improvements that run more efficient hardware, such as 7th generation Tensor Processing Units and more efficient LLM models, we can continue to create value as enterprise users increasingly implement AI technology.
Extreme Networks’ Markus Nispel, writing for techradar.com in September, said it is IT infrastructure that is undermining enterprises’ AI visions. He blames the failure of AI projects on the high demands that AI workloads place on legacy systems, the need for real-time and edge capabilities (often lacking in today’s enterprises), and the continued existence of data silos. “Even when projects get off the ground, they are often hampered by delays caused by poor data availability or system fragmentation. Without clean, real-time data flowing freely throughout the organization, AI models cannot operate effectively, and the insights they generate arrive too late or are ineffective,” he said.
“With 80% of AI projects struggling to meet expectations around the world, primarily due to infrastructure limitations rather than the AI technology itself, what matters now is how we respond.”
His views are shared by decision makers at major technology providers. Capital spending by Google, Microsoft, Amazon, and Meta is expected to exceed $380 billion this year, with much of it focused on AI infrastructure.
The message from hyperscalers is clear. If we build it, they will come.
Addressing the infrastructure challenges that organizations experience is a key component to the successful implementation of AI-based projects. Agile infrastructure as close as possible to the compute point and integrated datasets are seen as key parts of the recipe for extracting maximum value from next-generation AI projects.
While we expect some degree of market consolidation across the AI space in the next six months, companies like Google are among those expected to consolidate in the market and continue to deliver innovative technology based on advances in AI.
(Image source: “Construction Site” by tomavim is licensed under CC BY-NC 2.0.)
Want to learn more about AI and big data from industry leaders? Check out the AI & Big Data Expo in Amsterdam, California, and London. This comprehensive event is part of TechEx and co-located with other major technology events. Click here for more information.
AI News is brought to you by TechForge Media. Learn about other upcoming enterprise technology events and webinars.
