In order to meet the massive demand for AI, Google wants to double the overall size of its servers every six months, a growth rate that would create a 1000x greater capacity in the next four or five years.
The statement came from the head of Google’s AI infrastructure, Amin Vahdat, during an all-hands meeting on November 6, according to CNBC. Alphabet, Google’s parent company is certainly performing well, so such a requirement may be within its financial capabilities. It reported good Q3 figures at the end of October, and has raised its capital expenditure forecast to $93 billion, up from $91 billion.
Vahdat addressed one employee’s question about the company’s future amid talk of an ‘AI bubble’ by re-stating the risks of not investing aggressively enough. In its cloud operations, such investment in infrastructure has paid off. “The risk of under-investing is pretty high […] the cloud numbers would have been much better if we had more compute.”
Google’s cloud business continues to grow at around a 33% per year, creating an income stream that enables the company to be “better positioned to withstand misses than other companies,” he said.
With better infrastructure running more efficient hardware such as the seventh-gen Tensor Processing Unit and more efficient LLM models, Google is confident that it can continue to create value for its enterprise users’ increased implementation of AI technologies.
According to Markus Nispel of Extreme Networks, writing on techradar.com in September, it’s IT infrastructure that’s making companies’ AI vision falter. He places the blame for any failure of AI projects on the high demands AI workloads place on legacy systems, the need for real-time and edge facilities (often lacking in current enterprises), and the continuing presence of data silos. “Even when projects do launch, they’re often hampered by delays caused by poor data availability or fragmented systems. If clean, real-time data can’t flow freely across the organisation, AI models can’t operate effectively, and the insights they produce arrive too late or lack impact,” he said.
“With 80% of AI projects struggling to deliver on expectations globally, 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 the large technology providers: Capital expenditure by Google, Microsoft, Amazon, and Meta is expected to top $380 billion this year, the majority of which is focused on AI infrastructure.
The message from the hyperscalers is clear: If we build it, they will come.
Addressing the infrastructure challenges that organisations experience is the key component to successful implementation of AI-based projects. Agile infrastructure as close as possible to the point of compute and data sets that are unified are seen as important parts of the recipe for getting full value from next-generation AI projects.
Although some market realignment is expected across the AI sector in the next six months, companies like Google are among those expected to be able to consolidate on the market and continue to offer game-changing technologies based on AI as it evolves.
(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 AI & Big Data Expo taking place in Amsterdam, California, and London. The comprehensive event is part of TechEx and co-located with other leading technology events. Click here for more information.
AI News is powered by TechForge Media. Explore other upcoming enterprise technology events and webinars here.
