Google Exec Fikes company needs to double AI serving ‘every six months: report
Tech companies are racing to build their infrastructure as their products expand resources that increase energy, clean up ChipMakers, and they need more energy. Google, rebranded as “The king of the web“It is one of those companies, and the High-Level Level for the Gig g has reportedly told employees that the company needs to improve its service capabilities significantly if it wishes to keep up with the demand for its AI services.
CNBC got its hands on a recent presentation by Amin Vahdat, VP of machine learning, systems, and cloud ai at Google. The presentation includes a slide on “AI compute demand” that Google says “should double in six months…. 1000x in the next 4-5 years.”
“The AI infrastructure race is the most critical and expensive of the AI race,” Vahdat reportedly said at the all-hands meeting where the presentation took place. Google’s “job” is actually to build this infrastructure, but you shouldn’t exclude the competition, really,” he added. “We’re going to spend a lot,” he said, in an effort to build an AI infrastructure that’s “more reliable, more organized and more manageable than what’s available.”
Since the CNBC story was published, Google has understood the reporting. While CNBC quoted Vahdat as saying that the company would need to “repeat” its manipulation capacity every six months, a Google spokesperson said that GIZMODs were taken out of context. The spokesperson also explained that Vahdat “It wasn’t about building a large amount of money for anything close to what was suggested. In fact, he just noted that the need for AI services means that we’re being asked about the efficiency of all the hardware, software, and more architecture.”
CNBC has since updated its reporting from “compute” to “work” power. Serve capacity will refer to Google’s ability to handle the increasing definition of user requests, while compute capacity refers to the company’s entire infrastructure dedicated to AI and other necessary resources. When asked for more clarification on the difference between the two, the spokesman for the organization said that the first title “Clearly Learn What We Mean We Show How We Are Repeated in How We Multiply the Number of Computes We Have – Instead, “the increasing power Amin described will be achieved in many ways, including new chips with efficiency and efficiency and optimization,” he added.
Whatever happens under the hood, it will appear that Google – like its competitors – needs to scale its operations to support its nascent ai infrastructure business. Vahdat’s comments came not long after reporting some chunky profits from its cloud business, which the company announced plans to spend next year.
During his presentation, Vahdat was reported to have said that Google needs to “be able to deliver 1,000 times more power, compute, network storage [than its competitors] Basically the same cost and more, the same power, the same level of power. ” He admitted that “it won’t be easy” but said that “working together and cooperating, we will get there.”
The race to build data centers — or “AI Infrastructure” as the tech industry calls it — is maddening. Like Google, Microsoft, Amazon, and Meta all said that they will build their capital expenditure in an effort to build the future of computing (cumulativel, big technology is expected to spend at least $ 400 billion in the next twelve months). As these places rise, they cause all kinds of drama in the communities where they live. Environmental and economic concerns are common. Some communities have begun protesting data center projects — and, in some cases, successfully overturning them. Still, given the amount of money invested in the industry, it’s going to be an ongoing battle for Americans who don’t want an AI colossus in their backyard.
