LinkedIn will not expand its data centers next year

Unlike many others And the big tech platforms, LinkedIn has decided that it will not spend aggressively on expanding its AI Data centers This fiscal year. Executives at the professional social network told WIRED that it plans to keep its investments in GPUs stable, and its compute and storage footprint remains steady as well.

The spending calculations apply to LinkedIn’s fiscal year, which began last month and ends next June. The company says it was able to avoid spending too much on AI hardware because it found ways to use its existing GPUs twice as efficiently over the past six months. LinkedIn’s plan could still fall apart as hardware requirements for AI change rapidly, but executives say the company has already taken into account higher prices. Memory chips.

“One of the goals we set is to try to keep our compute footprint flat or as close to it as possible while shipping more compute-hungry stuff to production,” says Erran Berger, chief technology officer for engineering at LinkedIn. “This is a very bold statement to make in today’s world.”

Berger and Raghu Hirimagalur, LinkedIn’s chief technology officer for infrastructure, say they want to be cautious about spending and that the new restrictions will spur engineering teams to get more creative when developing many of the new generative AI features LinkedIn plans to launch. Berger says he believes the efficiency gains could compound over time, enabling LinkedIn to get more out of data center expansions when they eventually increase their budgets again.

“I really want to reiterate that for a company of our size, saying we’ll do this for a full year without additional storage and compute is no small feat, but it takes a lot of work to get there,” Hiremagalur says.

Companies like OpenAI, Meta, and Google Searching for all the money they can find and Pairing up in unexpected partnerships To build, equip and operate huge data centers filled with the latest computer chips. Labor and spare parts shortages have disrupted many projects, and many companies have been forced to limit customers’ use of some AI tools. But there is Questions also increase On whether continued investment in AI is sustainable. LinkedIn, with more than 1.3 billion users, is perhaps the largest company yet that has not publicly addressed spending concerns by resisting the construction boom.

“It’s encouraging for the industry,” says Song Yi-yeon, managing partner of Principal Venture Partners and a board member of server maker HP. “It signals that AI is beginning to move from experimental to productive discipline. The winning companies will not simply be the ones that spend more on infrastructure.”

Own it

A few years after Microsoft acquired LinkedIn in 2016, the company tried to move to its parent company’s Azure cloud service, but it didn’t make economic sense to squeeze the giant social network into general-purpose data centers. “Microsoft Azure was growing like crazy, the level of customer demand was very high, and at the same time we saw rapid growth on the LinkedIn side,” says Hirimagalur.

In 2022, LinkedIn pulled out all the stops in its data centers in Oregon, Texas, and Virginia. Ownership has given LinkedIn significant control over every detail of its technology, setting itself up well to face the realities of the new age. Around the same time, LinkedIn began developing AI-based assistants that could help users write messages, find jobs, and recruit candidates. The endeavor was not cheap. “Every inquiry that comes to our site gets more expensive over time,” says Hirimagalur, adding that the amount of data stored on LinkedIn doubles annually. “This is not a sustainable place.”