Unlike many others enormous technology platforms LinkedIn has decided not to spend aggressively on expanding its AI data centers this fiscal year. Executives at the professional social network tell WIRED they plan to keep GPU investment steady, while demand for processing power and storage also remains unchanged.
The expense calculations are for LinkedIn’s fiscal year, which began last month and ends next June. The company says it has managed to avoid substantial spending on AI hardware because it has found ways to employ existing GPUs twice as efficiently over the past six months. LinkedIn’s plan may not come to fruition yet as AI hardware requirements change rapidly, but executives say the company has already taken into account rising memory chip prices.
“One of the goals we’ve set for ourselves is to basically keep the compute area flat, or as flat as possible, when sending higher compute-demanding components to production,” says Erran Berger, LinkedIn’s chief technology officer for engineering. “That’s quite a bold statement in today’s world.”
Berger and Raghu Hiremagalur, LinkedIn’s chief technology officer for infrastructure, say they want to be conservative with spending and that the fresh restrictions will motivate engineering teams to be more imaginative when developing the many fresh generative AI features LinkedIn plans to launch. Berger says he believes the efficiency gains could accelerate over time, allowing LinkedIn to better take advantage of its data center expansion when it eventually increases its budgets again.
“I really want to reiterate that for a company of our scale, let’s say an entire year of doing this without incremental storage and compute is no mean feat, but it took a lot of work to get there,” says Hiremagalur.
Companies like OpenAI, Meta, and Google are raking in all the money they can and forming unlikely partnerships to build, equip, and operate massive data centers filled with the latest computer chips. Labor and parts shortages have held up many projects, and many companies have had to limit customer employ of some AI tools. However, there are growing questions about whether continued investment in artificial intelligence is sustainable. LinkedIn, with more than 1.3 billion users, is perhaps the biggest company yet to publicly address spending concerns by bucking the construction boom.
“This is encouraging for the industry,” says Songyee Yoon, managing partner at Lead Venture Partners and a board member of server maker HP. “This suggests that AI is starting to move from experimentation to production discipline. The companies that win will not simply be the ones that spend the most on infrastructure.”
Owning 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 enormous, and at the same time we saw explosive growth on the LinkedIn side,” says Hiremagalur.
In 2022, LinkedIn took full advantage of its own data centers in Oregon, Texas and Virginia. This owner has given LinkedIn significant control over every detail of its technology, well-preparing it to meet the realities of the fresh era. Around the same time, LinkedIn began developing AI-powered assistants that could assist users write messages, find jobs, and recruit candidates. The undertaking was not low-cost. “The cost of each query coming to our site increases over time,” says Hiremagalur, adding that the amount of data LinkedIn stores is doubling annually. “It’s not a sustainable place.”