Meta’s Shares Surge on Cloud Business Plans
Meta’s stock jumped close to 10% on Wednesday after news emerged that the company is looking to sell off surplus computing power, aiming to recover some of the considerable investment it has made in artificial intelligence.
The tech giant, based in Menlo Park, California, is quickly moving to expand its data centers and chip acquisition, as well as develop a new cloud service that would provide access to its AI technology and computational resources. As highlighted by a report from Bloomberg, this is a significant shift for the company.
This decision appears to be a timely reassurance for investors, especially given their rising concerns over whether Meta can deliver adequate returns from the vast sums it has spent amassing a powerful computing infrastructure aimed at creating “superintelligence.”
Comments from Meta’s leadership haven’t been made public yet.
The introduction of this cloud service promises to generate income from unnecessary capacity while also positioning Meta to go head-to-head with established players like Amazon, Microsoft, Google, Coreweave, and SpaceX.
The report further mentions that Meta is contemplating whether to offer access to its AI models directly or instead to provide raw computing power. It should be noted that plans might evolve as the situation develops.
If Meta opts to sell access to its AI models, it would mirror the strategy used by Amazon: managing the infrastructure and charging clients to utilize these services.
SpaceX, led by Elon Musk, has a similar approach after its acquisition of the AI firm xAI. They’ve secured high-value lease agreements with companies such as Anthropic and Google, which paid $1.25 billion and $920 million per month, respectively, for access to their expansive Memphis data center.
Alternatively, Meta might decide to follow CoreWeave’s example and sell direct access to its computing power.
The rapid development of AI spurred by the launch of OpenAI’s ChatGPT in 2022 has prompted many firms to urgently accumulate significant computing resources as demand grows while capacity remains limited.
Interestingly, Meta’s stock had faced declines following an increase in its spending forecast to $145 billion back in April. This rise in forecasts occurred amid fears that AI stocks might be inflating, reminiscent of the dot-com bubble in the early 2000s.
Mark Zuckerberg, CEO of Meta, has been vocal about the necessity of maximizing the company’s computing capabilities due to supply constraints, but he also hinted in May at the possibility of monetizing excess power.
“That’s definitely something to consider,” Zuckerberg mentioned during a shareholder meeting. “Almost every week, we’re approached by different companies asking us to initiate an API service or check if we have available capacity to sell at a premium.”
He added, “We haven’t pursued it yet because we see potential uses for this computing power. Nonetheless, if we reach a point of overbuilding, it remains an option we can explore. This gives us some confidence to invest in these developments.”
Last summer, in a major move, Meta recruited AI innovator Alexander Wang, investing a staggering $15 billion in a 49% share of his startup, Scale AI.
Under Wang’s leadership, the company introduced its first AI model in April, named Muse Spark, although it failed to meet expectations for being a groundbreaking tool.
Wang defended the model, suggesting it was simply an “appetizer” while others were still finalizing the main dish.
