On April 19, The Information published an article stating that for most of the past year, Meta has been trying to get other technology giants to fund its Llama AI model. This reflects rising spending on AI technology, even as its potential returns remain uncertain.
AI data center
Tech giants have deep pockets, but spending on AI is draining their cash. Meta generated $91.3 billion in cash from operations last year, with $39 billion spent on capital expenditures and $35 billion on stock buybacks and dividends. This year, Meta expects capital spending to rise 60% to between $60 billion and $65 billion. We can't assume that its operating cash flow will also grow by 60%, so the problem is that Meta shareholders may already be questioning the value of investing in AI.
Amazon, Microsoft and Alphabet reported operating cash flow of between $116 billion and $125 billion in fiscal 2024. Alphabet spent nearly all of its cash flow on capital spending, dividends and stock buybacks last year, and it expects capital spending to rise to $75 billion this year. Microsoft and Amazon have a little more of a cushion between their cash flow and spending, but that cushion could soon be exhausted as capital spending ramps up. Of course, all of these companies have plenty of cash in the bank, but that cash could quickly dwindle as AI spending increases. Meta, for example, has discussed building a $200 billion data center campus.
Borrowing money to fund AI is one option. Bloomberg reported a few weeks ago that Meta was working with Apollo on a $35 billion debt financing for AI data centers. There is also the possibility that a consortium similar to the "Stargate" project will emerge - Microsoft is partnering with Blackstone Group, Nvidia and xAI to establish a data center fund. But it's a $30 billion plan that the group says will eventually grow to $100 billion, and the money could soon be spent.