Big Tech is shovelling cash into AI furnaces so fast that even its old money-printing tricks are starting to wheeze.
According to the Washington Post, giant tech companies are “feeding every available dollar into the cash-incinerating maw of AI machines.”
It warned that “Tech superstars that once had oodles of cash left over at the end of each year are now flipping into the red.”
Optimists still reckon AI will unleash massive corporate profits and spread wealth and wellbeing across society. The awkward bit is that doubts about this vision are getting louder as the bills pile up.
Apollo Global Management chief economist Torsten Slok said: “This AI thing better work out because if it doesn’t, we’re going to have a problem.”
AI costs and doubts are spreading through the market. The US stock market has swooned this summer on fears that the AI bubble might go pop.
The American AI punt is being led by tech outfits splashing cash on huge data centres stuffed with chips and kit. These warehouses are needed to train advanced AI models and flog them to customers.
Google, Microsoft, Meta and Amazon told investors last week that AI-related sales and business deals were climbing. Advertisers are using the stuff to sharpen marketing pitches.
Corporations and start-ups are buying chatbot access and other AI software to squeeze more productivity from their workers. That sounds lovely until the invoice arrives.
The spending can keep going only if AI creates a much bigger avalanche of revenue to pay for it. The latest financial results suggest the costs are trampling much of the sales boost.
At Google, every dollar of cash generated in the past three months was met by $1.15 in spending. That went on AI chips, equipment, data centre land and other pricey toys.
The company is plugging the gap partly by borrowing money and selling more stock. That is a curious look for a firm once treated as a cash gusher.
Next year, five leading AI players, Google, Amazon, Microsoft, Meta and Oracle, are projected to have negative free cash flow. That measures what is left after expenses and AI infrastructure costs.
The numbers, compiled from investment analyst projections by S&P Global Market Intelligence, show a sharp reversal. These were meant to be some of the world’s most reliable cash machines.
By standard accounting measures, the companies remain profitable because AI infrastructure costs are spread across many years. That trick makes the burn look tidier than it feels.
Pessimists see a bet so huge that it cannot possibly pay off and they are getting noisier as the spending grows and the payoff stays conveniently vague.
The Bank for International Settlements, a usually calm Swiss outfit advising government bankers, recently warned of “economy-wide recessions” if the AI boom falters.
That could mean pain for workers and communities across the US. The corporate AI binge is no longer just a Silicon Valley vanity project.
Jefferies global head of equity strategy Christopher Wood said: “I’m not saying AI is going to go away, it’s just not clear to me these guys are going to make money on it.”
