Big Tech’s massive spending on artificial intelligence is set to continue unchecked in 2025 after Amazon topped its rivals with a planned $100bn-plus investment in infrastructure this year.
Spending by the four leading US tech companies had already surged 63 per cent to historic levels last year. Now executives are vowing to accelerate their AI investments, dismissing concerns about the vast sums being bet on the nascent technology.
Microsoft, Alphabet, Amazon and Meta have reported combined capital expenditure of $246bn in 2024, up from $151bn in 2023. They forecast spending could exceed $320bn this year as they compete to build data centres and fill them with clusters of specialised chips to remain at the forefront of AI large language model research.
The scale of their spending ambitions — announced alongside their fourth-quarter earnings — has surprised the market and exacerbated a sell-off caused by the release of an innovative and cheap AI model from Chinese start-up DeepSeek in late January.
Microsoft and Google parent Alphabet each saw $200bn wiped from their market value after reporting weaker than expected growth in their cloud computing divisions alongside steep increases in capital spending. Google’s 8 per cent drop on Wednesday was its fifth-worst trading day in the past decade.
Some content could not load. Check your internet connection or browser settings.
“The unbridled enthusiasm across the entire ‘Magnificent Seven’ has been replaced by pockets of scepticism and created some ‘show me’ situations,” said Jim Tierney, head of the concentrated US growth fund at AllianceBernstein. “The concerns that I’ve had since summer are magnified today.”
Amid the hype about AI’s transformative potential, shareholders worry that doubling down on spending without a commensurate increase in revenues could eat into capital that would otherwise be returned in the shape of buybacks and dividends, whilst starving non-AI business lines.
Google has been opaque about usage and revenue from its Gemini chatbot, while companies have been wary of adopting Microsoft’s glitchy and costly Copilot “agents” to improve workforce productivity.
“If or when we see the cloud growth acceleration at Google or [Microsoft’s] Azure, or see Copilot uptake improve, investors will be more comfortable with spending at Alphabet or Microsoft,” said Tierney. “Cheaper and more commoditised AI models will probably amplify investor concerns in the meantime.”
DeepSeek’s R1 model was emblematic of such fears. The Chinese AI lab’s claim to have built a reasoning model with similar capabilities to Google and OpenAI’s products at a fraction of the price — and without access to Nvidia’s most advanced graphic processing units — caused the chipmaker’s stock to plunge 17 per cent, erasing $600bn in one day, from which it has only partially recovered.
Some content could not load. Check your internet connection or browser settings.
Big Tech chiefs have held their nerve. On Tuesday, Google’s Sundar Pichai said in defence of his plan to spend $75bn in 2025 — up 42 per cent from $53bn last year — that the AI opportunity was “as big as it comes, and that’s why you’re seeing us invest to meet that moment”. DeepSeek would add to demand by showing how new techniques could make it cheaper and spur new lines of research, he said.
Microsoft’s Satya Nadella said two weeks ago in Davos: “I am going to spend $80bn building out Azure, customers can count on Microsoft.” He reiterated his belief in the folly of slowing down and failing to capitalise on its early backing of start-up OpenAI.
And on Thursday, Amazon CEO Andy Jassy topped Google and Microsoft by forecasting more than $100bn in capital expenditure this year, up from $77bn in 2024 and more than double the $48bn of the previous year. The vast majority will go towards data centres and servers for Amazon Web Services, and Jassy said he was simply responding to “significant signals of demand”. The stock fell as much as 7 per cent in after-hours trading.
“Growth is cooking along a little bit, but the appetite to invest hasn’t been curtailed,” said Jeff Pearson, vice-president of cloud strategy at consultancy Presidio. “They are ploughing ahead even if the return on investment seems distant.”
Meta received a more positive reception to its earnings, with its shares rising even as chief Mark Zuckerberg pledged to spend “hundreds of billions” more on AI, on top of the $40bn invested in 2024.
“Investors have embraced Meta, even though their capex is growing, because there is a real-time return-on-investment improvement in client spending that is measurable,” said Tierny, referring to Meta’s use of AI to improve ad targeting on Facebook and Instagram.
Meta’s success in showing tangible returns from AI investment stood in contrast to Google, which faces new competitors and the difficult task of integrating AI into search without cannibalising its core advertising business.
The search giant has introduced brief answers, or “AI overviews”, at the top of search results, but these are displacing its lists of links, the first of which are often lucratively sponsored.
Nevertheless, “if there’s meant to be cracks in Google’s search empire, it certainly isn’t showing up yet”, said Bernstein analyst Mark Shmulik, pointing to a 13 per cent growth in ad revenue to $54bn in the final three months of 2024 alone. “Google hasn’t missed search expectations even once since ChatGPT launched nine quarters ago.”
Spending among the “Magnificent Seven” — which also includes Apple, Nvidia and Tesla — dwarfs the rest of the US benchmark S&P 500. Their capital spending rose 40 per cent in 2024 compared with 3.5 per cent among the remaining 493 companies, according to Société Générale. Profits among the elite group soared by a third in the same period, versus 5 per cent among the rest.
The spending spree is not limited to publicly listed companies, and neither Deep Seek nor fears of an AI bubble have slowed the flow of capital into Silicon Valley start-ups.
OpenAI’s Sam Altman has formed a partnership with SoftBank and Oracle to invest $100bn in AI-related US infrastructure, potentially rising to half a trillion over time. The Japanese investor is in talks to invest $25bn in the start-up at a valuation of $260bn.
“Could there be an AI winter at some point? Sure,” said Rishi Jaluria, an analyst at RBC Capital Markets. “But if you’re in a position to be a leader, you can’t take your foot off the gas.”
