Big Tech groups say their $100bn AI spending spree is just beginning
Technology stocks have been volatile as Microsoft, Meta, Amazon and Google report huge increases in their investments in artificial intelligence
 From left: Microsoft’s Satya Nadella, Alphabet’s Sundar Pichai, Meta’s Mark Zuckerberg and Amazon’s Andy Jassy © FT montage/Bloomberg/AP
current progress 94%
Stephen Morris(opens a new window) and Hannah Murphy(opens a new window) in San Francisco and Camilla Hodgson(opens a new window) in London
38 minutes ago
19Print this page
Stay informed with free updates
Simply sign up to the Big Tech myFT Digest -- delivered directly to your inbox.
Big Tech companies have boosted their capital spending by 50 per cent to more than $100bn this year, as they race to build the infrastructure supporting artificial intelligence, despite growing scepticism from Wall Street about the returns on the unprecedented investment.
Microsoft, Alphabet, Amazon and Meta all revealed massive increases in spending in the first six months of 2024 — totalling $106bn — in their latest quarterly earnings reports, as their leaders brushed off stock market jitters to pledge further investment hikes over the next 18 months.
“At this point, I’d rather risk building capacity before it is needed, rather than too late,” Meta chief Mark Zuckerberg said this week, as he predicted the Facebook parent’s capital spending could hit $40bn this year.
Their collective forecasts mean Big Tech’s AI-related investment could more than double by year-end. Analysts at Dell’Oro Group now expect as much as $1tn could be channelled into infrastructure such as data centres within five years, even though the companies have so far failed to convince investors that their customers are prepared to spend big on AI products and services.
“Tech management teams are pressing their bets on spending,” said Jim Tierney, head of US growth at AllianceBernstein. “Investors are still unclear what all the business models and pay-offs are. This is creating a ‘trust us’ environment which is not overly comforting given the aggregate spending.”
Big Tech’s latest round of earnings reports collided with a broader dive in Wall Street sentiment, as the Nasdaq fell into a correction on Friday on weaker US jobs numbers. Shares in semiconductor companies, including the leading AI chipmaker Nvidia, have been particularly volatile this week as investors grew more sensitive to Big Tech’s comments on spending plans.
Nvidia gained or lost around $200bn in each of three successive trading sessions, while Intel, which is yet to capture much of the spending on AI infrastructure owing to its lack of competitive products, lost more than a quarter of its value on Friday after announcing massive job cuts.
Big Tech companies ramp up infrastructure investment to support AIQuarterly capital expenditures ($bn)
Amazon figures are business-wide purchases of property and equipment, including for its retail business
Microsoft
Meta
Amazon*
Alphabet
Q2 2024Q1 2024Q4 2023Q3 2023Q2 2023Q1 2023Q4 2022Q3 20220604020
Source: Company reports
Yet even as the stocks of Google, Microsoft and Amazon sold off in the immediate wake of their earnings reports, Big Tech executives made no apologies for their spending spree.
Zuckerberg estimated the amount of computing power required to train its next large language model would be “almost 10 times more” than the previous version, even while conceding that it would be “years” before some of its AI features, such as its Meta AI chatbot, made any money “by themselves”.
“In tech when you are going through transitions like this?.?.?.?the risk of underinvesting [in AI] is dramatically higher than overinvesting,” said Google chief executive Sundar Pichai.
After Google’s parent Alphabet last week reported a 90 per cent surge in capital spending in the first two quarters of 2024 to $25bn, Microsoft on Tuesday responded with a 78 per cent increase to $33bn. Amazon’s investments in property and equipment during the first half of the year — which includes spending for its vast ecommerce and logistics network — jumped 27 per cent to $32.5bn, it disclosed on Thursday.
Amazon said in a subsequent filing it expected capital spending overall to “meaningfully increase” in 2024, which chief financial officer Brian Olsavsky said would mostly be channelled into new cloud infrastructure. Generative AI was now a “multibillion dollar business” for the company, he added. But amid the broader market rout and signs of weakness in its consumer business, shares in Amazon fell by 9 per cent on Friday.
Google executives had pointed to resilient revenues from advertising — up 11 per cent to $64.6bn in the second quarter alone — as evidence the company’s core business was healthy enough to shoulder the spending burden.
But while that failed to stem a drop in Alphabet’s stock, Microsoft investors appeared more reassured after chief financial officer Amy Hood said its data centres were long-term assets that would be “monetised over 15 years and beyond”.
If demand for AI products and services was less than expected, Microsoft could slow the pace at which it fills data centres with costly AI hardware, said Hood. In the latest quarter cloud growth had been constrained by a lack of capacity, not customer demand, Microsoft said.
Capital expenditure is being discussed far more on recent Big Tech earnings callsMentions of 'capex' and related phrases on quarterly earnings calls, 'Magnificent 7' companies excluding Nvidia
Microsoft
Tesla
Meta
Apple
Amazon
Alphabet
202420232022202120200806040202722221918222419225423396042416560
Source: AlphaSense
Much of the investment from Big Tech groups is going towards buying land and constructing new data centres for their cloud computing businesses. Huge sums are also being spent on hardware including the specialised clusters of chips — mainly made by Nvidia — needed to train and run large language models that underpin chatbots.
Demand for cloud services has surged as companies trial generative AI services to automate processes and improve productivity, even if most of those experiments are yet to be put into full production. Meanwhile, start-ups like OpenAI, Anthropic, Elon Musk’s xAI and France’s Mistral are competing for scarce computing resources to train ever more advanced LLMs.
This week’s stock gyrations follow a historic bull run. The tech-dominated Nasdaq 100 index has climbed by around 70 per cent since the start of 2023, when AI fervour first began to take hold, making Apple, Microsoft, Nvidia, Alphabet and Amazon the five largest public companies in the world.
AI was discussed hundreds of times during this quarter's Big Tech earnings callsMentions of AI and related phrases on quarterly earnings calls, 'Magnificent 7' companies excluding Nvidia
Microsoft
Tesla
Meta
Apple
Amazon
Alphabet
202420232022202120200300200100661016292943396861123188244244275292271
Source: AlphaSense
“I think the natural comparable that jumps to mind for many investors is the telecom bubble of the late 1990s and early 2000s,” said Michael Hodel, an analyst at Morningstar. “Most of the companies involved in that buildout went bust. This buildout seems similar in some ways?.?.?.?The main difference, though, is that the firms doing most of the building have massively profitable existing businesses and fortress-like balance sheets.”
Charts by Clara Murray and Stephanie Stacey
Copyright The Financial Times Limited 2024. All rights reserved.
Reuse this content(opens in new window)(opens a new window) CommentsJump to comments section
Latest on Big Tech
Follow the topics in this article
Comments
Commenting is only available to readers with an FT subscription
Please login or subscribe to join the conversation.
Sort by
NewestOldestMost repliesMost Recommended
Crema not Creme
29 minutes ago
$100bn for chips while we start a recession.. nice.
Share
Report
SpendItLikeBeckham
32 minutes ago
Testament to how much cash these companies generate that they can indefinitely spend unlimited billions on a hype job vanity project.
Share
Report
DRW
35 minutes ago
While the music is playing they all have to get up and dance.
Share
Report
GroEL
49 minutes ago
Are these the same companies that invested billions in self-driving cars, blockchain, NFTs, metaverse and so on ad absurdum? A bad case of ‘You hype it and they’ll come’
Share
Report
M
1 hour ago
Hype
Hype
Hype
They have to perpetuate the hype. Like a ponzi scheme.
Share
Report
Neil at home
2 hours ago
In the UK a company recently bought a
polyhalite mine, which required a large investment even though the product didnt have a proven market. They recently stopped mining and wrote down the value of the investment.
Share
Report
Backlight
2 hours ago
Didn't the US government want to encourage US companies to increase their Capex in the US, and planned to introduce incentives for this? Bring back cash to the US and invest etc etc...
I guess the up shot of this, or one cynical way to look at it, is there will be in the short term less unrealised capital gains for the future Democrat administration to tax. As these capital light firms become at least in the short term slightly more capital heavy as they buy land and construct their new data centres!
I wonder if there is an inflation hedge in this too?
Share
Report
Badabing
2 hours ago
I’d be interested in finding out where the data centre are getting built. Are they in the USA or Ireland?
Share
Report
F2020
1 hour ago
In reply to Badabing
Many in Malaysia apparently
archive.ph
Share
Report
spin spin spin
7 minutes ago
In reply to Badabing
Google is building and expanding data centres in Virginia (which is popular with a variety of tech companies including Roblox given its nuclear power), Iowa, and Ohio and in Malaysia for the first time.
Share
Report
viewer discretion is advised
2 hours ago
FOMO at it's finest ...
Share
Report
SAK
2 hours ago
AI is for battyboys
Share
Report
Fives
2 hours ago
In reply to SAK
It certainly is a bet on human behaviour, if that’s what you mean.
Share
Report
SAK
27 minutes ago
In reply to Fives
I agree
Share
Report
John M
2 hours ago
...The main difference, though, is that the firms doing most of the building have massively profitable existing businesses and fortress-like balance sheets.”
And stock prices that fully reflect the TRILLIONS in money printing by CBs.
Share
Report
donniedarko
2 hours ago
The dilemma for Google is that, on the one hand their search engine business model is currently threatened by generative AI, which is also trashing search results with digital hallucinations - and if the AI bubble bursts they could end up staffing vast sums of money up the wall.
Share
Report
F2020
3 hours ago
(Edited)
I’d rather risk building capacity before it is needed,
$100bn
The catch is these things depreciate at 20 to 25% pa (or faster). That is $25bn of expenses.
Ok as long as the ads money keeps rolling in..
Share
Report
Rather sceptical
2 hours ago
In reply to F2020
Not according to Microsoft:
Microsoft investors appeared more reassured after chief financial officer Amy Hood said its data centres were long-term assets that would be “monetised over 15 years and beyond”.
Share
Report
F2020
2 hours ago
(Edited)
In reply to Rather sceptical
The DCs might be, the H200 cards are out of date in probably 3 years ??
If they did amortise the cards over 15, SEC may want to take a look.
Share
Report
Top of commentsTop of page
Useful links
Support
View Site Tips Help Centre Contact Us About Us Accessibility myFT Tour Careers
Legal & Privacy
Terms & Conditions Privacy Policy Cookie Policy Manage Cookies Copyright Slavery Statement & Policies
Services
Share News Tips Securely Individual Subscriptions 4-b62d-cb33-4c278e6fdf61&cpccampaign=B2B_link_ft.com_footer]Professional Subscriptions Republishing Executive Job Search Advertise with the FT Follow the FT on X FT Channels FT Schools
Tools
Portfolio FT App FT Digital Edition FT Edit Alerts Hub Business School Rankings Subscription Manager News feed Newsletters Currency Converter
Community & Events
FT Community FT Live Events FT Forums Board Director Programme
More from the FT Group
Markets data delayed by at least 15 minutes. © THE FINANCIAL TIMES LTD 2024. FT and ‘Financial Times’ are trademarks of The Financial Times Ltd.
The Financial Times and its journalism are subject to a self-regulation regime under the FT Editorial Code of Practice.
 |
