Q2 capex for Meta, Alphabet, Amazon, and Microsoft

Sherwood News

Big tech’s huge AI spending isn’t slowing down. The revenue? Uhhhhh...

Today is a good time to remind you that almost every major tech company is dropping large amounts of cash on AI tech that isn't making them anything back.

Rani Molla

8/5/24 10:11AM

Last quarter was a huge quarter for capital expenditure at big tech companies.

This one was somehow even bigger.

Amazon, Microsoft, Alphabet and Meta spent a combined $52.8 billion on capex in the second quarter — or nearly 60% more than the same quarter last year, according to standardized data from FactSet. The main expense? AI, as big tech hopes to invest big on the next big thing.

And according to their forward-looking statements, that spending is expected to go up even more.

Amazon CFO Brian Olsavsky

Looking ahead to the rest of 2024, we expect capital investments to be higher in the second half of the year. The majority of the spend will be to support the growing need for AWS infrastructure as we continue to see strong demand in both generative AI and our non-generative AI workloads.

Meta CFO Susan Li

We anticipate our full year 2024 capital expenditures will be in the range of $37 billion to $40 billion updated from our prior range of $35 billion to $40 billion. While we continue to refine our plans for next year, we currently expect significant CapEx growth in 2025 as we invest to support our AI research and our product development efforts.

Microsoft CFO Amy Hood

To meet the growing demand signal for our AI and cloud products, we will scale our infrastructure investments with FY 2025 capital expenditures expected to be higher than FY 2024.

Alphabet CFO Ruth Porat

Our reported CapEx in the second quarter was $13 billion. Once again, driven overwhelmingly by investment in our technical infrastructure with the largest component for servers followed by data centers. Looking ahead, we continue to expect quarterly CapEx throughout the year to be roughly at or above the Q1 CapEx of $12 billion.

What’s changed this quarter is that investors are more interested in when all this AI spending might actually yield returns.

When asked about AI capex ROI, Alphabet CEO Sundar Pichai said:

“The one way I think about it is when you go through a curve like this, the risk of underinvesting is dramatically greater than the risk of overinvesting for us.”

Microsoft CEO Satya Nadella said the company is pegging its AI spending to demand signals.

It's more important to manage, to capture the opportunity with the right product portfolio that's driving value.

Meta’s Porat:

We don't expect our GenAI products to be a meaningful driver of revenue in 2024. But we do expect that they're going to open up new revenue opportunities over time that will enable us to generate a solid return off of our investment

In other words: They’re spending now in hopes of bigger returns down the line.

More Tech

See all Tech

Jon Keegan22h

WSJ: Anduril’s weapons systems have failed during several tests

Autonomous drones by sea, land, and air. Futuristic AI-powered support fighter jets, and swarms of networked drones controlled by sophisticated software. These are some of the visions for the future of warfare pitched by defense tech startup Anduril. Cofounded by Oculus founder Palmer Luckey, the Peter Thiel-backed startup has landed some major national security contracts based on this futuristic outlook for battlefield AI.

But according to a report from The Wall Street Journal, the company’s tech is failing key tests in the real world, raising concerns about the viability and safety of Anduril’s systems within the military command.

Anduril’s Altius drones proved vulnerable to Russian jamming while deployed in Ukraine and have been pulled from the battlefield, per the report.

More than a dozen sea-based drone ships powered by Anduril’s Lattice command and control software recently shut down during a Navy test, creating a hazard for other vessels in the exercise.

And this summer, during a drone intercept test, Anduril’s counter-drone system crashed and caused a 22-acre fire at a California airport, the report found.

Anduril told the WSJ that the failures are just part of its rapid iterative development process:

“We recognize that our highly iterative model of technology development — moving fast, testing constantly, failing often, refining our work, and doing it all over again — can make the job of our critics easier. That is a risk we accept. We do fail… a lot.”

Exclusive | ‘We Do Fail … a Lot’: Defense Startup Anduril Hits Setbacks With Weapons Tech

Anduril’s Altius drones proved vulnerable to Russian jamming while deployed in Ukraine and have been pulled from the battlefield, per the report.

More than a dozen sea-based drone ships powered by Anduril’s Lattice command and control software recently shut down during a Navy test, creating a hazard for other vessels in the exercise.

And this summer, during a drone intercept test, Anduril’s counter-drone system crashed and caused a 22-acre fire at a California airport, the report found.

Anduril told the WSJ that the failures are just part of its rapid iterative development process:

“We recognize that our highly iterative model of technology development — moving fast, testing constantly, failing often, refining our work, and doing it all over again — can make the job of our critics easier. That is a risk we accept. We do fail… a lot.”

Jon Keegan11/28/25

OpenAI’s partners shouldering $100 billion of debt, taking on all the risk

OpenAI’s ambitious plans for global AI infrastructure projects — like its series of massive Stargate AI data centers — will require tens of billions of dollars funded by debt, but you won’t find much of that on OpenAI’s balance sheet.

According to a new analysis by the Financial Times, OpenAI has somehow convinced its many partners to shoulder at least $100 billion in debt on its behalf, as well as the risks that come with it.

Partners Oracle, SoftBank, CoreWeave, Crusoe, and Blue Owl Capital are all taking on debt in the form of bonds, loans, and credit deals to meet their obligations with OpenAI for infrastructure and computing resources.

Having close ties with OpenAI has been an anchor for many publicly traded companies in recent weeks. The company’s cash burn and the rise of Gemini 3 have seemingly darkened its outlook and fostered guilt by association for many of its close partners and investors. Most notably, Oracle’s aggressive capital expenditure plans to support demand from OpenAI have sparked a sell-off in its stock while widening its credit default swap spreads.

A senior OpenAI executive told the FT: “That’s been kind of the strategy. How does [OpenAI] leverage other people’s balance sheets?”

OpenAI partners amass $100bn debt pile to fund its ambitions

A senior OpenAI executive told the FT: “That’s been kind of the strategy. How does [OpenAI] leverage other people’s balance sheets?”

David Crowther11/27/25

Chinese tech giants are training their models offshore to sidestep US curbs on Nvidia’s chips

Nvidia can’t sell its best AI chips in the world’s second-largest economy. That’s an Nvidia problem. But it’s also a China problem — and it’s one that the region’s tech giants have resorted to solving by training their AI models overseas, according to a new report from the Financial Times.

Citing two people with direct knowledge of the matter, the FT reported that “Alibaba and ByteDance are among the tech groups training their latest large language models in data centers across south-east Asia.” Clusters of data centers have particularly boomed in Singapore and Malaysia, with many of the sites kitted out with Nvidia’s latest architecture.

One exception, per the FT, is DeepSeek, which continues to be trained domestically, having reportedly built up a stockpile of Nvidia chips before the US export ban came into effect.

Last week, Nvidia spiked on the news that the Trump administration was reportedly considering letting the tech giant sell its best Hopper chips — the generation of chips that preceded Blackwell — to China.

China’s tech giants take AI model training offshore to tap Nvidia chips

One exception, per the FT, is DeepSeek, which continues to be trained domestically, having reportedly built up a stockpile of Nvidia chips before the US export ban came into effect.

Millie Giles11/27/25

Alibaba unveils its first AI glasses, taking on Meta directly in the wearables race

Retail and tech giant Alibaba launched its first consumer-ready, AI-powered smart glasses on Thursday, marking its entrance into the growing wearables market.

Announced back in July, the Quark AI glasses just went on sale in the Chinese retailer’s home market, with two versions currently available: the S1, starting at 3,799 Chinese yuan (~$536), and the G1, at 1,899 yuan (~$268) — a considerably lower price than Meta’s $799 Ray-Ban Display glasses, released in September.

Alibaba's AI glasses to rival Meta go on sale for $500

Jon Keegan11/26/25

Musk: Tesla’s Austin Robotaxi fleet to “roughly double” next month, but falls well short of earlier goals

Yesterday, Elon Musk jumped onto a frustrated user’s post on X, who was complaining that they were unable to book a Robotaxi ride in Austin. Musk aimed to reassure the would-be customer that the company was expanding service in the city:

“The Tesla Robotaxi fleet in Austin should roughly double next month,” Musk wrote.

While that sounds impressive, there are reports that Austin has only 29 vehicles in service.

But last month, Musk said the Robotaxi goal was to have “probably 500 or more in the greater Austin area” by the end of the year.

Meanwhile, Google’s Waymo has more than 100 autonomous taxis running in Austin, and 1,000 more in the San Francisco Bay Area.

Elon Musk slashes Tesla Robotaxi fleet goal from 500 to ~60 in Austin

“The Tesla Robotaxi fleet in Austin should roughly double next month,” Musk wrote.

While that sounds impressive, there are reports that Austin has only 29 vehicles in service.

But last month, Musk said the Robotaxi goal was to have “probably 500 or more in the greater Austin area” by the end of the year.

Meanwhile, Google’s Waymo has more than 100 autonomous taxis running in Austin, and 1,000 more in the San Francisco Bay Area.

Latest Stories

Sherwood Media, LLC produces fresh and unique perspectives on topical financial news and is a fully owned subsidiary of Robinhood Markets, Inc., and any views expressed here do not necessarily reflect the views of any other Robinhood affiliate, including Robinhood Markets, Inc., Robinhood Financial LLC, Robinhood Securities, LLC, Robinhood Crypto, LLC, or Robinhood Money, LLC.