Article
AI News Cloud Google

31 hours and counting: Big Tech's earnings test. Is the AI spending paying for itself?

Alphabet, Microsoft and Amazon report within 31 hours, with about $600 billion of planned capital spending behind them.

by TechDefused Newsroom · Editor IL
The image features a distorted view of a wall clock, showcasing its face with numbers clearly visible. The abstract distortion adds a surreal and artistic effect to the depiction of time. — Credit: Photo by Marko Brečić on Unsplash c Photo by Marko Brečić on Unsplash

Alphabet and Microsoft go first, after the US market close tonight, with calls at 8:30 p.m. and 9:30 p.m. UK time. Amazon follows on Thursday, with its call at 9:00 p.m. UK time.

Between them the three account for more than a third of the S&P 500's earnings.

Investors want to know one thing: is the spending on artificial intelligence (AI) finally paying for itself?

The scale of the bill

The three companies could spend roughly $600 billion in 2026 on data centres, chips, servers and networking.

Microsoft accounts for about $175 billion, Alphabet for $195 billion to $205 billion and Amazon for $220 billion.

Amazon has already raised its budget once this year, from $200 billion to $220 billion. Chief executive Andy Jassy says it still would not have enough capacity to meet expected demand next year.

The strain is showing in cash flow. Alphabet's free cash flow turned negative in the second quarter, reportedly for the first time since its 2004 flotation.

Savita Subramanian, a strategist at Bank of America, says hyperscalers' 2027 capital spending estimates have been revised up by nearly 90% this year.

"All eyes will be on AI monetization vs. spending," she says.

Cloud has to deliver

The argument will be won or lost in cloud computing.

Market expectations ahead of the results call for roughly 45% growth at Azure, Microsoft's cloud platform.

They call for 100% growth at Google Cloud, which grew 82% year on year in the second quarter, and 38% at Amazon Web Services.

One estimate puts combined hyperscaler AI spending at close to $850 billion in 2026. That compares with more than $2 trillion of cloud backlog, meaning work customers have contracted but not yet been billed for.

Analysts say the gap supports continued investment without guaranteeing near-term profit.

Google gave some detail on where the money goes at its Cloud Next event in April. It said 330 Google Cloud customers had processed more than a trillion tokens, the units of text AI models read and write, over the previous 12 months.

Throughput had risen to 16 billion tokens a minute from 10 billion the previous quarter. Roughly 60% of its latest quarterly infrastructure spending went on servers alone.

How the market reacts

The market has shown what happens when spending scares the livestock.

Meta narrowed its 2026 capital spending range to $135 billion to $145 billion last week. Its third-quarter capex of $34.9 billion outstripped operating cash flow.

Its shares fell about 7% the next day, despite a 26% rise in revenue. Days later it said it would raise up to $95 billion through share sales to fund the build-out.

Alphabet shares fell more than 7% after it raised its 2026 capital spending outlook to as much as $205 billion.

Microsoft has faced similar unease, with some analyst estimates putting its current-year capex above $190 billion.

Analysts expect meaningful returns to take roughly two to three years to arrive, not this earnings season.

So the test over the next 31 hours is not whether revenue grows, which it almost certainly will.

It is whether the companies can show that growth is coming from AI, and not from the advertising and retail businesses that built them.

by TechDefused Newsroom