Big Tech earnings raise an uncomfortable question: Who pays for AI?

Amazon, Alphabet and Intel all reported Q3 results that fell short of expectations. Cloud growth was strong, but AI capex keeps climbing, and free cash flow AI spending is turning negative—testing whether the AI trade can hold up.

Amazon (AMZN), Alphabet (GOOGL) and Intel (INTC) all reported Q3 results that fell short of expectations on key metrics. Cloud growth was strong, but AI capex keeps climbing, and free cash flow is weakening. Investors are now questioning whether AI spending can be monetised fast enough, and the AI trade is feeling the pressure.

The key theme of this earnings season is not slowing growth but the cost of that growth. Google Cloud revenue surged 103%, and AWS grew 39%, yet investors focused on AI capex and cash flow instead. The idea that AI spending quickly turns into profit, which has powered the AI trade this year, is now being tested—a shift we explored in our article on why Big Tech earnings changed everything, where investors stopped rewarding AI capex unless it came with real cash flow or proven demand.

Amazon: AWS accelerates, but cash flow stays negative

Amazon reported an adjusted EPS of 1.89 USD, below the 1.96 USD consensus. Its revenue of 201.94 bln USD also fell slightly short of the 202.31 bln USD estimate, though it grew 12%. Net income was 20.69 bln USD, down from 21.19 bln USD a year earlier.

AWS revenue rose 39%, edging past the 38% consensus. The core online stores business grew just 2% to 68.5 bln USD, its slowest since early 2023, and advertising revenue growth also cooled to 12%.

The pressure is on the spending side. Q3 AI capex came in at 75.9 bln USD, far above estimates. Free cash flow was negative for a second consecutive quarter, with 42.4 bln USD flowing out over the past 12 months—one of the clearest signs yet of how AI spending is reshaping Amazon’s free cash flow. CEO Jassy said he expects AWS growth to stay above 40% in both 2027 and 2028, and that 2027 capex will rise substantially from here.

Summary:

Management’s message is that demand is strong. But what the market wants to see is when rising AI capex starts coming back as cash flow.

Alphabet: Cloud up 103%, but free cash flow turns deeply negative

Alphabet reported adjusted EPS of 2.67 USD, well below the 3.02 USD estimate. Revenue of 127.58 bln USD also missed the 128.22 bln USD consensus, despite a 25% increase. Net income fell to 26.17 bln USD (2.12 USD per share) from 34.98 bln USD (2.87 USD per share) a year earlier, partly because of 8.7 bln USD in losses on equity investments.

Google Cloud revenue grew 103%, well above the 90% consensus. Search revenue grew 12% against expectations of 15%, extending a slowdown from 19% in Q1 and 17% in the second.

The cash flow picture is the bigger concern. Free cash flow was -9.07 bln USD, far worse than the -3.5 bln USD consensus and the second negative quarter in a row—a clear sign that AI spending is now hitting free cash flow even at the strongest-performing names. Q3 AI capex was 51.36 bln USD, below the 55.09 bln USD estimate, but the full-year outlook was held at 195 bln to 205 bln USD. CFO Anat Ashkenazi said Alphabet expects capital spending to increase significantly in 2027.

AI capex at Google is forecast to keep climbing through 2027, reaching 69.78 bln USD in Q1 2027.

Summary:

Cloud shows that AI investment is converting into revenue. But slowing search growth and weaker free cash flow show that the cost of that spending is already visible in the cash flow statement.

Intel: AI tailwinds and a weak PC market

Intel reported an adjusted EPS of 0.52 USD, above the 0.40 USD estimate. Revenue of 16.51 bln USD was slightly below the 16.56 bln USD consensus, though it grew 21%. Data center revenue jumped 78%, reflecting the same hyperscaler-driven demand we tracked in our article on how long the semiconductor rally can last, while the client computing segment fell 5%.

Guidance was the disappointment. Q4 EPS guidance of 0.25 USD was well below the 0.45 USD estimate, and revenue guidance of 15.0 bln to 16.0 bln USD missed the 17.29 bln USD consensus. CEO Tan said results would have been better without some raw material shortages and a factory closure. He also named memory as one of the biggest challenges for the PC market, saying prices have risen sharply and that it will be a very serious issue for everyone in 2027. Intel expects the PC market weakness to last at least the first half of 2027, with a 12% decline in 2026 followed by a further 10% decline next year.

Intel shares (USD) plunged to 64.74 USD on 30 October after the earnings shock, even as AI-driven data center demand stayed strong.

Summary:

Intel captures both sides of the AI cycle, with gains from data center demand and pressure on PCs from higher memory prices. The AI capex boom is not lifting the entire semiconductor chain evenly.

Final thoughts: Demand is proven, but the payback timeline is not

From on-cloud numbers alone, AI demand is hard to doubt—a dynamic that’s reshaping broader index-level sentiment well beyond these three names. But management at all three companies pointed to higher AI capex next year, and none offered a date for when free cash flow pressure from AI spending will ease. The next key points to watch are 2027 capex guidance on the Q4 earnings calls and when cloud growth starts to translate into better free cash flow. Until then, the AI trade is likely to react more to spending concerns than to fundamentals.

Disclaimer: This article is for informational purposes only and does not constitute financial or trading advice. Always conduct your own research or consult a licensed financial advisor before making any investment decisions.