When AI capital spending eats cash, how long will the market wait?

Why did growth and cash burn show up in the same quarter?

In the last week of October, the three big cloud companies reported earnings within a few days of each other. All three saw cloud revenue grow quickly. But the same announcements also showed, in numbers, the burden of AI infrastructure investment.

Amazon reported third-quarter revenue of $201.9 billion, up 12% from a year earlier. According to a report by Electronic Times source, AWS revenue grew 39%, its highest growth rate in about five years. The same report said the free cash flow deficit widened as AI infrastructure investment increased source. Free cash flow is the cash earned from operations minus capital spending. Alphabet was similar. AI Times source reported that Google Cloud revenue grew 103% from a year earlier, and Ars Technica source reported that Google's cash flow deficit grew to $9.06 billion.

The reason both numbers grow together can be read as a timing gap. Cloud revenue comes from data center capacity already built. Capital spending goes out first for capacity to be sold later. The greater the demand, the more a company builds in advance, and the more its cash on hand shrinks, it can be argued.

For practitioners, this means it is hard to judge a provider's situation from growth rate alone. Revenue and cash run on different clocks.

How different are the growth rates of the three clouds?

All three companies posted high growth rates, but the gaps between them are wide. Placing the growth rates side by side makes them easy to read as a ranking. They only become meaningful when viewed together with scale.

Microsoft reported revenue of $92.2 billion for the first quarter of fiscal 2027 (July–September 2026), up 19% from a year earlier (The Electronic Times, s10). According to TechCrunch source, Azure revenue grew 48%, and the company disclosed Azure's quarterly revenue of $33.2 billion for the first time. It added that its commercial dealings with OpenAI, including revenue sharing, produced $10 billion in revenue in a single quarter (TechCrunch, s9). AWS third-quarter revenue was $46 billion, up 39.3% from a year earlier (Artificial Intelligence Newspaper, s7).

With Azure's revenue now disclosed, AWS and Azure can be compared on the same basis. Google Cloud, which has the highest growth rate, may be growing from a relatively small base, while AWS's 39% at its much larger scale can be read as carrying a different weight in absolute terms. Another feature of this round of results is that the share of quarterly revenue coming from a single customer relationship is now visible as a number.

For readers choosing a provider, a more practical criterion than growth rate is when, and how much, of the capacity they need they can actually get.

Where does the investment money come from, and how is the payback explained?

Continuing to invest takes money. In this week's news, each company gave a different account of where that money will come from and when it will be paid back.

According to The Electronic Times source, citing Bloomberg, Microsoft plans to raise at least $25 billion in investment-grade corporate bonds in the US bond market. It is Microsoft's first corporate bond issue in the US since 2017, split into 10 tranches with maturities from 2 to 40 years source. Amazon CEO Andy Jassy said AWS will be the first business to reach $1 trillion in annual revenue, and that capital expenditure must keep expanding until then (TechCrunch, s6). Jassy also projected that AWS growth will exceed 40% in each of 2027 and 2028 (Artificial Intelligence Newspaper, s7). This is the company's own forecast. Google acknowledged that AI search earns less than traditional search and said it will make up the difference by increasing ads (Ars Technica, s3).

In short, three lines of explanation emerged: external borrowing, long-term growth forecasts, and shoring up existing revenue sources. The inclusion of 40-year bonds can be read as a funding structure that assumes a long payback period.

Practitioners should keep in mind that changes in a provider's revenue model can show up in pricing plans and service offerings.

What is the market trying to confirm, and what should practitioners watch?

According to SCMP source, Alphabet shares, after the company reported earnings on the 29th (local time), fell 12.6% in regular trading and pulled the Nasdaq Composite down 1.4%. Intel fell about 23% and Amazon 7.3% in after-hours trading source. Amazon's revenue and earnings per share missed market expectations (The Electronic Times, s1). The report attributed this to growing investor anxiety over swelling capital expenditure and delayed returns on investment source.

This reaction can be read as a sign that the metrics the market watches are shifting from revenue growth to payback timing and cash flow. The investment burden does not stop at chips. The UK Treasury projected that the UK will need about 6GW of AI data center capacity by 2030 and said it will apply standards to new data centers to manage power costs, environmental effects and local community impact (Artificial Intelligence Newspaper, s11). It is an example showing that data center costs also include power and site conditions.

Practitioners may want to review their contracts and cost structures on the assumption that capacity supply terms and pricing policies could change while providers continue to invest.