Alphabet and Amazon Pull Ahead of Meta in the AI Race as Quarterly Earnings Confirm Cloud Strength

Quarterly earnings from Alphabet, Amazon and Meta showed where AI investment is already turning into revenue and where the market still sees mainly rising costs. Alphabet and Amazon looked like the day's big winners: their cloud units accelerated as demand for generative AI compute grew.

Where AI is already paying off

At Alphabet, the key signal was Google Cloud. Cloud revenue rose 63% year over year to $20.02 billion, well above analyst expectations. The company says outright that demand is now limited not by customer interest but by available capacity: quarterly capital expenditure reached $35.7 billion, and the full-year 2026 capex forecast was raised to $180–190 billion. The cloud backlog exceeded $460 billion and nearly doubled over the quarter.

"In the near term, we are constrained by compute capacity," is how

Sundar Pichai described the state of cloud demand.

Amazon's picture is similar, but on a larger scale. AWS grew 28% to $37.6 billion in revenue, its fastest growth in 15 quarters. The company also reported EPS of $2.78 against a consensus of $1.62, and its chip business passed $20 billion in annual run rate. The downside is the same as for its rivals: infrastructure is expensive. Free cash flow over the trailing 12 months fell to $1.2 billion, mainly because of the sharp rise in AI infrastructure spending.

Why the market punished Meta

Operationally, Meta reported solid results: quarterly revenue was $56.31 billion, above expectations, and its second-quarter guidance of $58–61 billion is broadly in line with market consensus. But the number investors focused on was capex, not revenue. Meta raised its 2026 target to $125–145 billion from the previous $115–135 billion, citing more expensive components and additional data center costs. After the release, the shares fell about 6% in after-hours trading.

The difference from Alphabet and Amazon lies in how the market reads this spending. The growth of Google Cloud and AWS reflects external demand: thousands of enterprise customers already pay for models, accelerators, inference and agent services. At Meta, most AI spending still works inward, on recommendation systems, generative ad tools and the Llama model family. That matters strategically for the business, but in the short term it looks more like a cost center than a separate revenue line.

Key numbers of the quarter

The main takeaway from this round of earnings is not that companies are spending abnormally much, but that the market has begun to judge equally aggressive capex differently. Where extra data centers quickly convert into cloud and enterprise AI sales, investors are willing to tolerate enormous bills. Where the infrastructure mainly serves internal products, the reaction is noticeably harsher. That is why the size of the investment itself is no longer the main story.

• Google Cloud grew 63% and brought in $20.02 billion in quarterly revenue for the first time
• AWS grew 28% to $37.6 billion, its fastest pace in 15 quarters
• Alphabet raised its 2026 capex forecast to $180–190 billion
• Meta raised its capex forecast to $125–145 billion and got a negative market reaction
• Combined AI spending by the five hyperscalers in 2026 may now exceed $650 billion

This matters more than the quarterly surprises themselves. A year ago, the market was debating whether a new bubble was inflating around AI construction. The answer now looks more concrete: demand for compute is real, but the financial return is distributed unevenly. Where AI is sold as a cloud service, the effect already shows up in revenue. Where companies build mostly for themselves, investors still demand more patience. This is where the new line runs between growth and simply an expensive construction project.

What it means

The first wave of the AI race is shifting from a "build at any cost" mode to a "show me the money" mode. Alphabet and Amazon are already demonstrating that the cloud has become the main channel for monetizing the AI boom, while Meta will have to prove that its giant investments in its own infrastructure will also deliver a comparable return.

*Meta is designated an extremist organization and banned in Russia.