Four tech giants just reported quarterly earnings, and among all the headlines we could write, one idea stands out: three Big Tech firms are building a new moat — and one may become the biggest client. Alphabet, Microsoft, and Amazon are deploying capital like sovereign wealth funds, building chips, clouds, data centers, and supercomputers. And investors are rewarding them.
Meta, meanwhile, is trying to keep up — but without owning the infrastructure. Instead, it is increasing spending to rent it, and only now beginning to build its own capacity. Markets appear unconvinced without a clear AI strategy, and shares dropped 12%.
All four firms delivered exceptional results. But this quarter clarified something important for antitrust and regulatory observers: AI is splitting Big Tech into builders and buyers, and the “buyer” risks becoming a client-platform in a world where infrastructure is the moat.
The Builders: Alphabet, Microsoft, Amazon
If AI is the new electricity, these three are building the power grid. Read their earnings presentations and you’ll see it clearly: massive investment in AI, from data centers to custom chips, large language models, and — crucially — consumer-facing interfaces like apps, browsers, and AI agents that deliver this power directly to users and enterprises.
With record backlogs and a strong market reaction, this bet is paying off. But it comes with regulatory consequences. The prize for these companies is far too big — and the risks of antitrust intervention far too small and uncertain — for regulation to meaningfully influence strategy at this stage.
Alphabet: Generative Search + Massive Capex
Alphabet posted $102.3 billion in revenue, up 16%, and almost $35 billion in net income, even after a large EU fine — which had virtually no impact on the P&L. CAPEX is set to reach $91–93 billion this year, almost entirely for AI compute and data centers.
Search is being rebuilt around Gemini — AI Overviews and AI Mode are not experiments, they are the future. “Our full-stack approach to AI is delivering strong momentum,” Google’s CEO Pichai said.
Revenue growth is not just coming from ads anymore. Cloud’s 34% year-on-year jump to $15.2 billion sets a faster pace than AWS and Microsoft Azure this quarter, driven by AI Infrastructure and Generative AI Solutions. Google Cloud’s operating income more than doubled to $3.6 billion, proving the business has hit profitable scale without pulling back on growth.
Microsoft: A Planet-Scale Compute Strategy
Microsoft Cloud revenue reached $49 billion, up 26%, with Azure growing 40% — driven by generative AI workloads. The company will double data center footprint in two years, with $34.9B in capex this quarter alone and more coming.
Satya Nadella described Microsoft as a “planet-scale cloud and AI factory”, adding that demand still exceeds capacity. The OpenAI partnership now includes exclusive rights and $250B in Azure commitments, locking the supply chain into Microsoft’s orbit for years.
If you look at the table below, you only need to find the word “cloud” to see where the biggest growth is happening. And if you notice that the first lines don’t say “cloud,” those refer to the recent €250B partnership with OpenAI — so they count too.
Amazon: The Silent AI Superpower
Amazon delivered $180.2 billion in revenue, up 13%. AWS sales rose 20% to $33 billion, and AI adoption is surging. Amazon added 3.8 gigawatts of power capacity and is deploying chips (Trainium2, Graviton4) while building mega-clusters to train Claude.
Capital intensity is soaring — property and equipment spend hit $35.1B in 9 months — driven by data center expansion and silicon strategy.
It launched Project Rainier, a cluster with nearly 500,000 AI chips, and is scaling agentic software for enterprise, developers, and retail. The message is similar to Alphabet and Microsoft, just less theatrical: Amazon is building the global AI mercantile infrastructure — clouds, chips, fulfillment, logistics, satellites.
It is not surprising that Amazon’s shares rose 11% after publishing 3Q given not only the good numbers but the logic on its AI strategy.
Meta’s AI Bill Without the Moat
Meta’s quarter was strong on paper: $51.2B revenue, $25B in operating profit from the Family of Apps, huge usage growth, and nearly 1B monthly users for Meta AI.
But the tone — and market reaction — differed sharply.
Meta’s AI vision is expensive, ambitious, but still vague. It is investing heavily in compute and talent, yet it doesn’t own the full stack. Unlike its peers, it cannot rely on proprietary chips or hyperscale cloud to amortize the explosion in AI demand.
Instead, Meta risks becoming one of the largest customers of Amazon, Microsoft, and Google. The company has pledged to build more of its own capacity, but it will still rely on external providers for a significant portion of its needs. Building superintelligence labs is admirable — but doing so on someone else’s infrastructure?
And there is another significant difference between Meta’s earnings call and those of the other three: the tone on regulatory risk. While Google and Amazon treated European Commission and FTC actions as routine — a cost of doing business — Meta explicitly warned investors about more serious regulatory headwinds:
“We cannot rule out the Commission imposing further changes… that could have a significant negative impact on our European revenue, as early as this quarter.”
Meta explicitly warns of near-term European action on its Less Personalized Ads model and litigation risk in the U.S. Alphabet shrugged off another €3.5B EU hit. Microsoft talks about capacity constraints, not the DMA. Amazon booked a $2.5B FTC settlement same quarter — and moved on.
Regulatory friction is not equal. Meta is still defending its old moat, just as regulatory pressure intensifies and enforcers close in. Meanwhile, the other three are racing to build a new AI-infrastructure moat — and they are not slowing down to wonder whether regulators will approve. At this point, the scale of the opportunity dwarfs the regulatory risk.
Meta Net Income (in million $)
