Big Tech Quad Kill earnings recap: Google Cloud surges 63%, Meta drops on CapEx concerns

Big Tech Quad Kill Earnings Recap: Google Cloud Surges 63%, Meta Drops on CapEx Concerns

Google's earnings crushed expectations, driven by two distinct narratives that together paint a picture of AI infrastructure acceleration. The core search business remains durable—Google Search and other revenue grew 19% year-over-year, decisively answering the question of whether AI chatbots are cannibalizing the search moat. The real story, though, is Google Cloud. At $20 billion in revenue growing 63% year-over-year, Google Cloud has become so material that analyst commentary on the earnings season pivoted entirely to cloud metrics: GCP at 63%, AWS at 28%, Azure at 40%. Google Cloud's operating income reached $6.6 billion, showing the unit is not just growing fast but generating real margin. The backlog signal matters more. Google Cloud backlog nearly doubled to more than $460 billion, and critically, more than half is expected to be recognized within 24 months—this is near-term demand, not multi-year noise.

The stock is up 10%. The CapEx bet is working.

Microsoft's enterprise play is executing, not transforming. Revenue hit $82.9 billion, up 18% year-over-year—a clean beat on headline numbers. The stock slipped 2%, a muted reaction despite solid results. The tension is adoption velocity. Microsoft has 450 million paid Microsoft 365 seats. Copilot, the AI agent layer, sits at 20 million seats. The math is stark: if every seat eventually gets a Copilot subscription, the company goes from 20 million to 450 million TAM. The question is time. How quickly do enterprises adopt?

The OpenAI renegotiation introduces a structural tension. Microsoft lost exclusivity when OpenAI became available on AWS. On the sales side, that's painful—Azure can no longer be the exclusive channel for GPT models. On the equity side, it's neutral to positive, since more OpenAI adoption means more value flowing to Microsoft's stake in the company. Satya Nadella renegotiated the deal, and by most accounts it landed cleanly. The AGI clause that allowed OpenAI to stop sharing models is gone. The deal extends to 2032 with revenue sharing and IP provisions. The market digested this as a resolved risk, not a dealbreaker.

Amazon executed on the hardest part of cloud growth at scale. AWS grew 28% year-over-year, beating the 25% expectation—material for a unit that size. Amazon overall hit $181.5 billion in quarterly revenue, putting the company on track for $1 trillion annual run rate. The ads business churned $17.2 billion. AWS as a profit engine is reaccelerating despite relentless CapEx investment. Amazon is agnostic to whose models run on its infrastructure. They partner with both OpenAI and Anthropic without the competitive tension a Microsoft or Google faces with their own frontier models. That positioning—being the infrastructure play—justifies higher CapEx because the demand signal comes from multiple AI labs, not just internal consumption. The market moved slightly upward. The CapEx king gets a pass when AWS beats on growth and margin.

Meta faces a credibility gap between business strength and narrative clarity. Revenue hit $56.3 billion, up 33% year-over-year. Ad impressions rose 19%, price-per-ad rose 12%—the core advertising machine is working. But the stock dropped 9% on CapEx guidance. Meta raised its full-year CapEx range from $115–$135 billion to $125–$145 billion, a $10 billion upward shift on both ends. The question that haunts investors: is Meta buying more compute, or just paying more for the same compute as input costs inflate?

The deeper issue is narrative risk. Google, Microsoft, and Amazon each have a story for why CapEx drives cash flow. Google Cloud sells infrastructure directly. Microsoft sells Copilot seats and enterprise AI software. Amazon sells cloud services to OpenAI, Anthropic, and others. Meta has no such direct revenue lever. CEO Mark Zuckerberg framed the investment as a long-term bet on frontier AI—the company is "recursive self-improvement pilled," as one analyst put it. The pitch is speculative: if Meta's internal AI research pays off and the company builds a competitive frontier model, it becomes a player in an AI oligopoly alongside OpenAI and Anthropic. That's a call option, not a cash flow story.

One earnings note: Meta's daily active people metric declined sequentially for the first time since the company began reporting it. Meta attributed this to internet disruptions in Iran and WhatsApp restrictions in Russia. There's no evidence of a new social platform competing with Meta's family of apps. TikTok, Snapchat, and Sora all failed to dent Meta's core user base. The risk is structural—that LLMs will generate content for existing platforms rather than spawn new ones, which would entrench Meta further.

The AI narrative is fracturing into distinct strategies. Over the last year, the story was simple: bigger CapEx, bigger deals, bigger growth. Each hyperscaler now reports massive CapEx, but the justification varies. Google is the full-stack platform, selling cloud infrastructure and models. Microsoft is the enterprise distribution play, leveraging 450 million Microsoft 365 seats to sell AI software. Amazon is the agnostic infrastructure play with partnership momentum. Meta is the ad optimization machine with a speculative bet on frontier AI. The market is sorting between near-term cash flow evidence and long-term optionality. This quarter's filter: can you justify your CapEx?

One check against a .com bubble narrative: the MAG 4 trades at 16–25x forward earnings while generating hundreds of billions in annual cash flow. During the .com peak, Microsoft traded at 73x earnings, Cisco at 200x+, Yahoo at 800x. The Nasdaq as a whole sat at 200x. Today's giants are priced conservatively in absolute terms. The caveat is real: private AI startups with no earnings are marked at 20–500x multiples. AMD trades at 130x, Tesla at 350x, Palantir at 220x, Intel at 900x. The bubble, if it exists, is in the tail, not the core. The core is anchored.