Google Cloud's Q2 Revenue Jumps 82% to $24.8 Billion as AI Demand Rips Through the Ceiling
Alphabet reported second-quarter results on Wednesday that blew past Wall Street's expectations, powered by a stunning 82% year-over-year surge in Google Cloud revenue to $24.8 billion. The cloud unit's backlog — contracted revenue not yet recognized — hit $514 billion, up 390% from the $106 billion it reported in the same quarter last year.
The numbers tell a clear story: AI infrastructure spending is no longer hypothetical. It's happening right now, at a scale that's reshaping the entire cloud computing market.
"I've been saying for years that Google Cloud's growth rate deserves more attention than it gets," said Bob Evans, founder of Cloud Wars, in his analysis of the quarter. "After this Q2, there's no more room for doubt."
Google Cloud's $4.8 billion sequential revenue jump — the quarter-over-quarter increase alone — became a new benchmark that AWS and Microsoft Cloud will have to match when they report their own numbers later this month.
What's Driving the Acceleration
The headline number — 82% growth — is remarkable enough on its own. But the composition of that growth matters more. CEO Sundar Pichai broke it down on the earnings call: demand for AI infrastructure, enterprise AI solutions, and core Google Cloud Platform services are all firing at once.
"Cloud revenue grew 82%, powered by strong demand for AI infrastructure and AI solutions," Pichai said. "It's great to see the wide adoption of Gemini Enterprise, with nearly 90% of the Fortune 100 using it."
Gemini Enterprise, the company's AI assistant for corporate workers, now starts at $21 per person per month and has found its way into PepsiCo for analytics, Intel for process automation, HSBC for wealth management, and Macy's for commerce. Almost 500 cloud customers have each consumed over 1 trillion tokens in the last year, Pichai noted.
The Gemini App itself now has 950 million monthly active users, with daily active users tripling over the past year. Models process 22 billion API tokens per minute, up from 16 billion last quarter. And the company's Antigravity coding platform crossed 2.4 million weekly active users.
Capex Spiral: $195 Billion and Climbing
Google's gargantuan spending on AI infrastructure dominated the earnings call from a different angle. The company raised its 2026 capital expenditure forecast for the second quarter in a row, now expecting between $195 billion and $205 billion for the full year.
That's up from the $180 billion to $190 billion range it gave just three months ago.
"Given the supply constraints environment, we plan to expand the use of third-party capacity in Q3 as a bridging strategy while we build up more internal capacity," CFO Anat Ashkenazi told analysts. "It will create modest margin pressure in the near term."
The company spent $44.9 billion on capex in Q2 alone, roughly 100% more than a year ago. About 60% went to servers, 40% to data centers and networking gear.
Google also acknowledged it's now renting outside compute capacity — a striking admission from a company that builds its own TPUs and runs one of the world's largest infrastructure networks. A June deal with Elon Musk's SpaceX calls for $920 million per month in AI compute. Ashkenazi said third-party capacity serves as a "bridging strategy" while Google builds out its own internal capacity, which will take months.
The spending isn't expected to slow next year. Ashkenazi warned in April that 2027 capex would "significantly increase" from 2026 levels. Evercore analysts have penciled in $262 billion for 2027; the FactSet consensus sits around $257 billion.
What This Means for the Cloud Wars
Google Cloud's $24.8 billion quarter places it firmly in the conversation with AWS and Microsoft Azure, even though those two remain larger in absolute revenue. The gap is narrowing fast.
Last quarter, Google Cloud grew 63% year-over-year. AWS grew 28%. Azure grew roughly 40%. This quarter, Google accelerated to 82% while analysts had only modeled 64%. The $4.8 billion sequential revenue increase is the kind of number that changes market-share trajectories.
Part of that growth came from Google's $29.5 billion acquisition of cloud security company Wiz, which closed earlier this year and is already contributing to revenue. Security has become one of Google Cloud's main differentiators, alongside its AI portfolio.
"Overall, we believe Google Cloud is the best-positioned of the three major clouds to take AI-market share over the next few years," wrote a team of Evercore analysts in a recent note, pointing to the company's TPU supply, deep integration of Gemini across its stack, and growing enterprise customer base.
Google also released several new AI models this week aimed at holding its position: Gemini 3.6 Flash for coding with 17% fewer tokens, and Gemini 3.5 Flash Cyber for automated vulnerability patching — a direct answer to Anthropic's security lead. The company has started its "most ambitious pre-training run yet" for Gemini 4.
The Stock Market's Mixed Reaction
Despite the blowout cloud numbers, Alphabet's stock dipped in after-hours trading. The reason: investors are worried about the sheer scale of the capex buildout and what it means for margins.
Gene Munster of Deepwater Asset Management pointed to search revenue and operating margins as the dampeners. Google's search revenue of $63.3 billion came in just shy of the $63.4 billion StreetAccount estimate, and capital spending shows no sign of peaking.
Still, Munster argued the stock should rally. "The Cloud number is the most important number and it was a massive beat," he wrote on X.
Alphabet's total revenue hit $119.8 billion, up 24% year over year. Net profit reached $112.11 billion, inflated by a $99 billion gain on equity securities — largely reflecting the soaring valuation of its Anthropic stake. YouTube ad revenue rose 13% to $11.06 billion, topping estimates.
For the broader cloud industry, Google's performance confirms a pattern that's been building all year: AI demand is real, it's massive, and it's pulling a disproportionate share of growth toward the hyperscalers that can actually deliver compute at scale. The question hanging over the next quarter is whether AWS and Microsoft can match Google's trajectory — or whether Google Cloud is pulling away.
Google's AI Model Pipeline and the Gemini Push
Beyond the raw infrastructure numbers, Google is betting that its own AI models will keep customers hooked on its cloud. The company on Tuesday released three new Gemini models including Gemini 3.6 Flash for coding tasks, which uses up to 17% fewer tokens than its predecessor, and Gemini 3.5 Flash Cyber — a specialized model designed to find and patch software vulnerabilities, available initially to governments and trusted partners.
Pichai confirmed the company has started its "most ambitious pre-training run yet" for Gemini 4, investing compute now to compete with frontier labs like OpenAI and Anthropic. "We wanted to compete at the frontier level of where the frontier will be when Gemini 4 comes out," he said.
That model pipeline is already driving customer decisions. Nearly 500 Google Cloud customers have each consumed more than 1 trillion tokens in the last 12 months. More than 2,000 enterprises have burned through over 100 billion tokens each. The Gemini Flash series, which balances performance and cost, is seeing the most demand.
Google's open Gemma models — small enough to run on local devices — have been downloaded over 900 million times. The latest Gemma 4 models, released in April, account for 300 million of those downloads.
On the enterprise agent front, Google's Agent Development Kit hit nearly 70 million total downloads in Q2. The company frames this as the next phase of cloud competition: not just raw compute, but the platforms that let enterprises build AI agents atop the infrastructure.
Third-Party Capacity and the SpaceX Deal
One of the quarter's more telling details is Google's decision to rent outside compute. The company inked a $920 million per month deal with Elon Musk's SpaceX for Nvidia GPU capacity — a stopgap while it builds out its own infrastructure.
"This is the first time Google has had to publicly acknowledge it can't build fast enough," noted a Bernstein analyst in a client note. "It's a good problem to have, but it signals just how far demand is outstripping even Google's massive supply."
Ashkenazi said third-party capacity would create "modest margin pressure" in Q3 but argued the lifetime customer value justifies the short-term cost. "The incremental opportunity they are bringing to us, while a short-term cost over a few months may be very high, in the lifetime of the deal, as we bring more capacity on, is highly ROI positive," Pichai added.
Microsoft has been more aggressive on this front, renting from CoreWeave, Iren, Lambda Labs, Nebius, and Nscale. CoreWeave and Nebius shares jumped 4% and 5% respectively after Ashkenazi's remarks suggested Google would follow a similar playbook.
The net effect: all three major hyperscalers are now competing not just for customers but for the same finite pool of Nvidia GPUs and data center real estate. With Google's 2027 capex already flagged for another jump, the AI infrastructure arms race shows no signs of cooling down.
--
This article was updated with details from Alphabet's Q2 2026 earnings call. Read more about AI cloud infrastructure or explore our full Cloud & Edge Computing coverage. Source: CNBC - Alphabet Q2 2026 earnings