DruckFin

Alphabet's Cloud Revenue Doubles to $30.7 Billion as TPU System Sales Turn Google Into a Hardware Vendor, and Free Cash Flow Goes Negative

Third-quarter 2026 earnings call, October 28, 2026

Alphabet reported Q3 2026 revenue of $127.6 billion, up 25%, but the more important story is a change in the shape of the business. Google Cloud revenue rose 103% to $30.7 billion, and management acknowledged for the first time how much of that growth comes from selling TPU systems rather than renting compute. The quarter also produced negative free cash flow of $9.1 billion, a result the company tied directly to building TPU inventory ahead of revenue recognition. Search grew 12%, a deceleration that management attributed partly to the lower monetization of AI Mode queries.

TPU system sales change the Cloud model and the cash profile

Management said TPU system sales were a "meaningful contributor" to Cloud growth in the quarter, though it said Google Cloud Platform still accounted for "the vast majority" of revenue. It declined to quantify the TPU contribution or give a margin for the product. CFO Anat Ashkenazi said only that growth excluding TPU systems "was again incredibly strong."

She laid out the mechanics. Once a customer signs, the contract appears in backlog, and Alphabet then builds inventory ahead of billing. "You see cash going out, and we're building up the inventory base. And then as we start delivering and billing customers for these TPU systems, then you see cash coming in," she said. This explains why operating cash flow was $42.3 billion in the quarter while free cash flow was negative. A $5.2 billion payment for a 2018 EC fine also weighed on the result. Trailing-twelve-month free cash flow fell to $19.7 billion against $179.6 billion of operating cash flow. Ashkenazi guided that TPU system sales will keep scaling in Q4, followed by "a meaningful revenue ramp in 2027," and that the inventory build will keep pressuring free cash flow.

Cloud chief Thomas Kurian described a deliberate widening of distribution. TPUs are now offered in customers' own data centers and through third parties, including a partnership with Oracle and Crux AI, a joint venture with Blackstone. He said Alphabet buys components only against signed contracts, so "all of the inventory for TPU systems is associated with committed contracts." He also argued that placing systems in third-party facilities "offsets the need for us to have to prebuild facilities." The customer list is broadening beyond Anthropic to Apple, which is scaling pretraining of its foundation models on the latest TPUs, as well as Citadel Securities, Jump Trading, Physical Intelligence and others. Nearly 100 customers now use TPUs. A new TorchTPU layer, built with Meta, lets PyTorch models run on TPUs, which lowers the main software barrier to adoption.

Management gave no margin disclosure for this business. Kurian said only that Alphabet is "very focused on continuing to improve margins as we scale." Investors should treat TPU system economics as an open question. Hardware resale typically carries lower margins than cloud services, and Alphabet is taking inventory and contract-execution risk it did not carry before. CEO Sundar Pichai framed the strategy as one of scale economics: "In the semiconductor industry, economies of scale matter." He also defended selling to rivals, saying Alphabet treats TPUs as a platform that must compete "against market forces."

Cloud backlog reaches $561 billion, but margins face near-term pressure

Cloud operating income more than tripled to $11.1 billion, and operating margin rose to 36.2% from 23.7% a year earlier. Backlog stood at $561 billion, mostly from typical GCP contracts, and Alphabet expects to recognize just over 55% of it over the next 24 months. Kurian said 18 product lines now exceed $1 billion in annual revenue, and 2,600 customers have each processed at least 100 billion tokens on Gemini Enterprise.

The margin trajectory is less favorable. Because Alphabet remains supply constrained, it began using third-party compute capacity at the very end of Q3 and will ramp that usage in Q4. Ashkenazi said, "We expect margin headwinds in the near term as we utilize this capacity." The 36.2% margin therefore likely reflects a peak for now, since third-party capacity contributed little to the quarter.

Capital intensity keeps climbing

Q3 capex was $51.4 billion, with about 65% going to servers and networking equipment that have useful lives of 6 years. Full-year 2026 capex guidance of $195 billion to $205 billion was unchanged, but Alphabet again said capex will "increase significantly in 2027," with details on the Q4 call. Future purchase commitments reached $924 billion, plus $152 billion of pre-commencement leases and other obligations. Management noted that some contracts carry pricing and volume flexibility that could reduce actual spend.

Cash and marketable securities were $265.1 billion, including $90.4 billion of marketable equity securities, against $125.1 billion of long-term debt. Net income and EPS declined year over year because other income and expense was negative $8.2 billion, mainly an unrealized loss on the SpaceX stake. Ashkenazi said Alphabet expects ROIC to stay "meaningfully above our cost of capital." The company did not give a timeline for when free cash flow recovers, and the combination of rising capex, inventory builds and $1 trillion-plus of commitments leaves little near-term cushion.

Search decelerates as AI Mode monetizes below legacy queries

Search and other revenue grew 12% to $63.3 billion, slower than Q2, even though Pichai said query growth was the highest in five years. Analyst Doug Anmuth of JPMorgan asked about the disconnect, and Chief Business Officer Philipp Schindler gave a more candid answer than in prior quarters. FX was "the single largest driver" of the slowdown. Other factors included APAC retail advertisers hit by the EU's removal of its de minimis exemption, and media and entertainment spend normalizing after a World Cup-driven Q2. Schindler also said "the strong growth in AI Mode queries had an impact because of the lower monetization we were seeing in AI Mode."

Ashkenazi added that Alphabet began lapping the Search acceleration of late 2025 in Q3, which cost about 3 points of growth, and that Q4 will see "a couple of points" of similar drag. She also guided to a slightly larger FX headwind in Q4, concentrated in Search and YouTube ads. Taken together, the guidance implies that Search growth will not rebound quickly.

The mitigation is to scale ads in AI Overviews and AI Mode. Alphabet is expanding AI Mode ads to international markets for the first time and adding multiturn ads in the U.S. Schindler said click-through rates on sponsored and organic product listings in AI Mode are already similar, and that most AI Mode queries are incremental. AI Mode queries run 3x longer than traditional searches, and 1 in 5 uses voice, images or documents. The monetization gap is real, and the company has not quantified how fast it expects it to close.

Gemini 4 Carbon launches as agentic products move toward consumers

Alphabet announced Carbon, the second model in the Gemini 4 series, which it says ranks first on the Artificial Analysis Intelligence Index at 59. It is going first to a subset of trusted cyber defenders and is being tested by the U.S. government. Argon, launched two weeks ago, saw first-week developer usage ramp 3.5x faster than Gemini 3 despite capacity constraints. First-party model APIs now process nearly 32 billion tokens per minute, up 40% sequentially and more than 4x year over year, and the Gemini app has reached 1 billion users. Pichai said Alphabet teams using Argon are shipping 35% more code with 30% fewer errors, and he expects "continued acceleration because of RSI," a reference to recursive self-improvement in model development.

On agents, Pichai said the Gemini agent for enterprise came first because of its higher security and governance bar, and that "later this year, we will bring that same experience to consumers, starting with the 1 billion users in our Gemini app." He said distilled versions of Carbon are being built to serve that scale at acceptable cost and latency. He was candid about the reliability threshold: "When you're doing it at 70%, it looks like it doesn't work until you cross a certain threshold, and it really takes off."

Schindler said agentic experiences are additive to advertising and argued that users want help with "the grunt work" but still value choice during discovery. The Universal Commerce Protocol now enables direct checkout for hundreds of thousands of brands, and Alphabet launched Universal Cart this quarter. Whether agent-mediated shopping preserves Search ad economics is unproven. Management offered a conceptual example of an agent presenting a sponsored alternative, not data. Pichai also acknowledged that safety and trust are "a very, very high factor" in how consumers adopt agents.

Services and other segments

Google Services revenue rose 11% to $96.3 billion, with operating margin of 40.7%. YouTube advertising grew 9% to $11.2 billion, helped by living-room viewing, which exceeded one-third of YouTube ad revenue. Network revenue grew only 1% to $7.5 billion. Subscriptions, platforms and devices rose 11% to $14.3 billion, supported by YouTube Premium and Music, which passed 150 million subscribers including trials, and by AI plans in Google One. Amazon is now joining YouTube Shopping.

Other Bets lost $2 billion on revenue of $372 million. Waymo has surpassed 300 million fully autonomous miles and operates in 15 U.S. cities, with planned launches in Canada, Germany, Singapore and Spain. Unallocated Alphabet-level costs, mostly shared AI R&D, produced a $7.7 billion operating loss. R&D expense rose 36%, driven by AI talent and depreciation. Total operating margin was 31.8%, with operating income up 30% to $40.6 billion.

Disclaimer: This article is for informational purposes only and does not constitute investment advice or a recommendation to buy, sell, or hold any security. Our analysts provide detailed coverage of corporate events but can make mistakes, always conduct your own due diligence. The views and opinions expressed do not necessarily reflect those of DruckFin. We have not independently verified all information used herein, and it may contain errors or omissions. Before making any investment decision, consult a qualified financial advisor. DruckFin and its affiliates disclaim any liability for any losses arising from reliance on this content. For full terms, see our Terms of Use.