Cloud at 30% Margin, Apple Picks Gemini, and Capex Doubles Again — A Generational Print With a Generational Bill

Results vs. Consensus

Quality of the Beat

• Revenue: Strongest of the year. Search +17% (vs. +15% Q3, +12% Q2, +10% Q1) is the most meaningful acceleration — Search is GROWING FASTER as AI Overviews and AI Mode scale. Cloud accelerated to +48% (from +34% in Q3) — a genuine inflection driven by capacity coming online and Gemini 3 demand.
• Operating margin: 31.6% headline includes the $2.1B Waymo SBC charge (~190bps drag). Excluding it, margin would have been ~33.5% — strong but slightly compressed from Q3's 33.9% ex-EC fine. Services margin of 41.9% is the cleanest operational read.
• YouTube ads miss: +9% Y/Y vs. consensus of +13-14% reflects the lapping of strong 2024 U.S. election spend, especially on YouTube. Excluding the election comp, underlying YouTube growth is in the low-teens — consistent with prior-quarter trajectory.
• Cloud margin: 30.1% is a watershed number. The Q3 print at 23.7% suggested 25-27% by year-end; instead Q4 jumped to 30.1%, meaning Cloud is now operating at a higher segment margin than many of Alphabet's pre-AI Services businesses.
• FCF: Q4 $24.6B is healthy; FY $73.3B reflects $91.4B of capex absorbing a meaningful portion of operating cash flow. The tax law expensing change continues to be a positive contributor.

Segment Performance

Google Services — Search Reaccelerates Again on Gemini 3

Search & other revenue grew 17% to $63.1B — the fourth consecutive quarter of acceleration (10% → 12% → 15% → 17%). Two distinct catalysts drove the Q4 step-up: (1) Gemini 3 Pro, which launched in December and now powers AI Overviews and is "consistently processing 3x as many daily tokens on average as 2.5 Pro" since launch; (2) AI Max in Search continuing to expand monetization on previously-difficult-to-monetize long-tail queries.

"Search saw more usage in Q4 than ever before, as AI continues to drive an expansionary moment. We have executed with incredible speed; we shipped over 250 product launches within AI Mode and AI Overviews just last quarter." — Sundar Pichai, CEO

"The launch of Gemini 3 was a major milestone... Gemini 3 Pro has consistently processed three times as many daily tokens on average as 2.5 Pro." — Sundar Pichai, CEO

The behavioral data continues to support the bull thesis. AI Mode daily queries per user doubled since launch. Queries in AI Mode are now 3x longer than traditional search. ~1 in 6 AI Mode queries are non-text (voice or images). Circle to Search is on 580M+ Android devices.

Schindler walked through three pillars of AI ad monetization: (1) Gemini-powered ad quality improvements ("almost a launch per month for the last two years"); (2) Gemini-powered advertiser tools (Q4 alone saw nearly 70M creative assets generated via AI Max and PMax); (3) New AI user-experience monetization — "We have significantly increased our focus on AI Mode and are in the early stages experimenting with AI Mode monetization, like testing ads below the AI response."

YouTube ads grew 9% to $11.4B — light vs. consensus, but the lapping of 2024 U.S. election spend was a known and well-flagged headwind. Subscriptions/Platforms/Devices grew 17% to $13.6B; YouTube annual revenue exceeded $60B for the first time across ads + subs; 325M+ paid subscriptions across consumer services.

Assessment: Four consecutive quarters of accelerating Search growth in the face of the most aggressive competitive AI rollouts in tech history is the most decisive single piece of evidence against the cannibalization thesis. The Gemini 3 launch creates a fresh adoption curve that should sustain the pace into 2026.

Google Cloud — 30% Margin, $240B Backlog, and the Apple Anchor

Cloud's transformation in Q4 is the financial story of the quarter. Revenue accelerated to +48% (from +34% in Q3) to $17.7B. Operating margin expanded to 30.1% from 17.5% Y/Y — a +1,260bps Y/Y expansion. Operating income more than doubled to $5.3B.

• Backlog: $240B at quarter-end — up $85B sequentially from $155B (+55% Q/Q), and more than double Y/Y. This is the largest quarterly addition to Cloud backlog in Alphabet's history.
• $1B+ deals in 2025: Surpassed the prior three years COMBINED.
• New customer velocity: Doubled vs. Q1.
• Existing customer expansion: Outpacing initial commitments by 30%+.
• 14 product lines at $1B+ annual run-rate (vs. 13 at Q3).
• ~75% of Cloud customers use AI products; AI customers consume 1.8x as many products as non-AI customers.
• December alone: ~350 customers each processed >100B tokens; Q4 generative-AI-built product revenue +400% Y/Y (vs. +200% in Q3).
• 120,000+ enterprises use Gemini, including 95% of top 20 SaaS and 80%+ of top 100 SaaS companies.

And the marquee announcement:

"I'm pleased that we are collaborating with Apple as their preferred cloud provider and to develop the next generation of Apple Foundation Models based on Gemini technology." — Sundar Pichai, CEO

Apple selecting Gemini for its next-gen Apple Foundation Models is a strategic earthquake. Apple has over 2 billion active devices; Apple Intelligence runs across them; the Foundation Models are the brain. Apple's pre-Q4 stance (small partial OpenAI integration, parallel Anthropic talks) flips to a primary Gemini partnership. This deal probably represents tens of billions of dollars of TPU/Cloud commitment over its life, and more importantly, it brings Gemini distribution into the Apple consumer ecosystem.

Reliance Jio partnership announced separately: 500M+ Indian consumers receive an 18-month free trial of Gemini products and 2TB cloud storage; Reliance Enterprise gets Gemini Enterprise + TPUs. Distribution unlock at scale in the world's most populous AI market.

"95% of the top 20 and over 80% of the top 100 SaaS companies use Gemini, including Salesforce and Shopify. Gemini is becoming the AI engine for the world's most successful software companies." — Sundar Pichai, CEO

Gemini Enterprise sold 8M+ paid seats across 2,800 companies in just four months since launch. Customer interactions managed by Gemini Enterprise: 5B+ in Q4 alone, up 65% Y/Y, for customers including Wendy's, Kroger, Woolworths, BNY, Virgin Voyages.

Assessment: The Cloud thesis is now substantially de-risked through 2027. With $240B of backlog (~13.5x Q4 revenue) and the Apple anchor, the question is no longer "can Cloud scale" but "can capacity be built fast enough." The 30.1% operating margin places Cloud above many large-scale infrastructure peers and validates the multi-year capex investment.

Other Bets / Waymo — The Funding Round Crystallizes Value

Waymo announced a $16B investment round (the "significant majority" funded by Alphabet) just days before the call. The Q4 results include a $2.1B SBC charge tied to the Waymo valuation step-up — primarily reflected in R&D expense. Operationally:

• 20M+ cumulative fully autonomous trips (passing the milestone in December).
• 400,000+ rides per week (vs. 250K/week at Q1) — 60%+ growth in 9 months.
• Sixth market: Miami launched.
• Coming soon: London, Tokyo, multiple additional U.S. cities.
• Public service to airports and freeways "opening up."

Assessment: The $16B funding round at an implied valuation step-up makes the Waymo option substantially more visible to the market. Combined with 400K weekly rides and 8+ city expansion, Waymo is becoming material to the Alphabet sum-of-parts. The $2.1B SBC charge is non-cash and one-time-in-character. We continue to view Waymo as upside optionality that we don't underwrite explicitly but is increasingly hard to ignore.

Key KPIs

Key Topics & Management Commentary

Overall Management Tone: Confident, ambitious, and unmistakably on offense. Pichai opened with "tremendous quarter" and proceeded to detail the most aggressive capex ramp in company history. The language framing 2026 capex was deliberately structural — "investments required to further capitalize on the AI opportunity" — not opportunistic. Ashkenazi was the most disciplined, repeatedly grounding the spend in "supply availability, pricing of components, and timing of cash payments." There was no hedging on the Apple deal — it was front-and-center, named twice on the call, with both Pichai and Schindler giving it dedicated airtime.

The 2026 Capex Reset — $175-185B

This is the singular new variable in the Alphabet equity story. Q4 2025 capex was $27.9B; full year 2025 was $91.4B; 2026 guide is $175-185B — implying quarterly capex averaging ~$45B. The Street was modeling roughly $120B; the guide is $55-65B above consensus.

"Our 2026 CapEx investments are anticipated to be in the range of $175 to $185 billion." — Sundar Pichai, CEO

"We're investing in AI compute capacity to support frontier model development by Google DeepMind, ongoing efforts to improve the user experience and drive higher advertiser ROI in Google services, significant cloud customer demand, as well as strategic investments in Other Bets." — Anat Ashkenazi, CFO

The mechanical implications: 2026 D&A will increase "meaningfully" beyond the +38% growth in 2025 ($21.1B). At a baseline 35-40% asset useful life and 2025+2026 capex weighted late-2026, D&A expense alone could exceed $35-40B in 2026 — a $14-19B headwind to operating income.

The offsets:

• Cloud at $70B+ run-rate growing 48%, with margin at 30.1% — could exit 2026 at $100B+ run-rate.
• Gemini serving costs declined 78% in 2025 — the unit economics of AI inference are improving rapidly.
• $240B backlog provides revenue visibility into 2026-2028.
• Apple deal adds a multi-year anchor customer.
• Tax law (R&D expensing, accelerated depreciation) continues to provide cash flow tailwind.

Assessment: The capex shock is real, and the stock will likely de-rate near term as the Street recalibrates 2026 EPS and FCF. But the math is defensible if Cloud sustains its current trajectory. The risk-reward turns asymmetric on demand — if Cloud grows another 30%+ in 2026, the capex was right; if growth softens to 25%, the depreciation drag overwhelms.

The Apple Deal

Apple selecting Gemini as the foundation for its next-generation Apple Foundation Models is the most meaningful strategic announcement Alphabet has made in years. The deal positions Google as Apple's preferred cloud provider AND model partner — historically Apple has built Foundation Models in-house (with limited OpenAI integration). The shift to Gemini implies Apple judged Gemini superior for its on-device + cloud Foundation Model architecture.

Strategic implications:

• Distribution: Gemini gets access to ~2B Apple devices, the wealthiest consumer base in tech.
• Competitive: OpenAI and Anthropic lose what was their most likely Apple integration path.
• Economics: Multi-year Cloud+licensing revenue, likely $5B+ annually at scale.
• Validation: Apple's brand reinforces the Gemini quality narrative — alongside Anthropic, SSI, Salesforce, Shopify.

Assessment: The Apple deal materially shifts the AI competitive landscape. Combined with Anthropic's TPU commitment (Q3) and the SaaS-100 penetration, Gemini has effectively become the consumer-AI infrastructure layer for the largest distribution surfaces outside of Microsoft. The move is incremental to our model — we add ~$3-5B of annualized Cloud revenue from the Apple deal starting H2 2026.

Gemini 3 — A Discontinuous Adoption Curve

Gemini 3 launched in December and is "the fastest adoption of any model in our history." Specifically, Gemini 3 Pro processes 3x daily tokens of 2.5 Pro since launch. The Gemini app reached 750M MAU (from 650M in Q3), and "significantly higher engagement per user" since the Gemini 3 launch.

Importantly, Gemini 3 was integrated into AI Mode AND AI Overviews in early 2026 — meaning Search Q1 2026 will get a fresh boost from the model upgrade. AI Overviews "upgraded to Gemini 3 last week" before this earnings call. Project Genie 3, Imagen, VO, Liria — the model portfolio is the broadest in the industry.

Google Antigravity (autonomous coding agent platform) launched in November and reached 1.5M weekly users in two months. Universal Commerce Protocol launched at NRF in January with retailer partners.

Assessment: The Gemini 3 launch creates the conditions for Search and Cloud growth to sustain or accelerate further in Q1 2026. We expect Q1 2026 Search growth in the +18-20% range and Cloud growth in the +48-52% range.

The Waymo $2.1B SBC Charge

The $16B Waymo funding round triggered a $2.1B stock-based compensation charge in Q4 (mostly to R&D), as Waymo employee equity was repriced to the new round valuation. This is non-cash but reduces reported operating income by ~190bps. Without the charge, operating income would have been ~$38B and operating margin ~33.5%.

Assessment: Mechanically dilutive on a one-quarter basis, strategically validating: the $16B round implies a Waymo valuation step-up that is meaningful to the Alphabet sum-of-parts. We treat the charge as one-time and the funding as net positive.

Guidance & Outlook

Implied Q-over-Q ramp: Q1 2026 should see Search growth in the +17-19% range, Cloud growth in the +48-55% range, and consolidated revenue growth in the +14-16% range with FX tailwind. Capex will ramp through the year, weighted to H2 — implying Q1 capex of ~$35B and Q4 capex of ~$50B+.

Analyst Q&A Highlights

Capex & Capacity

• Eric Sheridan (Goldman): Asked about the pathway to closing the demand-capacity gap given the $175-185B 2026 capex. Pichai: "I do expect to go through the year in a supply constrained way" — even after the doubling. Ashkenazi reiterated efficiency focus across infrastructure, real estate, and 50% AI-generated code.
  Assessment: Pichai explicitly confirming continued supply constraint at $175-185B is the strongest bull signal possible. Demand visibility is exceptional.

LLM Frontier & TPU Distribution

• Doug Anmuth (JPMorgan): Asked about LLM leapfrogging risk and whether TPUs could move outside Google Cloud as a separate revenue stream. Pichai punted on TPU externalization — "I would think about it as part of what makes Google Cloud an attractive choice." On model leadership, he expressed confidence in continued progress across pre-training, post-training, test-time compute, multimodality, and agentic capabilities.
  Assessment: TPU externalization (selling TPUs to AWS/Azure customers, or as standalone hardware) is the next $50-100B optionality. Management is not signaling it but is not ruling it out either.

Agentic Commerce & YouTube Genie

• Brian Nowak (Morgan Stanley): Asked Pichai about agentic commerce progress and YouTube/Genie integration. Pichai framed 2025 as the foundation year (model/protocol building) and 2026 as the year consumers experience agentic commerce in production. Universal Commerce Protocol is the underlying standard. Genie/Veo/Imagen will be tools for YouTube creators with creators staying "at the center."

Search Monetization Future

• Schindler in his prepared remarks: Disclosed that AI Mode monetization is "in the early stages — testing ads below the AI response, with more underway." Direct Offers in AI Mode pilot launching with select retailers.
  Assessment: AI Mode at 75M+ DAU still doesn't have full ad monetization. The 2026 storyline is closing that gap. Even modest monetization expansion is a meaningful incremental revenue line.

What They're NOT Saying

1. 2027 capex: Pichai/Ashkenazi did not extrapolate beyond 2026. Implicitly the message is "$175-185B is the new normal, not a one-year peak" — but explicit 2027 commentary was avoided.
2. Apple deal economics: No revenue figure or contract duration disclosed for Apple. This is contractually expected (Apple notoriously keeps deal terms private), but the lack of even directional sizing leaves the Street to guess.
3. RPO duration on $240B backlog: Still not disclosed. With $85B added in Q4, the duration mix is critical for revenue forecasting and is being withheld.
4. TPU external sale strategy: Pichai declined to commit. The Anthropic 1M-TPU deal raises the question of whether Alphabet will eventually sell TPU capacity outside its own cloud.
5. DOJ Search remedy / Chrome: Zero call mentions for the fourth consecutive quarter. Management is choosing silence; the remedies process timeline implies decisions in mid-2026.
6. 2026 operating margin guide: No specific direction. With D&A "meaningfully" accelerating, margin will compress unless Cloud margin expansion fully offsets — which is plausible given 30.1% Q4 print.
7. Buyback pace: Q4 repurchases were just $5.5B — the lowest of 2025 (Q1: $15.1B, Q2: $13.6B, Q3: $11.5B, Q4: $5.5B). FY 2025 repurchases ~$45.7B. The Q4 slowdown coincides with capex ramp; this is the real near-term capital allocation signal — buyback pace is being modulated downward to fund capex.

Market Reaction

• After-hours move (Feb 4, 2026): Stock initially fell as much as 3% on the $175-185B 2026 capex guide, before paring losses to ~0.4-0.5% down by close of next trading day.
• Net read: The print itself was a clean beat; the capex guide was the source of selling pressure. Multiple sell-side analysts raised price targets despite the capex shock — recognizing that the operational fundamentals are stronger than ever, but trimming near-term FCF and EPS estimates to absorb the capex.
• Volume: Elevated; the print was the most-watched single-stock event of the week.

The muted decline (vs. what one might have expected from a $55-65B Street-vs-guide capex gap) is itself telling: investors recognize the capex is being raised in response to demand, not deployed speculatively. The $240B backlog is the offset; without it, the reaction would have been -10%+.

Street Perspective

Debate: Is the 2026 capex guide a buying opportunity or a structural rerating signal?

Bull view: Capex is purely demand-following. $240B of backlog (~13.5x quarterly Cloud revenue) provides multi-year visibility. Cloud at 30.1% margin proves the unit economics. The 2027-2028 EPS power is dramatically higher with this capacity in place than without.

Bear view: Doubling capex in one year inevitably outruns demand somewhere. The depreciation curve will pressure margins for 2-3 years. The Apple deal economics are unproven. AI capacity is at risk of an industry-wide oversupply if multiple hyperscalers ramp simultaneously.

Our take: The capex is justified at the moment. The risk is not "is the spending wise" but "will demand sustain 12-24 months out." The $240B backlog gives us 18-24 months of visibility; beyond that, we're in macro / AI-cycle risk. We treat near-term multiple compression as a buying opportunity.

Debate: Has Search permanently inflected higher?

Bull view: Four straight quarters of accelerating growth, Gemini 3 launch driving fresh adoption curve, AI Mode & AI Overviews monetization still ramping. Search +17% in Q4 is not a peak — it's the new run-rate.

Bear view: The acceleration partially reflects easy Y/Y comps; once the comps normalize in 2H 2026, Search will revert to high-single-digit growth. AI Mode has 75M DAU but is barely monetized — when it scales to AI Overviews scale, monetization will compress.

Our take: Bulls' case is stronger on the data. The 4-quarter acceleration with continued AI deployment is hard to reconcile with the bear thesis. We model Search at +14-16% in 2026 and +12-14% in 2027.

Debate: Is the Apple deal as transformative as it sounds?

Bull view: Apple has 2B+ active devices and chose Gemini as the foundation for next-gen Apple Intelligence. This is a multi-year, multi-billion-dollar commitment that locks in the largest consumer AI distribution surface outside of Microsoft.

Bear view: Apple deals are notoriously short-tenor and renegotiable. Apple is also building its own chips and may eventually replace Google Foundation Model dependency. The economics are unknown and may be modest.

Our take: The deal is meaningful even at moderate economics — it validates Gemini as best-in-class in Apple's view, which carries reputational weight with enterprise customers. We add $3-5B of incremental Cloud revenue from H2 2026, scaling to $8-12B by 2028.

Model Update Needed

Valuation impact: Near-term multiple likely compresses 1-2 turns on the FCF reset. Medium-term (12-18 months), the operational power of the business (Search +17%, Cloud at 30%+ margin growing 40%+) supports a re-rating once the capex digestion is visible. We treat the post-print pullback as a buying opportunity.

Thesis Scorecard Post-Earnings

Overall: The operational thesis is the strongest it has been since we initiated coverage. The capex thesis is now stretched and creates near-term FCF/EPS pressure. Net: thesis materially strengthened, with one significant new-variable risk that the market is currently digesting.

Action: Maintain Outperform. The capex pullback is a buying opportunity for investors with a 12-18 month horizon. Q1 2026 (April 29) will be the next data point — and the first test of whether the capex pace is sustained by demand.