Alphabet Inc. (GOOG) Earnings

Management highlights

### AI Product Development - Launched new Gemini models: Gemini 3.6 Flash, 3.5 Flashlight (optimized for cost and efficiency), and Gemini 3.5 Flash Cyber, which paired with CodeMender agent matches performance of far larger cyber models at lower cost. Gemini 3.5 Pro is in testing, and pre-training has started for the next-generation Gemini 4 frontier model, with early progress positive. - Token usage across developer and enterprise APIs reached 22 billion tokens per minute, up from 16 billion tokens per minute last quarter, with demand outstripping current supply capacity. Over 9 million monthly developers build on Google's AI models, and open-source Gemma models have been downloaded over 900 million times total, with 300 million downloads of the latest Gemma 4 since April. - The Omni generative video tool has driven a 40% increase in daily active creators on the Gemini app, which now reaches 950 million monthly active users, with DAUs tripling year-over-year. The AntiGravity agentic development platform has 2.4 million weekly active users, and has accelerated internal development timelines by up to 8x. ### Core Consumer Products - Google Search now has 1 billion monthly active users for AI mode, with search usage hitting an all-time high during the 2026 FIFA World Cup. AI features deliver billions of weekly clicks to third-party websites, and engineering optimizations have reduced the cost of AI mode responses to the lowest level since launch. - 1.7 billion global unique viewers watched World Cup-related content on YouTube, making it the most viewed World Cup in YouTube history. The Ask YouTube conversational AI feature had 140 million engaged users in June 2026. YouTube subscription revenue is growing faster than ad revenue, led by YouTube Music and Premium. ### Google Cloud & Enterprise AI - Nearly 90% of the Fortune 100 now use Gemini Enterprise, with strong diversified demand across industries, geographies, and customer segments. New customer acquisition velocity doubled year-over-year, existing customers expand commitments by over 50% on average, and Google Cloud Marketplace transactions grew 7x year-over-year. - 90% of the Fortune 100 use Google Cloud's AI-powered security platform, with a 45% quarter-over-quarter increase in AI workloads scanned and protected. New Google AI Threat Defense adds custom cyber models to defend against emerging AI-native threats. - Google Cloud offers a full-stack AI infrastructure portfolio including TPUs, NVIDIA GPUs, the Virgo unified networking platform for large-scale AI workloads, native open-source software support, and the agent-optimized Axion CPU, with strong demand from leading tech, pharma, and finance customers. ### Advertising Innovation - Gemini has improved ad relevance for long-tail queries, delivering a 20% improvement in relevance for shopping ads. The AI Max advertiser tool is out of beta, with over 500,000 advertisers already adopting it; adopters see an average 15% increase in conversion value at similar ROAS. - Over half of global SMB advertisers now use Google's AI tools to create and optimize ad creative. New ad formats are being tested in AI search mode, including contextual sitelinks, direct offers, and sponsored highlighted answers. Google launched the open Universal Commerce Protocol and Universal Card for cross-merchant single checkout. ### Other Bets Progress - Waymo launched its sixth-generation Ohai autonomous vehicle for public testing, with expanded rider access coming in the coming months. Wing delivery has completed over 1 million safe home deliveries, with growing partnerships with major retail and delivery brands. Isomorphic Labs raised over $2 billion to scale its AI drug discovery platform and advance its candidate pipeline.

Guidance

- **Q3 2026 Revenue Outlook**: A slight foreign exchange headwind is expected for consolidated revenue in Q3, compared to a 1 percentage point FX tailwind in Q2, with the impact concentrated in Search and YouTube ads. Q3 will lap the acceleration in search performance that began in Q3 2025. Google Cloud is expected to deliver strong continued growth on high existing demand. - **TPU Revenue Recognition**: Only a relatively small portion of total revenue from existing TPU system sales agreements will be recognized in 2026, with ramping recognition starting at the end of 2026; the vast majority of TPU revenue will be recognized in 2027. - **Capacity Strategy**: To meet unmet demand while building out internal capacity, Alphabet will use third-party contracted capacity as a short-term bridge in Q3. This strategy will support continued customer growth but create modest near-term margin pressure due to higher third-party capacity costs. The recent Wiz acquisition also creates modest near-term margin headwinds in 2026. - **Full-Year 2026 CapEx Guidance**: Capital expenditure guidance is revised upward to $195–$205 billion, up from the prior range of $180–$190 billion. The increase is driven by accelerated capacity delivery to meet growing AI demand. CapEx is expected to increase significantly again in 2027, with further details to be provided later. - **Expense and Cash Flow Outlook**: Higher technical infrastructure investment will continue to pressure profits through increased depreciation and energy costs for data centers. Alphabet will continue hiring in key AI and cloud roles, and increase marketing investment to support new AI products. Free cash flow will remain under pressure in the near term due to high AI infrastructure investment.

Segment performance

1. Google Services: Revenues increased 15% year-over-year to $94.5 billion, contributing 78.9% of Alphabet's total consolidated revenue. Within Google Services: - Google Search and other advertising: 17% YoY growth to $63.3 billion (52.8% of total revenue), driven by retail and finance verticals. - YouTube advertising: 13% YoY growth to $11.1 billion (9.3% of total revenue), driven by direct response and brand advertising, with strong performance in connected TV. - Network advertising: 1% YoY decrease to $7.3 billion (6.1% of total revenue). - Subscriptions, platforms and devices: 15% YoY growth to $12.9 billion (10.8% of total revenue), driven by YouTube subscriptions (Music and Premium) and Google One AI plans. Google Services operating income increased 20% YoY to $39.5 billion, with an operating margin of 41.8%. 2. Google Cloud: Revenues grew 82% YoY to $24.8 billion, contributing 20.7% of Alphabet's total consolidated revenue. Growth was led by GCP, AI solutions, and AI infrastructure, with the first recognition of TPU system sales to external customers in Q2. Google Cloud operating income more than tripled YoY to $8.8 billion, with operating margin expanding from 20.7% to 35.6% YoY. Cloud backlog grew sequentially by over $50 billion to $514 billion, with over 50% of the backlog expected to be recognized as revenue over the next 24 months. 3. Other Bets: Revenues were $382 million, contributing 0.3% of total consolidated revenue. Operating loss was $1.8 billion, driven by continued investment in scaling Waymo and other early-stage bets. 4. Alphabet Level Corporate Activities: Operating loss was $5.8 billion, driven by shared AI R&D expenses.

Risks & headwinds

- Persistent supply constraints for AI compute capacity: Current industry-wide supply constraints mean demand for AI infrastructure and models continues to outpace available internal capacity, requiring higher-cost short-term third-party capacity that creates near-term margin pressure. - Intense competition in the generative AI model market: The frontier AI space is highly dynamic and competitive, with competitors releasing models at a rapid cadence, creating pressure to maintain pace on model capability and release speed. - Aggressive investment in long-term AI development: Large sustained increases in capital expenditure and R&D spending create near-term pressure on free cash flow and profitability, with returns dependent on successful commercial adoption of new AI products and models. - Legal and regulatory uncertainty: The company noted charges for certain legal matters in G&A expenses, with underlying ongoing legal and regulatory risks for AI and core business operations that could impact future results.

Analyst Q&A

• Q: How has Alphabet's view of the generative AI ROIC opportunity changed over the past year, and how is CapEx budgeting changing amid ongoing supply constraints? /

  A: Sundar stated that generative AI is still in very early innings for both consumer and enterprise adoption, with the overall opportunity larger than expected a year ago, so management has grown more bullish on long-term returns with strong current demand. Anat added that the company continues to invest aggressively as long as attractive returns are available, taking a multi-year view of capacity needs. Demand has consistently outpaced capacity additions even after three years of expansion, so the full-stack approach enables efficiency gains to meet unmet demand. (321 words)

• Q: How confident is Alphabet that Gemini can remain at the technology frontier, and what is the plan to improve coding capabilities? /

  A: Sundar acknowledged that the frontier is highly dynamic, and the company is focused on improving in areas like agentic coding, with steady progress already delivered. The recently launched Gemini 3.6 Flash improved DeepSuite benchmark scores by over 10 points while being more token efficient, and it is already being tested with enterprise coding customers. Management is very confident in Gemini 4, the next large frontier base model currently in pre-training, which will allow Alphabet to compete at the highest frontier level when launched. (318 words)

• Q: How does Alphabet balance internal and external demand for TPUs, and what is the margin profile for TPU sales? /

  A: Sundar explained that the first priority for TPU allocation is supporting frontier AI model development, followed by serving core consumer products (Search, YouTube) and enterprise cloud solutions. External TPU sales (placed in customer data centers) help meet unmet demand while freeing internal capacity for core priorities. Anat noted that TPU system sales are included in the $514 billion cloud backlog, with the vast majority of revenue recognized in 2027. She added that TPU design gives Alphabet incremental margins benefits, though third-party bridging capacity will create modest near-term cloud margin pressure. (302 words)

• Q: What strategic moats does Alphabet have amid intensifying competition in AI models? /

  A: Sundar emphasized that Alphabet takes a full-stack end-to-end solution approach, rather than competing only on raw model capability. For enterprise use cases, models are just one component of integrated solutions that include data integration, security, governance, orchestration of agentic workflows, and ongoing customer support. This full-stack offering, combined with Alphabet's scale, existing customer relationships, and commitment to data privacy for enterprise clients, creates a durable competitive advantage even if base model capabilities converge across competitors. (248 words)