Second-quarter growth was driven collectively by cloud, search advertising, and AI products, but to secure computing power and model capabilities, the company is trading higher capital expenditures and short-term cash flow pressure for future market share in AI infrastructure and enterprise services.
Following the U.S. market close on July 22nd, Alphabet's latest earnings report revealed total Q2 revenue of $119.8 billion, a 24% year-over-year increase. Operating profit reached $40.8 billion, up 30%. Additionally, driven by a surge in equity gains from its investment portfolio, including stakes like Anthropic and SpaceX, the company's net income saw an astonishing fourfold increase.
Behind the strong revenue figures, market and investor focus is locked on two core issues: how much actual business growth AI is generating for Alphabet, and how much money this seemingly bottomless "AI arms race" will ultimately consume.
Alphabet's Q2 capital expenditures reached $44.9 billion, with the vast majority directed towards AI technology infrastructure. The company also raised its full-year 2026 capital expenditure guidance from the previous range of $180-190 billion to $195-205 billion. As a result, the company's free cash flow turned negative at -$5.9 billion for the quarter, whereas in the same period last year and for many years prior, Alphabet had been known for its robust free cash flow.
Following the earnings release, CEO Sundar Pichai, Chief Business Officer Philipp Schindler, and CFO Anat Ashkenazi participated in the Q2 earnings call. Pichai stated on the call: "Our AI investments are redefining what's possible across our entire business."
In response to a question from a Morgan Stanley analyst regarding the return on invested capital for generative AI, Pichai offered his core assessment:
"I think we are at a very, very early stage, this is a structural shift across many domains... Just like the cloud itself, a small percentage of enterprise workloads have moved to the cloud. Now think about what percentage of workloads are truly AI-native, AI-enabled—it feels very, very early again. Looking at the past year, if anything, we've become more optimistic about the opportunity ahead."
Cloud Revenue Surges 82%, $514 Billion Backlog Key to Future Realization
Google Cloud was the strongest growth engine this quarter.
The report shows Google Cloud Q2 revenue was $24.8 billion, an 82% year-over-year increase. Operating profit was $8.8 billion, more than doubling, with the operating margin rising from 20.7% in the same period last year to 35.6%.
Pichai stated that cloud growth was "driven by strong demand for AI infrastructure and AI solutions." He noted that Google Cloud's growth is not from a single product or a few customers, but from demand for "full-stack" AI services covering chips, models, data, security, and agent platforms.
Of greater market interest, Google Cloud's contract backlog increased by over $50 billion sequentially to $514 billion. CFO Ashkenazi stated that the vast majority of this backlog consists of standard GCP contracts from a broad customer base, and the company expects to recognize "slightly more than 50%" of this revenue over the next 24 months.
Pichai said enterprise customers are expanding their usage beyond existing commitments, exceeding original contract commitments by more than 50%. He stated: "We are winning new customers and deepening relationships with existing ones."
Regarding enterprise AI products, Alphabet said nearly 90% of Fortune 100 companies are now using Gemini Enterprise. Over the past 12 months, nearly 500 cloud customers each processed over 1 trillion tokens, and over 2,000 enterprise customers each consumed over 100 billion tokens.
However, a massive backlog does not mean immediate revenue conversion. Ashkenazi specifically noted that the company began recognizing revenue this quarter from delivering TPU systems to customer data centers for the first time, but will only recognize a "relatively small portion" of the revenue from existing TPU system sales agreements this year, with the "vast majority" of related revenue to be realized in 2027.
Capital Expenditures Raised Again, Short-Term Cash Flow and Margins Under Pressure
Of the $44.9 billion in Q2 capital expenditures, approximately 60% was for servers and about 40% for data centers and networking equipment. The company stated the vast majority of this investment serves the expansion of AI infrastructure.
Ashkenazi stated the company continues to face compute supply constraints: "We remain in a supply-constrained environment." She noted that although Alphabet has significantly increased compute capacity over the past three years, "demand still exceeds the pace of our investment growth."
Against this backdrop, Alphabet decided to further raise its capital expenditure budget. Ashkenazi said the upward revision to the full-year capex guidance is "primarily due to accelerating delivery of compute capacity to meet growing demand."
She also warned that AI infrastructure investment will continue to pressure the income statement and cash flow: "We expect free cash flow to continue to be pressured by technology infrastructure investments."
The report shows that in Q2, operating cash flow was $39.1 billion, but due to $44.9 billion in capital expenditures, free cash flow was negative $5.9 billion. Free cash flow over the trailing 12 months remained positive at $53.3 billion. As of quarter-end, the company held cash and marketable securities of $242.5 billion, including $87.1 billion in marketable equity securities. Long-term debt stood at $98.2 billion.
Addressing investor concerns about return on investment, Pichai stated the company will continue to use return metrics as a core constraint. "As long as we see attractive return on investment opportunities, we will continue to invest."
He also said that when allocating compute resources, the top priority is to ensure the company "can remain competitive in frontier AGI development." On that basis, compute will be prioritized for search, YouTube, Gemini, and cloud enterprise products.
Facing intense spending and short-term cash flow pressure, Pichai conveyed a firm signal to the market:
"When you have a multi-year opportunity like this, as we bring more capacity online, it drives a high return on investment over the life of the deal."
To Secure Customers, Q3 to Use Third-Party Compute as a "Bridge"
Alphabet acknowledged that internal compute build-out speed cannot fully meet customer demand in the near term.
Ashkenazi stated the company will expand the use of third-party compute in the third quarter as a "bridge strategy" until internal infrastructure is ready. This will help onboard more large cloud customers and multi-year contracts in the short term but will also lead to higher costs.
She said: "This strategy allows us to continue expanding our customer base and capturing larger overall value." However, due to the higher cost of third-party capacity, this will "put modest pressure on margins in the near term."
Pichai further explained that for some very large cloud customers, the company is willing to accept higher service costs for a few months in exchange for the returns from a multi-year contract. "In a multi-year opportunity, as more capacity comes online, both the margin and the return will be very attractive."
Beyond pressure from third-party compute, Alphabet also expects that increasing depreciation, data center energy costs, and hiring investments in AI and cloud talent will continue to elevate expense levels. The company also mentioned that the integration of the Wiz business will bring some short-term pressure to cloud business margins in 2026.
Search Revenue Grows 17%, AI Search Begins Creating New Commercialization Avenues
Amid concerns that AI could disrupt the traditional search business, Alphabet's core advertising business maintained growth.
The report shows that in Q2, Google Search and other advertising revenue was $63.3 billion, up 17%. YouTube advertising revenue was $11.1 billion, up 13%. Google Services total revenue was $94.5 billion, up 15%. Retail and finance were the primary drivers of search advertising growth.
Pichai stated that AI features are driving users to ask more and more complex search queries. Since its global expansion last October, AI Mode has surpassed 1 billion monthly active users. The Gemini App reached 950 million monthly active users, with daily active users tripling year-over-year.
He said AI Overviews and AI Mode are forming an "integrated search experience" and are driving incremental growth in search query volume. "Like AI Overviews, AI Mode is driving incremental growth in overall search query volume."
Chief Business Officer Philipp Schindler said Gemini has been deeply integrated into the advertising infrastructure, enabling the company to understand longer, more complex user search intents and match more relevant ads.
He stated: "Gemini has significantly enhanced our ability to understand user needs and match the right ads." Taking shopping ads as an example, after improving ad relevance with AI, the performance of highly relevant ad impressions improved by 20%.
Alphabet is also testing new ad formats within AI search, including text ad links generated from conversational content, direct offers within travel planning, and clearly labeled sponsored links inserted into answer lists.
As of Q2, 500,000 advertisers had adopted the AI Max tool. Alphabet said advertisers using AI-powered ad tools like AI Max or P Max average 15% higher conversions or conversion value at similar return on ad spend.
Gemini 4 Launches "Most Ambitious" Training, Model Release Cadence to Accelerate
As external scrutiny persists on whether Google can maintain frontier competitiveness in large models, Pichai responded that the company is accelerating model iteration and dedicating more compute to Gemini 4 training.
Alphabet this week released models including Gemini 3.6 Flash, Gemini 3.5 Flashlight, and Gemini 3.5 Flash Cyber. Pichai emphasized that the Flash series balances performance, cost, reliability, and response speed, making it the most in-demand "workhorse model" currently.
He stated: "We have kicked off the most ambitious pre-training to date for Gemini 4." Regarding market interest in coding and agentic coding capabilities, Pichai acknowledged Google still has room for improvement but said the team is iterating rapidly.
"We are very committed and very confident in staying at the frontier." Pichai stated that Gemini 4 will be a larger-scale foundational model, and Google wants it to compete at the frontier when released.
He also said the future model release cadence will accelerate significantly:
"As we build Gemini 4, you'll see us continue to accelerate, and releasing models at a near-monthly cadence will be part of our roadmap."
As of Q2, over 9 million developers use its models monthly via Alphabet's APIs and developer products. Model API throughput reached approximately 22 billion tokens per minute, up from 16 billion tokens last quarter.
Comments