Alphabet Posts First Negative Free Cash Flow on Record $44.9 Billion Quarterly Capex, Lifts Full-Year Outlook to $205 Billion

Google parent Alphabet's AI arms race is entering a deep "cash-burning" phase. After the US market close on July 22, Alphabet (GOOGL) reported its second-quarter 2026 earnings. While overall revenue and cloud business growth were strong, the company sharply raised its full-year capital expenditure guidance to as much as $205 billion and posted a rare negative quarterly free cash flow, sending shares tumbling more than 4% in after-hours trading.

Alphabet reported second-quarter revenue of $119.8 billion, up 24% year-over-year and exceeding the consensus estimate of $116.9 billion. However, behind the impressive top-line numbers, the market fixed its gaze squarely on soaring AI infrastructure spending. Capital expenditures for the quarter reached $44.9 billion, well above the expected $44.15 billion. Chief Financial Officer Anat Ashkenazi announced that the full-year 2026 capex outlook has been raised sharply from the previous $180 billion–$190 billion range to $195 billion–$205 billion, and warned that 2027 spending will increase "significantly" further.

Weighed down by the massive investments, Google posted negative free cash flow for the first time in its history, recording -$5.9 billion. This signal, combined with the delayed release of its flagship Gemini 3.5 Pro model, has stoked market concerns about the return cycle on AI investments.

Cloud Business Becomes Growth Engine, Backlog Tops $500 Billion

Among its various business segments, Google Cloud was undoubtedly the quarter's brightest spot. Fueled by surging demand for enterprise AI solutions and infrastructure, Google Cloud revenue reached $24.8 billion, soaring 82% year-over-year. This not only far exceeded the analyst estimate of $22.46 billion, but the segment's operating margin also jumped to 35.6% from 20.7% a year earlier.

A more critical indicator of future revenue—the cloud business backlog—grew by more than $50 billion sequentially to reach $514 billion. Ashkenazi stated that the vast majority of this backlog comes from a broad customer base, with slightly more than 50% expected to be recognized as revenue within the next 24 months.

CEO Sundar Pichai noted on the earnings call that enterprise customers are expanding their usage of existing commitments, exceeding original contract levels by more than 50%. Currently, nearly 90% of Fortune 100 companies are using Gemini Enterprise.

Advertising Business Steady, AI Drives Incremental Search Growth

In its core advertising business, Google maintained double-digit growth. Second-quarter Google Search & Other advertising revenue was $63.3 billion, up 17% year-over-year; YouTube advertising revenue reached $11.1 billion, up 13%, partly boosted by a traffic surge from the 2026 FIFA World Cup, where over 1.7 billion unique viewers globally watched related content on YouTube.

Pichai emphasized that AI features are driving incremental growth in search query volume. Since its global expansion last October, AI Mode has surpassed 1 billion monthly active users; the Gemini App has reached 950 million monthly active users, with daily active users tripling year-over-year.

Chief Business Officer Philipp Schindler revealed that Gemini has been deeply integrated into the advertising infrastructure, enabling the company to understand longer and more complex user search intent. Taking shopping ads as an example, after improving ad relevance through AI, the display effectiveness of highly relevant ads improved by 20%.

Insatiable Compute Hunger, Third-Party "Bridge" Strategy, and the Gemini 4 Bet

Despite massive investments, Google remains in a "supply-constrained" environment. Ashkenazi stated bluntly that even though compute capacity has been significantly expanded over the past three years, "demand continues to outpace the rate of investment." To bridge the gap, the company plans to expand its use of third-party compute as a "bridge strategy" in the third quarter, though this will put modest near-term pressure on margins.

Responding to a JPMorgan analyst's question about model competitiveness, Pichai acknowledged that Google still needs to improve in coding and agentic coding domains, but immediately announced that the company has begun "the most ambitious Gemini 4 pre-training to date." He said the team is directing massive compute resources toward this larger-scale foundation model and plans to accelerate the model release cadence to a near-monthly pace.

Just this week, Google released new models including Gemini 3.6 Flash, which improved by more than 10 percentage points over its predecessor on coding benchmarks while consuming fewer tokens. Currently, over 9 million developers use Google models via API each month, with the model API processing approximately 22 billion tokens per minute.

Investment Portfolio Contributes Massive Paper Gains

Notably, Alphabet's net profit for the quarter soared to $112.1 billion, a 298% year-over-year surge, but this was primarily driven by unrealized gains from its investment portfolio. The company's equity stakes in firms such as Anthropic and SpaceX saw significant valuation increases, contributing approximately $99 billion in equity investment gains. Excluding these non-recurring items, earnings per share came in slightly below market expectations.

Investing.com Senior Analyst Thomas Monteiro commented: "After a negative free cash flow quarter, Alphabet raising capex again makes the market uneasy. The most reliable cash generator in the market is now spending more than it earns." However, he also noted that the growth in cloud backlog is a clear signal that investments are translating into rapidly growing revenue, and contracts can go live immediately once capacity is built out.

With Microsoft and Amazon set to report earnings next week, Alphabet—as the first Big Tech company to deliver results—has cast new doubt on the sustainability of the broader AI infrastructure investment boom with its "high spending for high growth but pressured earnings quality" model.