Alphabet (GOOGL) shares fell more than 10% after the company delivered third-quarter revenue of $127.58 billion that fell short of Wall Street's expectations, even as sales grew 25% year over year. It was Alphabet's first simultaneous miss on revenue and earnings in almost four years, and heavy spending on artificial intelligence left investors worried about cash. The tension between booming AI demand and the rising cost of supplying it frames both the quarter's results and the company's plans.

Alphabet's AI bill keeps getting bigger

Analysts surveyed by LSEG had expected revenue of $128.22 billion and adjusted earnings of $3.02 per share. Alphabet instead reported adjusted earnings of $2.67 per share, according to its earnings release.

Net income fell to $26.17 billion, or $2.12 per share, from $34.98 billion, or $2.87 per share, a year earlier. Operating income reached $40.58 billion, below the $43.19 billion forecast compiled by FactSet.

The number that drew the most attention from investors was free cash flow. Alphabet recorded negative free cash flow of $9.07 billion, compared with the $2.70 billion deficit analysts had penciled in — a gap that raises questions about the payoff from the company's AI investments. Negative free cash flow means a company spends more cash than it generates in the period, a yardstick of how heavily the AI buildout is drawing on Alphabet's finances.

Quarterly capital expenditures were estimated at $51.36 billion, below the $55.09 billion forecast. The annual budget nonetheless remains very large: Alphabet maintained its full-year capital expenditure estimate at $195 billion to $205 billion, revised up from its previous $190 billion target in the second quarter. Spending is projected to exceed $200 billion for this year and to top $300 billion by 2027. Capex funds the data centers and custom chips behind Gemini and Google Cloud, making it a key gauge of the AI buildout's scale.

"We continue to expect our capex to increase significantly in 2027, and will provide further details on our Q4 call," finance chief Anat Ashkenazi told analysts.

There is also a growing debt tab. Alphabet's long-term borrowing jumped from $21.6 billion a year ago to $125.1 billion, an increase of roughly 479%.

And the investment portfolio was not much help. Equity holdings tied to companies including SpaceX and Anthropic helped produce a $99.03 billion gain in the second quarter; this time, that category registered an $8.66 billion loss. The swing shows how Alphabet's results now move with the value of its private-market stakes as well as with its operating businesses.

Google Cloud delivers, but Search misses its target

Alphabet's AI business is not struggling to find customers. Google Cloud revenue more than doubled to $30.7 billion, comfortably beating the $28.7 billion analysts had expected. Growth reached 103% as businesses rushed to buy computing power and AI tools, and Google is gaining ground against Amazon Web Services and Microsoft Azure. At roughly a quarter of Alphabet's quarterly revenue, Cloud is growing about four times faster than the company overall.

"We are seeing extraordinary demand and strong growth from both AI labs, capital markets, financial service institutions, and enterprises," Google Cloud CEO Thomas Kurian told investors. It was his first appearance on an earnings call.

Google Search, the company's long-standing core business, had a less convincing quarter. Search revenue rose 12% to $63.35 billion but missed the $64.96 billion estimate. Search still accounts for roughly half of Alphabet's quarterly revenue, keeping its growth central to the company's overall pace. CEO Sundar Pichai said AI features are bringing in more searches, including commercial queries, and that annual query growth is now its strongest in five years.

"We had a strong quarter and are now on a half-trillion dollar revenue run rate. Our deep full stack approach is driving momentum across the business, with Alphabet revenues growing 25% year-over-year," Pichai told analysts.

Usage milestones and an accelerating AI roadmap

There is plenty of usage data behind that push. The Gemini app now has one billion users, which Alphabet says makes it Google's fastest-growing product ever. YouTube Music and Premium have passed 150 million subscribers worldwide, including trials. Google's own AI model interfaces are handling nearly 32 billion tokens per minute, more than four times last year's volume, while demand for its custom TPU chips is rising.

Pichai also provided details on upcoming AI releases. The new Gemini 4 Argon has attracted many users, he said, but the even more advanced Carbon model outperforms its predecessor in nearly all tested categories, including programming, multimodal capabilities, scientific reasoning, and deep knowledge. Carbon scored 59 on the upgraded Artificial Analysis Intelligence Index, placing it in first position. Google is providing the first preview of the model to trusted security experts and the U.S. government. How widely Carbon becomes available beyond that first preview is one thing to watch, along with the capex details Ashkenazi promised for the fourth-quarter call.

Then there is Genie 4, Google's latest world model, which creates interactive, realistic digital environments with live audio. Google plans to use the technology to train robots and AI agents.

Pichai expects model development to speed up over the next year as AI helps improve AI research itself, a process he referred to as RSI. Smaller, cheaper, and faster models are also on the way for developers and Google's billions of users.

Source: Cryptopolitan