←News detail

TMTPost · AGIUpdated Oct 30 14:18

Gemini 4 Carbon's Top Ranking Fails to Mask Disappointing Earnings: The Bill for Google's AI Gamble

AI summary

Alphabet's third-quarter revenue rose 25% year-over-year to $127.58 billion, while net profit fell 25% year-over-year to $26.17 billion, free cash flow was negative more than $9 billion, and the stock fell nearly 10% in after-hours trading. Capital expenditures of $51.4 billion were mostly directed toward AI infrastructure, and CFO Anat Ashkenazi maintained the full-year 2026 capital expenditure range of $195 billion to $205 billion unchanged. Hours before the earnings release, Google launched Gemini 4 Carbon, which topped the Artificial Analysis Intelligence Index with a score of 59, but it has not yet opened a public API and is initially available only to some cybersecurity professionals.

Why read this

AI score 60

Examining the cost of Google's AI investment through its earnings cash flow, capital expenditures, and cloud business structure, and comparing it with the commercial posture of Gemini 4 Carbon released the same day.

Topics

Tags

Body · AI translation

Google Cloud's impressive growth in the third quarter was overshadowed by other financial results that disappointed Wall Street, especially the sluggish growth in Google search advertising revenue. Google's stock fell nearly 10% in after-hours trading.

Negative $9 billion in free cash flow

The earnings report shows that Alphabet's third-quarter revenue grew 25% year-over-year to $127.58 billion, slightly below Wall Street expectations; net profit fell 25% year-over-year to $26.17 billion. Due to heavy investment in artificial intelligence, third-quarter free cash flow was more than negative $9 billion, compared with positive more than $24 billion in the same period last year, marking the second consecutive quarter of negative free cash flow.

So where did the money go? Capital expenditures this quarter were $51.4 billion, most of it for AI technology infrastructure, of which 65% was servers and networking equipment and 35% was for data centers. It is worth noting that the inventory buildup from sales of Google's TPU systems and the impact of the $5.2 billion cash payment for the 2018 EU fine are also affecting the cash flow situation; third-quarter operating cash flow was $42.3 billion.

Alphabet CFO Anat Ashkenazi noted at the earnings call that the company's full-year 2026 capital expenditure guidance range of $195 billion to $205 billion remains unchanged, and capital expenditures are expected to increase significantly in 2027. This has made the outside world pay even more attention to whether Google's spending now and in the future is worthwhile, and the only bright spot in this quarter's earnings report was Google Cloud, not Google Search.

On the one hand, Alphabet's core search business grew only 12% to $63.35 billion, the lowest growth rate for that business since the second quarter of 2025. This result contradicts its claim that "people are using Google Search more than ever before." Alphabet CEO Sundar Pichai said Google's search query volume year-over-year growth rate was the highest in five years, while Alphabet Chief Business Officer Philipp Schindler explained that search volume in AI Mode was affected, causing AI Mode's profitability to decline at this stage. This statement means that AI is indeed changing users' search behavior, but how to commercialize it is still unclear.

On the other hand, Google Cloud's standout performance mainly came from its booming TPU business. Google Cloud revenue more than doubled this quarter to more than $30 billion, while operating profit tripled to $11.12 billion. This revenue figure is only slightly below Microsoft Azure's $33 billion in revenue for the same period.

However, Google did not disclose specific TPU sales for the quarter, but said TPUs were "a significant contributor to revenue growth." Alphabet management told analysts they expect TPU sales to have a greater impact on Google Cloud next year. Meanwhile, earlier this month, a TD Cowen report estimated that TPUs would generate $9.4 billion in sales in the second half of this year.

Externally, demand for TPUs from a large number of customers is surging, including OpenAI and Anthropic; the former has long been a major customer of Nvidia AI chips, while the latter has already announced plans to invest $114.5 billion in Google Cloud.

Strongest model tops the ranking on the same day

Just hours before the earnings release, Google launched Gemini 4 Carbon, which topped the Artificial Analysis Intelligence Index with 59 points, beating Claude Opus 5.5 and GPT-6 Astra. This is the second major model in a month—Google, which had not released a flagship model for nearly a year, suddenly released two in a row, seemingly eager to win back market share from OpenAI and Anthropic.

Google positions Carbon as "the strongest model for cybersecurity and long-horizon tasks." Google said Carbon has achieved a "major leap" in cybersecurity capabilities, and its ability to generate secure code is nearly twice that of Argon. Carbon's release comes as a cybersecurity crisis erupts, after some malicious AI programs exploited vulnerabilities in enterprises and governments.

Like Argon, Google said it will first release Carbon to select cybersecurity professionals, and then to partners in its Fairwind program. Google also said it is actively participating in the U.S. government's pre-release model testing program.

But for now, Carbon's commercial posture is awkward: no public API has been opened, it is available only to a small number of cybersecurity defenders, and the opening time for ordinary developers has not been set. At the same time, on pricing, it continues to follow the "low-price entry, double upon expiry" approach ($2 input/$10 output, doubling after the promotional period).

Tech giants' financing shift

As part of the "at-the-market offering program" announced earlier this year, Alphabet sold $12.4 billion worth of stock in the third quarter: 12 million Class A shares raised $4.3 billion, and 24 million Class C shares raised $8.1 billion. This is part of the $90 billion equity financing plan announced in June, with $27.6 billion still available as of September 30.

Over the past year, Alphabet has cumulatively issued more than $85 billion in bonds through six major currency markets, with total debt exceeding $100 billion. After entering 2026, the pace of bond issuance has clearly accelerated: in February it issued $20 billion in multi-tranche dollar bonds, in May it successively issued 9 billion euros and 8.5 billion Canadian dollars in bonds, and then sold about $3.6 billion in yen samurai bonds. This "equity and debt together" financing strategy forms a sharp contrast with its image over many years of preferring share buybacks and maintaining low debt.

In fact, it is not just Alphabet. As investment in AI infrastructure continues to increase, major tech companies have been racing to issue debt and equity to raise funds. Meta said on Tuesday that it plans to raise $97.5 billion through a stock offering for AI investment, higher than the previously announced $95 billion target. At the same time, Meta's third-quarter free cash flow also turned negative, at negative $2.12 billion.

Currently, several trillion-dollar tech giants are still continuously adding to their funding bets, and this still-unclear AI race continues. (This article was first published on TMTPost APP, written by TechHorizon)

Source

TMTPost · AGI