$400 billion in revenue as a foundation, Google's capital expenditure doubles to aggressively pursue AI

21st Century Business Herald Reporter Dong Jingyi

2025 will be a landmark year for Google.

After the market close on February 4th, Eastern Time, Google’s parent company Alphabet released its 2025 financial report, with annual revenue reaching $402.84 billion, surpassing the $400 billion mark for the first time. In the fierce AI competition, this Silicon Valley giant still maintains the initiative.

However, alongside impressive performance, there are huge investments. During the earnings call, Google also announced that capital expenditures are expected to reach $175 billion to $185 billion in 2026, nearly doubling from $91 billion in 2025. After the earnings release, Google’s stock price briefly plummeted in after-hours trading, with a maximum drop of over 7%.

Artificial intelligence is no longer just a future vision but a powerful engine driving current growth. Yet, maintaining this momentum comes at an increasingly high cost.

Two years ago, Wall Street was worried that the emergence of ChatGPT would erode Google’s search moat, with some predicting that this internet tech giant would face a “Kodak moment.” However, with the release of its Q4 and full-year financial reports for 2025, solid performance data have already responded to market doubts.

The annual report shows the company’s full-year revenue reached $402.84 billion, a 15% increase year-over-year. Net profit was $132.17 billion, up 32%, with profit margins rising rather than falling, which is particularly notable during a period of large-scale AI investment by tech giants.

Q4 revenue was $113.83 billion, up 18% year-over-year, with net profit of $34.46 billion, up 29.8%, exceeding market expectations. CEO Sundar Pichai called it “an extremely outstanding quarter” during the earnings call.

The core growth driver is Google Cloud. In Q4, Google Cloud revenue reached $17.664 billion, up 48% year-over-year, surpassing analyst expectations by over 9%. This growth far exceeds the industry average for cloud computing and is significantly faster than Microsoft’s Azure performance during the same period.

The growth is mainly driven by increased demand for enterprise AI infrastructure, enterprise AI solutions, and core Google Cloud platform products. Alphabet disclosed that by the end of 2025, Google Cloud’s annualized revenue exceeded $70 billion, with unfulfilled orders reaching $24 billion, more than doubling year-over-year.

Beyond cloud, search business continued to accelerate, with revenue up 17% year-over-year to $63.1 billion, surpassing the analyst forecast of $61.4 billion and remaining the main engine of total revenue.

AI continues to empower this growth. Pichai stated during the earnings call that, driven by AI-powered business expansion, search usage hit a record high in Q4. Bank of America Securities believes that the widespread adoption of AI-native search formats has increased user engagement and conversion rates, further unlocking incremental growth in search.

Regarding the commercialization of Google’s AI products, which investors are particularly focused on, Pichai provided several key indicators: Gemini Enterprise service sold over 8 million paid seats within just four months of launch; the number of tokens processed per minute via customer API calls has exceeded 10 billion; Gemini 3 became Google’s fastest-adopted model in history, with monthly active users surpassing 750 million; and in 2025, Gemini’s service unit costs were reduced by 78%.

Notably, progress in Google’s collaboration with Apple was confirmed. Pichai stated that both companies are working together to develop foundational models, and Google has become Apple’s preferred cloud service provider. This partnership could bring Google billions of dollars in additional cloud revenue and ensures that Google’s AI technology reaches global users through Apple devices.

Capital Expenditures Double

Beyond impressive results, massive investments are beginning to impact Google’s financial performance and strategic decisions.

In Q4 2025, Google’s capital expenditures accelerated to $27.9 billion, nearly doubling from $14.3 billion a year earlier, accounting for nearly one-third of the total $92 billion annual capital expenditure. This acceleration is expected to continue into 2026.

To meet the surging demand for AI computing power, Alphabet expects capital expenditures in 2026 to further expand to $175 billion–$185 billion. Funds will mainly be used to expand data center capacity, purchase NVIDIA chips, and develop its own TPU chips.

Investors have expressed concerns about such large-scale spending. During the Q&A session of the earnings call, several analysts asked about the return timelines and expected yields of these investments.

Pichai responded that “our capital expenditures this year are focused on the future, and supply chain cycles are also lengthening.” He noted that strong demand for services, Google DeepMind’s future development, and cloud business will continue, so the entire 2026 year may still face supply constraints.

To balance investment and profitability, Google is implementing efficiency improvement plans across the company. Alphabet CFO Anat Ashkenazi said nearly 50% of Google’s code is now generated by AI, significantly improving engineers’ productivity. Additionally, the company is deploying AI agents to optimize internal processes such as finance and operations to control costs.

Nevertheless, the market remains concerned that such scale of spending could lead to margin declines in the coming quarters. After the earnings release, Alphabet’s stock initially dropped 7.5% in after-hours trading, then briefly rose over 4%, before turning downward again, reflecting investor ambivalence toward the company’s AI strategy.

Bank of America analysts pointed out: “Google faces fierce competition from Microsoft Azure, Amazon AWS, and others. High capital expenditures are necessary but may compress short-term profitability.”

In the long run, this race for capital investment could reshape the entire tech industry. Google, Microsoft, and Amazon are all significantly increasing AI-related investments.

For Google, the risk is that if AI demand growth falls short of expectations, it could face overcapacity and asset impairments; but if the judgment is correct, these investments could help Google dominate the next wave of technological revolution. Pichai expressed confidence: “AI investments and infrastructure are driving revenue and growth across all business lines. We will continue investing to meet customer needs and seize future growth opportunities.”

Looking back at major moves, it’s clear how Google has been steadily paving the way for this “big gamble.” Starting in 2024, Google sharply reduced investments in some non-core or uncertain prospects, such as cutting AR hardware teams and optimizing personnel in its advertising and other business units, reallocating resources to key AI growth areas.

Meanwhile, Google accelerated TPU iterations in 2025. Compared to relying entirely on external suppliers like NVIDIA, developing its own chips has strengthened supply chain resilience amid computing shortages and significantly lowered unit costs. Analysts note that the “high investment, high profit” shown in the financials largely stems from its long-standing self-developed chip strategy, which bore fruit in 2025.

Another key move is Google’s “model-as-platform” ecosystem build. Pichai stated during the earnings call that self-developed models like Gemini are now processing over 10 billion tokens per minute via customer API calls. After establishing the most powerful computing infrastructure through massive capital expenditure, Google is offering its capabilities via Google Cloud to developers, sharing depreciation costs and locking the global AI innovation ecosystem into Google’s tech stack.

These advantages are clear, but this AI gamble is unprecedented in scale, with fierce competitors in the race. Investors are already wary of the huge expenditures by tech giants, and markets are relearning how to price them. For Google, the next few years will remain a test.