Citi significantly raised its forecast for Google’s capital expenditures to $111 billion, as strong AI demand may continue to drive up capital spending.

Citi significantly raised its forecast for Google’s capital expenditures to $111 billion, as strong AI demand may continue to drive up capital spending.

```

Citibank stated in its latest report on Tuesday that as demand for generative AI continues to exceed supply and Google's product iteration accelerates, the bank has raised its capital expenditure forecasts for Google in 2026 and beyond.

Citi expects Google’s capital expenditure in 2026 to reach about $111 billion, up from $86 billion in 2025. According to the latest estimates, the compound annual growth rate (CAGR) for capital expenditure from 2024 to 2029 will reach 26%.

Citi believes that this growth is driven by a larger-scale application of Google’s AI products and services in its core search and cloud businesses. Currently, usage of Gemini tokens continues to surge, surpassing one trillion per month, doubling since April this year. The number of Google Cloud (GCP) customers grew 28% quarter-over-quarter, with order backlogs accelerating, and Gemini is also driving more of Google’s core products.

Citi points out that Google’s capital expenditures are driving faster product development cycles, which will bring sustained growth to the company. Although competition in the search sector remains fierce, Citi believes Google has stronger execution and therefore maintains a “Buy” rating with a target price of $280.

On Tuesday, Google’s parent company Alphabet A shares fell nearly 1% intraday, trading at $241.94.

Accelerating Demand for Gemini and AI Products

According to the report, Gemini token usage increased to 9.8 quadrillion transactions per month by June 2025, up from 4.8 quadrillion in April. Demand for Google Cloud is rising in tandem, with new customers up 28% quarter-over-quarter, and contracts worth over $250 million doubling year-over-year. Gemini's monthly active users reached 450 million, and Citi believes this may have surpassed 500 million in September.

Meanwhile, Gemini's average daily requests grew 50% quarter-over-quarter in Q2, and new features like AI-Overviews and AI-Mode were rapidly deployed. Citi points out that about 60% of generative AI startups have chosen Google Cloud, and nine of the top ten AI labs use Google Cloud.

Moreover, this momentum continued in Q3. Previous media reports stated that Google Cloud has signed a six-year AI infrastructure contract with Meta, totaling $10 billion. Google Cloud's order backlog grew 17% quarter-over-quarter in Q2 to $15.8 billion; by comparison, Amazon AWS increased by $6 billion in the same period.

The report notes that Gemini is gradually being embedded into Google’s core products. For example, in the Chrome browser, AI mode is directly integrated into the address bar; YouTube has launched several new AI features, including fast integration with Veo 3, voice-to-song, and Ask Studio; Google is also testing a Windows version of the Gemini app, allowing users to use Gemini outside of Chrome and Search. Citi believes that Gemini's features are expanding quickly and usage continues to climb.

Infrastructure Expansion

Google recently launched the Model Content Protocol (MCP) server for data sharing, supporting agent-to-agent communication. The company also released an open-source Agent Development Kit (ADK) to help developers orchestrate among multiple model agents and introduced the Agent Platform Protocol (AP2), supporting autonomous agent payments for over 60 merchants and financial institutions. Citi notes that Google is building a broader infrastructure layer for the smart agent and model ecosystem.

Recently, Google announced an investment of over $50 billion in AI infrastructure over the next few years. This includes $25 billion in the tri-state area of Pennsylvania, New Jersey, and Maryland; $9 billion over two years in Virginia to build new data centers; $9 billion in Oklahoma over two years for infrastructure and skills training; $7 billion in Iowa to expand cloud and AI facilities; and $3 billion in Pennsylvania for hydropower facility upgrades.

Google has also announced global projects such as those in the UK. Citi believes this shows Google is ramping up its investments as OpenAI, Meta, and Amazon significantly expand their data center capacity.

Capital Expenditure May Remain High for the Long Term

Citi anticipates that Google’s capital expenditure in 2026 will reach $111 billion, up 29% year-over-year and $25 billion higher than the previous year. The compound annual growth rate from 2024 to 2029 is projected at 26%. About two-thirds will go toward servers, and one-third toward data centers and network equipment.

The report believes that capital expenditure has already produced significant results: Google Cloud revenue grew 32% year-over-year, an acceleration from 28% in Q1. At the same time, AI efficiency is improving. The energy consumption and carbon footprint of Gemini text prompts dropped by 33-fold and 44-fold year-over-year, respectively, and the median water consumption per request is about 0.26 milliliters. Citi emphasizes that demand still outpaces supply, and Google is seizing the opportunity to invest.

Due to the expectation that AI demand will continue to outpace capacity, Citi has raised its capital expenditure and depreciation & amortization (D&A) forecasts for 2026 and beyond. 2026 capital expenditures are raised 12% to $111 billion, while 2027 is raised 15% to $131 billion. Higher capital expenditures and D&A have led Citi to lower its 2026 GAAP EPS forecast by about 2.5% to $10.56, and for 2027 by about 3% to $11.90.

Citi maintains Google’s target price at $280, based on a 2026 GAAP EPS of $10.56 and a 26.5x P/E ratio (previously 26x).

The report notes that the slightly higher valuation multiple reflects the growth potential of AI demand, but is partly offset by higher capital expenditures and depreciation. Citi believes Google, with 15 products with more than 500 million monthly active users and seven with over 2 billion, will see its “product halo” drive continued growth in search traffic.

At the same time, Citi is impressed by Gemini 2.5 Pro and the expanded AI Mode. However, the report also cautions that increasing competition in search and GenAI products, as well as regulatory risks, remain uncertainties that cannot be ignored.

Risk Warning and DisclaimerThe market has risks, and investment needs caution. This article does not constitute personal investment advice and does not take into account the unique investment objectives, financial situations, or needs of individual users. Users should consider whether any opinions, views, or conclusions in this article fit their specific circumstances. Investments made accordingly are at your own risk. ```