Alphabet plans to invest USD 5.8 billion in Belgium over the next two years for cloud computing and AI infrastructure development.
According to Zhitong Finance APP, Alphabet (GOOGL.US), the parent company of US search engine and artificial intelligence giant Google, announced on Wednesday local time that it will invest approximately 5 billion euros (equivalent to about 5.8 billion US dollars) in cloud computing and artificial intelligence infrastructure in Belgium, a core European market, over the next two years. This move signifies Alphabet’s intensified efforts to build out its global artificial intelligence software and hardware ecosystem. Recent research reports from Wall Street financial giants indicate that Google's capital expenditure is expected to reach 111 billion US dollars by 2026, representing a year-on-year increase of 29%, or 25 billion US dollars more than the previous year. Citi forecasts a compound annual growth rate of approximately 26% between 2024 and 2029.
In a newly published blog post, Alphabet stated that this investment includes expanding its large data center campus in Saint-Ghislain, primarily for AI-related infrastructure development, and will create 300 new full-time jobs. The company also noted that it has reached new power supply agreements with European energy and power system suppliers Eneco, Luminus, and Renner to construct new onshore wind farms and provide clean energy supply systems for data centers.
Alphabet added in the blog: "This is an extraordinary moment for innovation in the European market and its digital and economic growth prospects. Our Google is deepening its roots in Belgium and across Europe, investing in local communities to unlock significant economic growth opportunities for the region and helping ensure the continent maintains leadership in cutting-edge technology and artificial intelligence."
It is reported that as this latest announcement was made, Alphabet is preparing to expand its investment in cloud computing and AI infrastructure in the Indian market to 10 billion US dollars. The company also recently stated that it plans to invest 4 billion US dollars in Arkansas, USA, by 2027 to build a new large-scale data center and intends to invest approximately 5 billion pounds in the UK market.
It is understood that during the Q2 earnings call in July, Alphabet's management raised the company’s full-year capital expenditure forecast by 10 billion US dollars to 85 billion US dollars. This increase will be mainly used for high-performance AI server clusters (based on Google's self-developed TPU and NVIDIA AI GPUs, among other AI accelerators), and supporting infrastructure related to data center networks and electricity. At the time, Alphabet stated that driven by strong demand from major clients for cloud computing services and cloud-based AI inference capabilities, as well as other growth opportunities, 2026 capital expenditures would exceed those of 2025, though specific figures have yet to be disclosed.
Strong AI Computing Power Demand! Wall Street Bets Big on the 'AI Spending War' Led by Tech Giants
In the view of Wall Street giants Citi, Loop Capital, and Wedbush, the wave of global artificial intelligence infrastructure investment centered on AI computing hardware is far from over; it is only just beginning. Driven by an unprecedented 'AI inference computing power demand storm,' this round of AI investment could reach a scale of 2 to 3 trillion US dollars.
Jensen Huang, CEO of NVIDIA, further predicted that AI infrastructure spending will reach 3 to 4 trillion US dollars before 2030, presenting significant long-term growth opportunities for NVIDIA due to the scale and scope of these projects. The demand for AI computing power driven by generative AI applications and AI agents on the inference side is vast, comparable to 'a sea of stars and oceans,' and is expected to propel the AI computing infrastructure market into exponential growth. 'AI inference systems' are also seen by Jensen Huang as NVIDIA’s largest future revenue source.
Recent research reports published by Wall Street financial giants have generally indicated that the wave of AI infrastructure investment and construction led by OpenAI and other AI pioneers, as well as major U.S. technology giants, is undergoing an extremely rapid expansion process. This has propelled the entire AI computing power industrial chain to remain the most globally vibrant industry, which explains why shares of leading companies in the AI computing power supply chain, such as NVIDIA, Taiwan Semiconductor, Broadcom, Micron Technology, and SK Hynix, have repeatedly hit record highs. The latest statistical data shows that global AI startups, including OpenAI, have collectively raised a record-breaking $1.927 trillion, largely indicating that the global-scale AI 'capital expenditure battle' remains in full swing. This also highlights that the bull market trajectory for the AI computing power supply chain is far from over.
As the demand for AI computing power driven by generative artificial intelligence and AI agents continues to outstrip supply, coupled with Google's accelerated iteration of its AI products, Citi has recently revised upward its capital expenditure forecasts for Alphabet in 2026 and beyond. Citi predicts that Alphabet’s capital expenditure in 2026 will reach approximately $111 billion, surpassing the $85 billion projected by Alphabet's management for 2025. According to Citi’s latest estimates, Alphabet’s compound annual growth rate (CAGR) of capital expenditures from 2024 to 2029 will reach 26%.
According to Citi’s research report, the usage volume of AI tokens for Gemini, Google’s flagship AI application software, surged to 980 trillion per month in June 2025, significantly higher than 480 trillion in April. Citi noted that demand for Google Cloud services has also risen sharply in tandem, with new customer growth increasing by 28% quarter-over-quarter, and large contracts exceeding $250 million doubling year-over-year. Citi expects that the number of monthly active users for Gemini could approach 500 million by September, second only to ChatGPT.
