Google to Invest $180 Billion in New Data Centers: Is AI’s Hunger Boundless?

Alphabet, the parent company of Google, has outlined a financial strategy for 2026 that leaves no room for interpretation: the company will double down on the generative AI bet.

During the fourth-quarter earnings presentation, the Mountain View giant announced a capital expenditure (Capex) plan expected to be between $175 billion and $185 billion for the year ahead.

It is a figure that nearly doubles last year’s investment, signaling how the hunger for compute power has reached unprecedented levels.

Where do Google’s billions go? Silicon and cement

The massive liquidity Google intends to mobilize has a precise destination: the construction of new data centers and the purchase of the hardware needed to fill them.

Anat Ashkenazi, Alphabet’s chief financial officer, offered a detailed breakdown of how these funds will be allocated. About 60% of the budget, a sum fluctuating between $105 and $111 billion, will be allocated to depreciating assets, namely servers.

This category includes the massive purchase of Nvidia GPUs and the in-house production of Tensor Processing Units (TPUs), including the new generation of accelerator chips “Ironwood”.

The remaining 40% of the budget, amounting to about $70-74 billion, will instead be absorbed by the physical construction of facilities and the network infrastructure.

It’s not just about supporting internal products like Search or YouTube, but ensuring the infrastructural stability needed by Google Cloud’s partners, including names as prominent as Apple, OpenAI and Anthropic. The compute infrastructure investment will be split evenly between Google’s internal workloads and the Cloud platform offered to external customers.

The risks of frenetic growth

Asked by analysts about the biggest near-term concerns, CEO Sundar Pichai highlighted that managing scalability is the real unknown.

The difficulty lies not in finding the funds, but in coordinating the expansion of compute capacity with real-world constraints tied to electricity availability, land acquisition, and supply-chain bottlenecks.

Meeting the demand for AI services requires a planning that goes well beyond code, touching critical physical infrastructure often slow to develop.

Despite the challenges, the strategy seems to be paying off. Alphabet has surpassed for the first time the $400 billion of annual revenue, with a fourth-quarter profit of over $34 billion.

Much of this success is driven by the Cloud division, which saw revenues climb 47% year over year, reaching $17.66 billion in the quarter, fueled by strong demand for AI products for businesses.

Impact on advertising and users

Artificial intelligence is not confined to research laboratories, but is permeating Google’s traditional business model. Philipp Schindler, chief business officer, explained how the Gemini models are improving ad relevance, enabling monetization of long and complex searches that were difficult to interpret in the past.

This greater understanding of user intent contributed to advertising revenues growing by more than 13%.

On the consumer side, the bet on the virtual assistant is paying off. Gemini has surpassed 750 million monthly active users, a jump of 100 million from the previous quarter, driven by the launch of Gemini 3.

Although Google still has to close the gap with ChatGPT, which is moving toward 810 million users, it has distanced Meta AI and is seeing daily AI queries double in AI mode.

The upcoming integration of “agentic” features, capable of completing purchases and concrete actions on behalf of the user, could mark the next step in consolidating this growth.