Google’s next big AI infrastructure move is not only about chips, models, or data centers. It is also about electricity. The company is now working with partners on a roughly $20 billion renewable energy push intended to help power several gigawatt-scale data centers.
The deal brings together Google, renewable developer Intersect Power, and investment fund TPG Rise Climate. It shows how quickly AI demand is reshaping energy planning for major technology companies.
Why Google is turning to renewable energy for AI
Tech companies have spent recent months drawing attention to nuclear power as they search for large, reliable electricity sources. Google has been part of that conversation. But the source article makes clear that nuclear power alone does not appear to be enough for the company’s AI ambitions.
Instead, Google is moving ahead with a broader energy strategy built around renewable power, battery storage, and grid upgrades. The goal is to create enough carbon-free power to support several gigawatt-scale data centers.
That matters because AI data centers are unusually power-hungry. As companies such as Google race to expand their AI capabilities, experts cited in the source expect new AI data centers might be underpowered by 2027. That prospect is pushing technology companies to invest directly in new energy supply, rather than simply waiting for existing grids to absorb demand.
In practical terms, this means the AI buildout is becoming an energy buildout. Data center plans now need to account for generation, storage, and connection to the grid. Google’s deal is an example of that shift.
What the $20 billion plan includes
Google announced Tuesday that it signed a deal with Intersect Power and TPG Rise Climate. The arrangement is designed to produce enough carbon-free electricity to drive several gigawatt-scale data centers. Altogether, the renewable power investment will run about $20 billion.
Intersect Power is already financing the first project, according to what the company told TechCrunch. The deal also includes an $800 million equity investment in Intersect Power. TPG is leading that round, with CAI, Google, and Greenbelt Capital Partners participating.
The structure described in the source pairs data centers with new renewable energy resources. For a hypothetical 1 gigawatt-scale data center, the plan would match it with an equivalent amount each of wind, solar, and battery storage. The battery storage would have enough capacity to last two to four hours.
Both the data center and the renewable power park would connect to the same point on the grid. Google said it will pay for any required grid upgrades.
The main components are straightforward:
- Renewable power: wind and solar capacity tied to new data center demand.
- Battery storage: storage designed to help support the renewable power park.
- Grid upgrades: improvements paid for by Google where needed.
- Phased deployment: a first phase operational by 2026 and fully completed by 2027.
The grid connection problem behind the deal
One of the most important details in the arrangement is where the projects connect. By linking both the data center and the renewable power park to the same point on the grid, the companies hope to move faster.
That speed matters because grid connection has become a bottleneck. Interconnection.fyi lists 11,860 active requests in the U.S. from power producers seeking to connect with the grid. In total, 2.05 terawatts of capacity are waiting, nearly double what is currently installed and connected.
The source says most of those requests are for solar and battery projects. That makes Google’s plan especially relevant. It is not just a clean power procurement effort; it is also an attempt to deal with the practical problem of getting new electricity resources connected in time to serve growing demand.
For AI infrastructure, delayed power can mean delayed data centers. A data center may be planned, financed, and built, but still need sufficient electricity and grid access before it can operate as intended. Google’s willingness to cover required grid upgrades suggests that grid readiness is now a central part of AI expansion.
Why the timeline matters
Google and Intersect are taking a phased approach. The first phase is expected to be operational by 2026, with full completion by 2027. The source frames that schedule as evidence of how quickly renewable power can be deployed.
That timing also creates pressure on nuclear power startups and developers. The nuclear projects mentioned in the source have longer timelines. Microsoft’s restart of a reactor at Three Mile Island is scheduled to come online in 2028. Google’s deal with small modular reactor startup Kairos has a 2030 deadline for the first of several power plants. Amazon’s contract with SMR startup X-Energy is targeting the early 2030s.
The source also notes that all of those nuclear timelines depend on the projects proceeding as planned and being completed on time. That has so far been difficult for the nuclear power industry.
The contrast is not that one energy source replaces every other. The clearer takeaway is that AI demand is moving fast, and companies need power on timelines that match their infrastructure plans. Renewable power and battery storage can be built into that schedule sooner in this case, while nuclear projects remain part of a longer-term bet.
What this signals for AI infrastructure
Google’s renewable energy deal shows that the next phase of AI competition will not be confined to software. Power availability is becoming a strategic constraint. Companies that want to build large AI data centers also need credible plans for electricity supply, storage, and grid connection.
The $20 billion figure underlines the scale of the challenge. So does the combination of partners: a technology company, a renewable developer, and an investment fund working together on infrastructure that sits outside the traditional boundaries of software and cloud services.
If the first phase reaches operation by 2026 and the full buildout is completed by 2027, the project would give Google a faster route to new carbon-free power for AI data centers. It would also provide a model for how large technology companies may approach energy development when existing grid queues and slower project timelines threaten their growth plans.
For now, the message is direct: AI needs more power, and Google is moving to build it.