TAR raised $120 million for off-grid AI power. The company will use its Series A funding to develop power systems for AI data centers, a business built around the growing strain between computing demand and available electricity.
The round values TAR at a $1 billion post-money valuation. Spark Capital led the investment, with existing investors Buckley Ventures and Align Fund also participating.
Pat Becker and Leonhard Soenke founded TAR to build off-grid power systems for AI data centers. The company’s central proposition is direct: data centers should not need to wait for the grid before they can add computing capacity.
That proposition arrives as electricity demand rises. The U.S. Energy Information Administration expects electricity use in the United States to reach 4,135 billion kilowatt-hours in 2026, almost 2% above 2025 levels, with demand set to rise by nearly another 2% in 2027.
AI’s Power Problem Is Becoming Harder to Ignore
Texas offers the clearest evidence of the pressure. Average hourly electricity demand on the ERCOT grid reached 74.5 gigawatts, while record electricity demand in Texas was recorded on August 22, 2026.
Texas authorities are reviewing data center connections as demand tests the limits of existing infrastructure. The U.S. Department of Energy is also reporting on transmission capacity, adding another layer to a problem that is no longer confined to data center operators and their utility providers.
Will Reed, a General Partner at Spark Capital, described the issue in blunt terms: “Power is becoming the main bottleneck to scaling compute. We will need innovation, unprecedented speed and exceptional companies unblocking it at every level.”
That statement explains the investment more clearly than a dozen polished funding announcements. AI companies can secure models, chips, and capital, but none of those assets run without electricity. The grid remains the part of the stack that refuses to accept a slide deck as infrastructure.
A Crowded Market for AI Capital
TAR’s financing came on September 11, 2026, one day after Arlequin AI raised €28 million and Clay raised $115 million on September 10. Those deals place TAR inside a busy stretch of AI fundraising, but its focus separates it from companies building software or model-based products.
TAR is targeting the physical constraint beneath AI expansion: power availability for data centers. Its approach does not replace the need for transmission capacity or change the electricity forecasts from the U.S. Energy Information Administration, but it aims to develop an alternative power path for AI facilities.
The scale of TAR’s valuation shows how investors are pricing that constraint. A $120 million Series A at a $1 billion post-money valuation gives the company substantial backing before it has solved the broader infrastructure problem—because apparently the next phase of AI requires both more intelligence and its own power supply.
The company now has funding to develop its off-grid systems while demand continues to climb. Whether that approach can match the pace of AI data center growth will depend on the systems TAR builds and the connections Texas authorities review.
