Electricity is becoming the AI industry’s scarcest input, and Velaura AI has raised $110M in a Series A that values the startup above $1B to attack the problem inside the chip.
Its Titan Core platform claims a 2x to 4x improvement in performance per watt on the mathematical operations inside AI accelerators, and the company says the technology already runs inside more than 30 million shipped ASICs.
Velaura is chasing two markets at once: hyperscale data centers fighting grid constraints, and physical AI systems such as robots and drones that must run demanding models on tight power budgets. It says leading hyperscalers are evaluating the tech for future chip roadmaps.
The cap table mixes fresh names with repeat backers. Seligman Ventures led; Capricorn Investment Group and Prosperity7 Ventures invested for the first time; Mayfield, Maverick Silicon, MARA, Premji Invest, Samsung Catalyst Fund and StepStone Group all returned.
CEO Rajiv Khemani, whose engineers come from Apple, Nvidia, Google, Qualcomm and Marvell, argues the next era of AI will hinge on “fundamentally better compute economics” as much as better models.
Chip startups face long qualification cycles, but the unicorn price tag shows investors treating efficiency as the next frontier of the AI race.
