What did Velaura AI announce about its $110 million Series A funding round—including its valuation, lead and participating investors, low-power chip and software technologies for AI data centers and physical AI applications, planned use of the funds, Titan Core chip-design platfo

Velaura AI announced a $110 million Series A financing at a valuation above $1 billion to develop ultra-low-power AI-compute infrastructure for hyperscale data centers and physical-AI systems such as robots, drones, and autonomous machines. 1

Funding and planned use

• Seligman Ventures led the round; new investor Capricorn Investment Group participated, alongside existing investors Mayfield, Maverick Silicon, MARA, Premji Invest, Samsung Catalyst Fund, and StepStone Group. 1
• The company said it would use the capital to develop and commercialize its AI-compute portfolio—particularly Titan Core—and expand engineering and customer-facing teams while deepening collaborations with strategic partners and customers. 1

Technology, customers, and commercial model

• Velaura develops silicon and software intended to reduce AI data-center power consumption and operating costs, while also serving power- and thermally constrained physical-AI applications. 1
• Its Titan Core proprietary digital chip-IP and design platform is intended to improve AI-accelerator mathematical-operations performance per watt by 2–4x without sacrificing performance; Velaura said the underlying technology has been validated through deployment in more than 30 million production ASICs. 1
• The company cited multiple hyperscaler engagements. Reuters also reported that CEO Rajiv Khemani said it was in active discussions with three of the four largest cloud-computing providers, though the companies were not named. 1
• Velaura’s model, as described in reporting on the deal, is an upfront payment combined with royalties tied to customers’ power savings. Insufficient evidence in the supplied material establishes the exact royalty formula or commercial terms.

Views from Khemani and Padval

• Khemani argued that AI’s binding constraint is increasingly electricity rather than demand for computation: advances in reasoning models and embodied intelligence require more compute and therefore more power. He said the next phase of AI will depend on “fundamentally better compute economics,” enabled by ultra-low-power silicon and software from hyperscale facilities to intelligent machines. 1
• Umesh Padval, Seligman Ventures’ managing partner, called physical AI a major next frontier that requires extreme power efficiency. He said Velaura was Seligman’s first physical-AI investment, citing its experienced leadership, low-power silicon, software and systems expertise, deployment in more than 30 million ASICs, and expanding opportunities in physical AI and hyperscale infrastructure as reasons to lead the investment. 1