Silicon Valley chip startup d-Matrix has secured a massive $275 million in Series C funding, reaching a $2 billion valuation as it races to solve one of artificial intelligence's most pressing challenges: the astronomical cost of running AI models at scale. The November 12, 2025 announcement positions the Santa Clara-based company as a serious challenger in the increasingly competitive AI hardware market.
The oversubscribed funding round attracted a global consortium of heavyweight investors, with BullhoundCapital, Triatomic Capital, and Singapore's Temasek co-leading the charge. Notable participants include the Qatar Investment Authority, Singapore's EDBI, and Microsoft's M12 Venture Fund, demonstrating strong international confidence in d-Matrix's approach. The round brings the company's total capital raised to $450 million since its founding in 2019.
What sets d-Matrix apart is its laser focus on AI inference rather than training. While training AI models costs millions, actually running them at scale costs billions. The company has developed what it calls Digital In-Memory Compute technology, embedding processing components directly into memory rather than separating them. This revolutionary architecture addresses the "memory wall" bottleneck that makes traditional GPU-based inference so expensive and power-hungry.
CEO and co-founder Sid Sheth emphasized that d-Matrix anticipated this shift years ago. The company predicted that once trained models needed to run continuously at scale, existing infrastructure would prove inadequate. Their solution the Corsair inference accelerator along with JetStream networking components and Aviator software reportedly delivers performance and efficiency gains that dramatically reduce operating costs for enterprises and data centers.
The technology claims are impressive: performance improvements up to ten times faster than existing solutions, costs reduced by three times, and energy efficiency gains between three to five times compared to conventional alternatives. These aren't just theoretical numbers d-Matrix has already begun shipping chips to early customers and has partnered with major players like Super Micro Computer, Arista, Broadcom, and Supermicro on its open standards-based reference architecture.
Looking ahead, d-Matrix isn't resting on current achievements. The company has already announced Raptor, its next-generation accelerator launching in 2026, which will utilize advanced 3D chip stacking and cutting-edge manufacturing processes. As AI deployment accelerates across industries, d-Matrix's timing couldn't be better the new funding will support global expansion and large-scale customer deployments at a moment when enterprises desperately need more cost-effective AI infrastructure solutions.
.jpeg)
0 Comments