AI Inference Powerhouse d-Matrix Secures $275 Million Series C Funding
November 13, 2025, 9:32 am
Location: United States, California, San Francisco
Employees: 11-50
Founded date: 2000
d-Matrix secured $275M Series C funding, pushing its valuation to $2 billion. Total capital raised now stands at $450M since 2019. This substantial investment powers the firm's global expansion, accelerates its product roadmap, and enables massive deployments. d-Matrix provides cutting-edge generative AI inference compute solutions for data centers. Its platform integrates compute-memory, high-speed networking, and optimized software. It delivers 10x faster performance, 3x lower cost, and 5x greater energy efficiency than traditional GPU systems. The technology excels at handling Llama 70B models, achieving 30,000 tokens per second. This innovation addresses critical sustainability and economic challenges in the rapidly expanding AI inference market. Hyperscalers and enterprises gain significant benefits. The funding round drew a global consortium of investors, signaling strong confidence in d-Matrix’s vision for next-generation AI infrastructure.
d-Matrix just secured a massive $275 million in Series C funding. This round values the company at $2 billion. Total capital raised now stands at $450 million. This significant investment fuels ambitious global expansion. It accelerates the company's product roadmap. It supports large-scale deployments of its cutting-edge AI inference technology. The market is taking notice.
The AI landscape is undergoing a fundamental transformation. Initial focus was on AI model training. That paradigm is shifting. Efficient inference is now paramount. Running sophisticated trained models at massive scale presents new challenges. Existing infrastructure, often GPU-centric, struggles with these demands. d-Matrix foresaw this impending bottleneck years ago. It proactively engineered a robust solution. The industry has clearly entered a new phase. This is the "Age of AI Inference." Economic realities are changing. Inference costs increasingly dominate overall AI system expenses. Companies need smarter compute.
d-Matrix developed a comprehensive, full-stack inference platform. Its core architecture boasts innovative compute-memory integration. This design is crucial. High-speed networking capabilities further enhance efficiency. Proprietary, inference-optimized software binds the system together. This integrated approach delivers groundbreaking performance metrics. It achieves speeds 10 times faster than conventional GPU-based systems. Operating costs drop dramatically, by a factor of three. Energy efficiency is also vastly improved, up to five times greater. These advancements directly address critical sustainability concerns. One d-Matrix-powered data center can effectively manage the workload of ten traditional facilities. This significant reduction in energy consumption is vital. It also substantially boosts profitability for hyperscalers and major enterprises deploying large-scale AI.
The company's product suite is engineered for demanding AI workloads. Corsair™ inference accelerators form the hardware backbone. JetStream™ NICs ensure ultra-fast data flow. The Aviator™ software suite provides critical optimization. This powerful combination yields impressive real-world results. Data centers leveraging d-Matrix technology can produce up to 30,000 tokens per second. This is achieved at an incredibly low latency of 2 milliseconds per token. Such performance has been validated on the demanding Llama 70B model. A key architectural advantage is its compute-dense design. This allows customers to efficiently run vast 100B-parameter models. All this power fits within a single rack. This level of performance and density is unprecedented. It fundamentally redefines the economic viability of advanced AI inference infrastructure.
The $275 million Series C funding round was significantly oversubscribed. This unequivocally signals strong investor confidence in d-Matrix's vision and technology. A diverse, global consortium co-led this crucial round. BullhoundCapital, Triatomic Capital, and Temasek were prominent co-leads. Key participation also came from the Qatar Investment Authority (QIA) and EDBI. Existing institutional investors reaffirmed their support. M12, Microsoft’s Venture Fund, was among them. Nautilus Venture Partners, Industry Ventures, and Mirae Asset also rejoined. Financial advisory was handled by Morgan Stanley, serving as the exclusive placement agent. Legal counsel was provided by Wilson Sonsini Goodrich & Rosati.
This substantial funding will accelerate d-Matrix's already impressive global expansion. The company maintains a strategic international presence. Its headquarters are in Santa Clara, California. Additional offices span Toronto, Sydney, Bangalore, and Belgrade. The global team exceeds 250 skilled professionals. This workforce continues to grow. Strong partnerships further solidify the ecosystem. Recent collaborations include industry leaders like Arista, Broadcom, and Supermicro. These partners integrate the d-Matrix SquadRack™ reference architecture.
Sid Sheth, CEO and co-founder, leads d-Matrix. Sudeep Bhoja serves as the highly capable CTO. Their combined vision targeted AI inference from the company's inception in 2019. They correctly predicted the looming infrastructure challenges. They understood the necessity for scalable and efficient inference compute. This foresight is now profoundly validated by market trends. The company's singular focus on inference sets it apart. It built a truly new architecture. This architecture enables ubiquitous, continuous AI operation. The proprietary digital in-memory compute architecture is purpose-built. It excels at processing low-latency, high-throughput inference workloads. This strategic clarity drives their market leadership.
d-Matrix's innovation path continues relentlessly. Its future roadmap includes groundbreaking advancements. Significant research focuses on next-generation 3D memory stacking technologies. Further enhancements in inference-optimized compute are also planned. The company firmly positions itself as a foundational pillar. It is building the future AI infrastructure stack. This caters to hyperscale deployments, large enterprises, and sovereign customers alike. The explosive demand for AI inference is undeniable. Efficiency and scalability are becoming primary drivers for revenue capture and profitability. d-Matrix uniquely addresses modern unit economics. It optimizes Large Language Model inference across a range of rapidly growing model sizes. Its in-memory product architecture provides sustainable Total Cost of Ownership benefits. It achieves leading metrics in both latency and throughput. The firm is actively shaping what is economically feasible within the rapidly evolving AI deployment landscape.
d-Matrix just secured a massive $275 million in Series C funding. This round values the company at $2 billion. Total capital raised now stands at $450 million. This significant investment fuels ambitious global expansion. It accelerates the company's product roadmap. It supports large-scale deployments of its cutting-edge AI inference technology. The market is taking notice.
The AI landscape is undergoing a fundamental transformation. Initial focus was on AI model training. That paradigm is shifting. Efficient inference is now paramount. Running sophisticated trained models at massive scale presents new challenges. Existing infrastructure, often GPU-centric, struggles with these demands. d-Matrix foresaw this impending bottleneck years ago. It proactively engineered a robust solution. The industry has clearly entered a new phase. This is the "Age of AI Inference." Economic realities are changing. Inference costs increasingly dominate overall AI system expenses. Companies need smarter compute.
d-Matrix developed a comprehensive, full-stack inference platform. Its core architecture boasts innovative compute-memory integration. This design is crucial. High-speed networking capabilities further enhance efficiency. Proprietary, inference-optimized software binds the system together. This integrated approach delivers groundbreaking performance metrics. It achieves speeds 10 times faster than conventional GPU-based systems. Operating costs drop dramatically, by a factor of three. Energy efficiency is also vastly improved, up to five times greater. These advancements directly address critical sustainability concerns. One d-Matrix-powered data center can effectively manage the workload of ten traditional facilities. This significant reduction in energy consumption is vital. It also substantially boosts profitability for hyperscalers and major enterprises deploying large-scale AI.
The company's product suite is engineered for demanding AI workloads. Corsair™ inference accelerators form the hardware backbone. JetStream™ NICs ensure ultra-fast data flow. The Aviator™ software suite provides critical optimization. This powerful combination yields impressive real-world results. Data centers leveraging d-Matrix technology can produce up to 30,000 tokens per second. This is achieved at an incredibly low latency of 2 milliseconds per token. Such performance has been validated on the demanding Llama 70B model. A key architectural advantage is its compute-dense design. This allows customers to efficiently run vast 100B-parameter models. All this power fits within a single rack. This level of performance and density is unprecedented. It fundamentally redefines the economic viability of advanced AI inference infrastructure.
The $275 million Series C funding round was significantly oversubscribed. This unequivocally signals strong investor confidence in d-Matrix's vision and technology. A diverse, global consortium co-led this crucial round. BullhoundCapital, Triatomic Capital, and Temasek were prominent co-leads. Key participation also came from the Qatar Investment Authority (QIA) and EDBI. Existing institutional investors reaffirmed their support. M12, Microsoft’s Venture Fund, was among them. Nautilus Venture Partners, Industry Ventures, and Mirae Asset also rejoined. Financial advisory was handled by Morgan Stanley, serving as the exclusive placement agent. Legal counsel was provided by Wilson Sonsini Goodrich & Rosati.
This substantial funding will accelerate d-Matrix's already impressive global expansion. The company maintains a strategic international presence. Its headquarters are in Santa Clara, California. Additional offices span Toronto, Sydney, Bangalore, and Belgrade. The global team exceeds 250 skilled professionals. This workforce continues to grow. Strong partnerships further solidify the ecosystem. Recent collaborations include industry leaders like Arista, Broadcom, and Supermicro. These partners integrate the d-Matrix SquadRack™ reference architecture.
Sid Sheth, CEO and co-founder, leads d-Matrix. Sudeep Bhoja serves as the highly capable CTO. Their combined vision targeted AI inference from the company's inception in 2019. They correctly predicted the looming infrastructure challenges. They understood the necessity for scalable and efficient inference compute. This foresight is now profoundly validated by market trends. The company's singular focus on inference sets it apart. It built a truly new architecture. This architecture enables ubiquitous, continuous AI operation. The proprietary digital in-memory compute architecture is purpose-built. It excels at processing low-latency, high-throughput inference workloads. This strategic clarity drives their market leadership.
d-Matrix's innovation path continues relentlessly. Its future roadmap includes groundbreaking advancements. Significant research focuses on next-generation 3D memory stacking technologies. Further enhancements in inference-optimized compute are also planned. The company firmly positions itself as a foundational pillar. It is building the future AI infrastructure stack. This caters to hyperscale deployments, large enterprises, and sovereign customers alike. The explosive demand for AI inference is undeniable. Efficiency and scalability are becoming primary drivers for revenue capture and profitability. d-Matrix uniquely addresses modern unit economics. It optimizes Large Language Model inference across a range of rapidly growing model sizes. Its in-memory product architecture provides sustainable Total Cost of Ownership benefits. It achieves leading metrics in both latency and throughput. The firm is actively shaping what is economically feasible within the rapidly evolving AI deployment landscape.