D-MATRIX

AI Focus Areas

• AI inference accelerators
• In‑memory computing architectures
• Data center AI hardware
• LLM and transformer inference optimization
• Low‑latency, low‑power AI serving

Key Products

• d-Matrix Corsair AI accelerator
• Inference compute platform for data centers
• Software stack for model deployment on d-Matrix hardware

Market Position

D-Matrix targets the lucrative inference segment of the AI hardware market with an in‑memory compute architecture designed to reduce data movement and improve energy efficiency. Backed by strategic investors such as Microsoft’s M12 and major sovereign funds, it aims to compete with Nvidia GPUs and emerging accelerators from startups and hyperscalers. Its niche is high‑throughput, cost‑efficient serving of large language models and other transformer workloads in data centers. Success depends on demonstrating significant total‑cost‑of‑ownership advantages and building a robust software ecosystem so that AI teams can adopt its hardware without major rewrites.

AGI Relevance

AGI‑scale systems will need highly efficient inference hardware to deliver capabilities to billions of users without prohibitive energy and capital costs. D-Matrix’s work on new architectures that minimize memory bottlenecks contributes to the broader search for post‑GPU AI hardware. If in‑memory computing proves practical at scale, it could enable much larger or more numerous models to be deployed within the same power envelope, accelerating experimentation with agentic and always‑on systems. Close collaboration with major cloud providers also means any breakthroughs can propagate quickly into the infrastructure used by leading AI labs.

Investment Highlights

In November 2025, D-Matrix raised $275M at a $2B valuation in a round co‑led by QIA, Temasek, Bullhound and others, with participation from Microsoft’s M12 and prior investors, bringing total funding to ~$450M per Bloomberg and TechCrunch.

Tags

• ai company