AI Landscape / AI Chips & Compute
d-Matrix
Builds in-memory compute chips for AI inference.
Company profile
d-Matrix builds memory-centric compute for generative AI inference in data centers, serving hyperscale, enterprise and sovereign customers. d-Matrix websiteprnewswire.com
Its products include Corsair inference accelerators, JetStream networking accelerators and the Aviator software stack; its investors include Microsoft's M12 venture fund. prnewswire.com
- Founded
- 2018
- Founders
- Sid Sheth, Sudeep Bhoja
- Headquarters
- Not publicly reported
- Chief executive
- Sid Sheth
- Employees
- Not publicly reported
- Ownership
- Not publicly reported
- Total funding
- $450 million
- Latest funding round
- Not publicly reported
- Valuation
- Not publicly reported
- Revenue
- Not publicly reported
Drawn from Wikidata, Wikipedia, the company's own site and news and web sources; each figure links to its source. Researched Oct 19, 2026.
Latest news
d-Matrix Cuts AI Token Latency in Half With Hybrid GPU-XPU Inference at OCP
An Interview with d-Matrix CEO Sid Sheth on Why 3D-Stacked Memory Compute Is a Game Changer
d-Matrix Launches Squadron AI Inference Clusters Scaling to 1,024 XPUs
Nvidia’s Bold d-Matrix Bet Could Help the Chip Leader Expand Its AI Dominance
d-Matrix Launches Squadron AI Inference Clusters, Scaling Corsair XPUs to Data Center Megawatt Levels
Nvidia to invest in rival chip startup d-Matrix at $5B valuation
Nvidia's Reported d-Matrix and Reflection AI Moves Sit Alongside a Confirmed Microsoft Push
d-Matrix - 2026 Funding Rounds & List of Investors
NVIDIA reportedly planning to invest in rival AI chipmaker d-Matrix
Nvidia reportedly plans to invest in rival AI chip startup d-Matrix
Nvidia to invest in rival d-Matrix: report
Report: Nvidia plans to invest in inference chip startup d-Matrix
Nvidia Plans to Invest in Chip Rival d-Matrix
In the same market
NvidiaMakes the GPUs and systems most AI is trained and run on.
AMDChipmaker whose Instinct GPUs compete for AI data-centre workloads.
TenstorrentDesigns AI processors and licenses RISC-V chip technology.
CerebrasBuilds wafer-scale AI chips and offers fast inference as a service.
LightmatterDevelops photonic interconnect technology for AI data centres.
EtchedBuilds a chip designed specifically to run transformer models.
SambaNovaBuilds AI chips and systems for training and inference.
GroqDesigns inference chips (LPUs) and runs a cloud for fast model serving.