Nvidia is planning to invest in d-Matrix, a startup building chips for AI inference, according to a report by The Information cited by Data Center Dynamics. The reported move would tie Nvidia more closely to a maker of rival accelerator designs, as the fight over how AI workloads are run reaches beyond training chips.
NVIDIA Watch analysis
What happened
Data Center Dynamics says The Information, citing three people with knowledge of the deal, reported that Nvidia’s planned investment is part of an effort to make its technology compatible with rival chip designs. The report does not state the size or terms of the proposed investment.
The account follows d-Matrix’s adoption of Nvidia’s NVLink technology, announced a month earlier, to connect its Raptor processors with Nvidia hardware in rack-scale systems. NVLink Fusion is designed to let partners link non-Nvidia accelerators with Nvidia hardware.
Why it matters
Inference is the work of running a trained AI model, and demand for it is drawing attention to systems beyond Nvidia’s own accelerators. If Nvidia invests in a company making rival chips while also helping connect those chips to its hardware, that points to a strategy built partly around keeping Nvidia technology in the wider data-centre mix.
The practical significance will depend on what the reported investment involves and what products or partnerships follow. For now, the proposed deal is a reported plan, not a completed transaction or a disclosed commitment.
Our read
There is a neat strategic tension here: Nvidia’s influence may grow not only by selling its own chips, but by making its hardware useful alongside other people’s. That could widen the reach of its technology, while giving d-Matrix another route into complex AI infrastructure. The details matter, though; a reported investment is not yet a working system or a signed cheque of known size.
What to watch
- Whether Nvidia or d-Matrix confirms the planned investment.
- Any disclosed terms, size or timing for the proposed deal.
- Whether further NVLink Fusion partnerships produce systems that combine Nvidia hardware with rival accelerators.
Discussion spark: If Nvidia makes its technology easier to pair with rival AI chips, does that strengthen its position as the industry’s platform, or make competitors less dependent on Nvidia?
Sources and evidence
- Report: Nvidia plans to invest in inference chip startup d-Matrix – Data Center Dynamics (9 October 2026, 12:02 UTC)
Independent WittyWires tracker for public updates about NVIDIA. Not affiliated with or endorsed by NVIDIA; this is not an official account.