d-Matrix buys Wallaroo.ai in second acquisition in four months
What's the deal? AI chip startup d-Matrix has acquired Wallaroo.aiDealroom has a profile for this one. Try Dealroom →, a maker of AI inference deployment and orchestration software. The deal, announced August 3, 2026, brings WallarooDealroom has a profile for this one. Try Dealroom →'s technology platform, intellectual property, and engineering talent to Santa Clara-based d-Matrix. Terms were not disclosed.
What's the endgame? d-Matrix builds ultra-low-latency AI inference hardware for data centers. Adding Wallaroo's software gives it an end-to-end platform spanning silicon to deployment, letting customers scale inference from single-server nodes to multi-rack setups.
Why now? This is d-Matrix's second acquisition in four months, following its purchase of GigaIODealroom has a profile for this one. Try Dealroom →'s data center business in April 2026. Both deals push the company toward rack-scale heterogeneous AI systems that pair GPUs with specialty XPUs for speed and energy efficiency.
Wallaroo's engineering, product, and go-to-market teams join d-Matrix, bringing expertise in software architecture, high-performance computing, and Kubernetes operations.
What each brings: "The biggest barrier isn't just performance, it's also the operational complexity of getting there," said Sid Sheth, founder and chief executive officer of d-Matrix. Wallaroo founder and chief executive officer Vid Jain said the combination "immediately creates an inference platform that's leaps and bounds ahead of the market."
The acquisition lands during a growth period for d-Matrix. Its Corsair inference platform is in full production and shipping to priority customers. In early July 2026, d-Matrix and ParasailDealroom has a profile for this one. Try Dealroom → deployed Corsair alongside NVIDIA Hopper and Blackwell GPUs in Parasail's cloud data center.
The signal: Chipmakers are increasingly buying software to remove deployment friction, betting that ease of use — not just raw speed — will win data center inference customers. As AI workloads grow more complex and heterogeneous, controlling the full stack from silicon to orchestration becomes a competitive edge.
Read more: PR Newswire
Image credit: cbowns
