Parasail is deploying d-Matrix Corsair inference accelerators alongside NVIDIA Hopper and Blackwell GPUs in a heterogene
Parasail, an inference cloud platform for AI-native startups, and d-Matrix, a provider of low-latency AI inference compute platforms, announced on July 8, 2026, that Parasail is deploying d-Matrix Corsair inference accelerators alongside NVIDIA Hopper and NVIDIA Blackwell architectures to deliver up to 10x faster, more cost-efficient inference services to its customers.
Parasail's Corsair deployment represents one of the first commercial-scale examples of heterogeneous disaggregated inference in production. NVIDIA AI infrastructure and d-Matrix's purpose-built Corsair inference accelerators will operate in concert, each performing computing tasks where they excel. The approach combines NVIDIA GPUs for compute-intensive prefill with d-Matrix Corsair accelerators for latency-sensitive decode, improving inference economics for select workloads.
With new data center approvals and construction taking multiple years, Parasail is leveraging heterogeneous compute with d-Matrix to obtain greater value from the NVIDIA AI infrastructure already deployed in its data centers. By pairing Corsair with Hopper and Blackwell, Parasail extends fleet performance and passes the gains directly to customers through faster, more economically efficient tokens.
Mike Henry, founder and CEO of Parasail, stated: We're relentless about delivering the best inference performance per dollar. Everyone's focused on the next GPU generation, but we run large Hopper and Blackwell fleets today and pairing them with Corsair lets us deliver our customers the edge they need while also extending the life of the hardware we've already deployed.
Sid Sheth, founder and CEO of d-Matrix, said: Parasail is proving what we've believed from the start: the future of inference is heterogeneous. GPUs and purpose-built inference accelerators aren't competitors; they're better together. By deploying Corsair alongside their NVIDIA fleet, Parasail is embracing a heterogeneous inference approach designed to optimize each phase of the workload.
Parasail's automatic kernel optimization technology dynamically routes workloads to the right hardware at the right time, extracting maximum performance and value from each model across a heterogeneous fleet. This delivers premium inference economics to Parasail's customers at scale.
d-Matrix Corsair's performance advantage stems from d-Matrix's Digital In-Memory Compute, or DIMC, chiplet architecture, which integrates compute with memory on the same silicon. In conventional systems, data must travel back and forth between separate memory and processor chips during inference, consuming time, power, and cost with every transfer. Corsair eliminates that penalty entirely, enabling up to 10x faster interactive inference and up to 3x better energy efficiency versus traditional approaches. Built on TSMC's N6 process with organic substrates and LP-DDR5 memory, Corsair is designed to support efficient, scalable inference deployments.
The companies plan to share detailed performance results and case studies following the first series of deployments and explore expanded integration across Parasail's global fleet of over 40 data centers in 15 countries. d-Matrix's Corsair inference platform is now available for select qualified customers, with inquiries directed to d-matrix.ai/contact-sales. Parasail inference services are available now at parasail.io/contact.