d-Matrix Announces JetStream I/O Accelerators Enabling Ultra-Low Latency for AI Inference at Scale

Paired with d-Matrix's Corsair accelerators and Aviator software, JetStream can deliver up to 10x speedups, 3x better cost-performance, and 3x higher energy efficiency than GPU-based offerings

SANTA CLARA, Calif., Sept. 8, 2025 /PRNewswire/ -- d-Matrix^® today announced the expansion of its AI product portfolio with d-Matrix JetStream™, a custom I/O card designed from the ground up to deliver industry-leading, data center–scale AI inference.

With millions of people now using AI services – and the rise of agentic AI, reasoning, and multi-modal interactive content – the industry's focus is quickly shifting from model training to deploying AI at ultra-low latency across multiple users. 

When combined with d-Matrix Corsair™ accelerators and d-Matrix Aviator™ software, JetStream I/O accelerators can scale to state-of-the-art models exceeding 100B parameters, delivering 10x the speed, 3x better cost performance, and 3x greater energy efficiency than GPU-based solutions.¹

JetStream's addition to the company's product portfolio makes d-Matrix one of the few AI infrastructure providers offering a complete platform that spans compute, software, and networking.

"JetStream networking comes at a time when AI is going multimodal, and users are demanding hyper-fast levels of interactivity," said Sid Sheth, d-Matrix co-founder and CEO. "Through JetStream, together with our already-announced Corsair compute accelerator platform, d-Matrix is providing a path forward that makes AI both scalable and blazing fast."

JetStream is a transparent NIC and streaming solution optimized for d-Matrix Corsair accelerators. Packaged in an industry-standard PCIe format and compatible with off-the-shelf Ethernet switches, JetStream NICs are easy to deploy within existing data centers – eliminating the need for costly infrastructure replacement.

Availability
JetStream NICs are full-height PCIe Gen5 cards delivering a maximum 400Gpbs bandwidth. Samples are available now, with full production expected by year-end.

To learn more, go to www.d-matrix.ai and follow d-Matrix on LinkedIn for the latest updates.

About d-Matrix
d-Matrix advances compute for modern AI workloads and breaks through key barriers that have stymied chips for AI, including performance, efficiency, and scalability. d-Matrix is a pioneer in Digital In-Memory Computing (DIMC) solutions that solve limitations on generative AI inference acceleration. d-Matrix creates flexible solutions for inference at scale using innovative circuits and packaging techniques, a chiplet-based DIMC architecture, innovative interconnect fabric, and hardware-software codesign. Founded in 2019, the company is backed by top investors and strategic partners including Playground Global, M12 (Microsoft Venture Fund), Triatomic Capital, SK Hynix, Nautilus Venture Partners, Marvell Technology and Entrada Ventures.

The terms d-Matrix, JetStream, Corsair and Aviator are trademarks and/or registered trademarks of d-Matrix, Inc. in the U.S. and other countries. All rights reserved.

¹ For Llama70B model on 8 Corsair servers. Performance, cost and power estimates are preliminary and subject to change. Results may vary.

SOURCE d-Matrix