AI BriefingFuriosaAI closes $620M Series D at $2.9B valuation to scale RNGD inference accelerator production
·2026.10.29 08:31
Key point
The company plans to produce 20,000 RNGD cards this year and more than double that in 2027, with the chip already in mass production and deployed by customers like Samsung SDS and LG Group.
Details
- FuriosaAI closed a $620 million Series D round at a $2.9 billion post-money valuation, with participation from Cisco Investments, Infinitum, and Korea’s National Growth Fund.
- The RNGD inference accelerator is in mass production, with plans to produce 20,000 cards in 2026 and more than double that output in 2027.
- RNGD is a 180-watt, air-cooled PCIe card that matches RTX PRO 6000 throughput at the same card count while serving twice as many users per kW of power.
- The company is developing a third-generation chip with Broadcom using a 2nm, multi-die design and HBM4/4E memory to serve frontier agentic models.
- SDK 2026.4, released in October 2026, raises token generation per watt across 18 models compared to version 2026.3, including a 138% increase on gpt-oss-120B.
- The architecture relies on the Tensor Contraction Processor (TCP), which uses higher-dimensional tensor contraction as the basic unit of work to improve efficiency and programmability.
This summary was generated automatically by AI. Check the original for the author's claims and context. Copyright belongs to the original author.
Our guide explains how the AI works. Report summary errors, attribution issues, or removal requests via Contact.
Related
Intel Announces D270 GPU with 480 GB LPDDR5X for Agentic AI Inference