FuriosaAI is a South Korean semiconductor company that develops purpose-built AI accelerator chips for data-center inference and other AI workloads. Its products, including the Warboy and RNGD processors, are designed for organizations deploying large-scale machine-learning applications and infrastructure.
Subscribe to see Furiosa AI’s full profile → the full roster, investors in every round, weekly hiring history and every corporate event
Furiosa AI is an ai/ml company headquartered in South Korea. Its most recent funding round, Series D, raised $623M in October 2026. Xcout tracks 2 named people at Furiosa AI, including June Paik (CEO). Xcout recorded 109 job postings from Furiosa AI in the last 90 days, naming tools such as PyTorch, vLLM, TensorFlow.
Every round on record:
| Date | Round | Amount | Investors |
|---|---|---|---|
| Series D | $623M | — | |
| Series C | $125M | 5 named | |
| — | $6.8M | 5 named |
Who invested in each round, lead investors, valuations and the co-investor graph — open the full funding history →
Xcout recorded 109 job postings from Furiosa AI in the last 90 days, from a hiring record Xcout has kept since August 2026.
The weekly hiring trend, the roles and locations behind it — see Furiosa AI's hiring signals →
6 corporate events on record, with sources — see the full timeline →
Based on 70 job postings in the last 8 weeks (7 Sep – 1 Nov 2026), compared with 400 scale-up technology companies.
This describes a hiring pattern — how often Furiosa AI’s job postings ask for AI skills, and how that changed against the previous 8 weeks, ranked among its peers. It says nothing about the company’s products or performance.
Signed-in members see the AI-skill share and its change with peer percentiles, role mix, posting momentum and advertised US pay against the market.
A few of Furiosa AI's closest competitors — see the full list of competitors and alternatives →
SiMa.ai is a physical AI company that scales machine learning applications across robotics, automotive, industrial automation, and smart vision sectors. The firm provides purpose-built hardware and software platforms to accelerate edge AI deployments.
femtoAI is a privately held semiconductor manufacturing and AI company headquartered in San Bruno, California. It provides a full-stack AI silicon and compression platform.
Tensordyne designs and manufactures next-generation AI inference systems and custom silicon optimized for data centers, hyperscalers, and enterprises. Its flagship product, the Tensordyne Napier system, utilizes proprietary logarithmic mathematics and low-latency interconnects to accelerate multi-modal generative AI, Mixture of Experts, and agentic workloads with high energy efficiency.
Agrani Labs builds AI GPUs and end-to-end software for enterprise and datacenter compute, operating as a semiconductor and technology provider. The company's primary corporate address and headquarters are located in Fremont, California.
SAPEON is a South Korean AI semiconductor company spun out of SK Telecom that develops AI accelerator chips, particularly for data-center inference workloads. Its products are intended for cloud providers, data-center operators, and enterprises needing energy-efficient high-performance AI computing.
HyperAccel is a fabless semiconductor company developing the world's first LPU AI chips and user-friendly software platforms to enable the widespread deployment of generative AI. The company specifically targets efficient inference over training to bridge modern AI applications and hardware.
Funding rounds, acquisitions, new executives, launches and expansions, each with a two-line summary and a link to the source. One short email a day, only when there is news.
Free: up to 5 companies, daily. Subscribers also follow whole sectors and saved filters (M&A in Israel, AI funding…), get instant alerts, in-app notifications and signed webhooks, and open the full profile behind every name. Subscribe →
Something wrong or missing about Furiosa AI? Suggest a correction · Work at Furiosa AI? Claim this profile
Post the link and the card appears automatically — each network fetches the image itself.