Test report DSG-2315 · Rev D · tested November 2, 2026

Processors & AcceleratorsDevice under test

FuriosaAI raises $620m Series D at $2.9bn valuation

Seoul-based FuriosaAI raised $620m in a Series D round at a $2.9bn valuation, with Korea's National Growth Fund putting in $280m ahead of a planned IPO.

Read
2 min
Words
498
Node
20nm
Operator
Marcus Bennett

Spec summary

  1. FuriosaAI raised $620m in Series D at roughly $2.9bn post-money valuation
  2. South Korea's National Growth Fund invested $280m; international investors added $110m
  3. RNGD output targets 20,000 units in 2026, rising to over 40,000 in 2027
  4. The company rejected an $800m Meta acquisition offer in March 2024
  5. Gen 3 accelerator is co-developed with Broadcom under a May 2026 partnership
FuriosaAI raises $620m at a post-money valuation of approximately $2.9bn
Fig. AFuriosaAI raises $620m at a post-money valuation of approximately $2.9bn — Igor Omilaev / unsplash

FuriosaAI has raised $620 million in a Series D round at a post-money valuation of roughly $2.9 billion. The round lands ahead of an anticipated IPO and more than covers the $300–500 million target the company was reportedly chasing in January 2026.

South Korea's National Growth Fund led the financing with $280 million. International investors added another $110 million combined, including QUDA Capital Partners, Cisco Investments, NeoFort Holding, Jericho Capital and Worldline Capital.

The numbers matter because they mark a step change for a company that had previously raised about $115 million across four funding rounds since its founding in 2017.

What does the money buy?

FuriosaAI says the funds serve three purposes:

  • expand production of its RNGD inference accelerators
  • strengthen its software ecosystem
  • accelerate development of its third-generation AI inference accelerator

The RNGD chip is a 180W, PCIe-based accelerator currently in mass production on TSMC's 5nm process node. Volume is the constraint the round addresses. The company expects to produce 20,000 RNGD units in 2026 and more than double that figure to over 40,000 units in 2027, citing growing customer demand.

Where does Gen 3 fit in?

The third-generation accelerator is the strategic piece. In May 2026, FuriosaAI announced a partnership with Broadcom for its development. The design targets high-bandwidth, rack-scale networking across large AI compute clusters, which places it in direct competition with the top tier of inference silicon.

CEO June Paik framed the round as validation from investors who know the sector. "This Series D financing marks an important milestone for FuriosaAI and reflects the confidence of investors around the world in our technology and ability to execute," Paik said. "We are particularly encouraged by the participation of investors with deep expertise and experience in the global AI semiconductor and infrastructure industries."

He added: "We are entering a new phase of bringing our technology to the global AI infrastructure market at scale. By accelerating the commercialization of RNGD and advancing the development of Gen 3, we aim to help shape the next generation of AI infrastructure."

What does this change for the market?

For buyers of inference hardware, it means one more independent supplier with funded capacity. Founded in 2017 and headquartered in Seoul, FuriosaAI has already shown it intends to stay independent: in March 2024 it turned down an $800 million acquisition offer from Meta, reportedly because it preferred to grow on its own rather than fold into a hyperscaler.

The Series D gives the company roughly six times its lifetime prior funding in a single round. That capital covers a production ramp, a software push and co-development with Broadcom — the standard checklist for an AI chip startup positioning itself for a public listing.

The IPO timeline remains anticipated rather than dated. The valuation jump to $2.9 billion, on paper at least, values the company at more than three times the Meta offer it walked away from two years earlier.

via Data Center Dynamics (Source)

Filed under

  • furiosaai
  • ai-inference-accelerators
  • rngd
  • series-d-funding
  • tsmc
Share this article:

More from Marcus Bennett

Marcus Bennett

Show full bio

News editor covering regulation, policy and their impact on hardware layer of artificial intelligence at Die Signal.

591 articles

Same lot · LOT-3609

« Previous articleNext article »