Report
Deal summary
Published
FuriosaAI announced a $620 million Series D on October 28, 2026, at a reported $2.9 billion valuation.[2] SiliconANGLE reports that South Korea’s National Growth Fund led with $280 million; Cisco Investments, NeoFort Holdings, QUDA Capital Partners, Jericho Capital and Worldline Capital together added $110 million; existing backers supplied the remainder.[1] The linked BusinessWire page was unavailable in the fetch, so these terms and investor allocations are supported here by the article’s report rather than independently confirmed against that announcement.
Source extracts (2)
- [1]
The investment was led by South Korea’s National Growth Fund, which injected $280 million of the total amount raised
Return to text - [2]
$620 million Series D
Return to text
Market impact
Published
The event aligns with a power-constrained data-center market: IEA-focused coverage published April 23, 2025 reports 415 TWh annual data-center electricity use in 2024 and projects more than 945 TWh by 2030, while noting grid constraints could delay up to 20% of planned projects and grid interconnection backlogs of four to eight years in some advanced economies.[1][2] That context supports a potential buyer case for efficient inference silicon at sites where power capacity is scarce, but does not establish FuriosaAI has won demand or displaced Nvidia. For FuriosaAI, the commercial test is whether its planned production can convert efficiency claims and customer deployments into repeatable shipments; incumbent GPU suppliers face an efficiency-driven alternative, not evidenced lost sales in the sources.
Source extracts (3)
- [1]
global electricity demand from data centers will more than double by 2030, surging past 945 TWh
Return to text - [2]
as much as 20% of planned global data center projects face potential delays due to electric grid constraints
Return to text - [3]
use less than half the power of Nvidia’s GPUs
Parties involved
Published
FuriosaAI is the South Korean inference-chip developer and Series D issuer; its record identifies it as founded in 2017 and operating. National Growth Fund led the financing, while Cisco Investments, QUDA Capital Partners and Jericho Capital are named among the new global investors; Korea Development Bank and NAVER are cited as existing backers, and NAVER Ventures is identified as a repeat participant since the 2017 pre-seed. SiliconANGLE says FuriosaAI develops energy-efficient Tensor Contraction Processor chips based on RNGD and reports deployments with LG Electronics’ EXAONE models, Upstage search summaries, and Samsung’s NPU-as-a-service. Broadcom is described as a co-development partner on the planned third-generation processor. The article casts Nvidia as the incumbent rival; these claims describe roles in the reported event and do not independently establish investor check sizes beyond those expressly reported.
Source extracts (3)
- [1]
The investment was led by South Korea’s National Growth Fund
- [2]
LG Group, the parent company of LG Electronics Co. Ltd., is using FuriosaAI’s chips to power its EXAONE models
- [3]
it’s also working on a third-generation processor that’s being co-developed with Broadcom Inc.
Outlook
Published
FuriosaAI says proceeds will scale next-generation Tensor Contraction Processor production. Its stated targets are 20,000 units in 2026 and 40,000 in 2027, with further scaling thereafter; it is also developing a third-generation processor with Broadcom.[1] The next observable checkpoints are production and shipment evidence against those annual unit targets and progress on Gen 3 commercialization. Failure to approach the 2026 target or delays to Gen 3 would weaken the scale-up thesis; the report does not establish these future milestones as completed.
Source extracts (2)
- [1]
It’s planning to produce 20,000 this year and double that to 40,000 in 2027
Return to text - [2]
By accelerating the commercialization of RNGD and advancing the development of Gen 3