Startup
PNPreferred Networks
株式会社Preferred Networks
Not ranked in the Japan Top 100 Startups: published total includes debt and the split is not disclosed
| Founded / HQ | 2014 · Tokyo |
|---|---|
| Equity raised to date | ≥ ¥38.4Bincl. debt/grants “10億円＋9億円（2015）＋約105億円（2017）＋20億円超（2017）＋240億円（2024–25ラウンド累計）” Summed by The Raising Sun from disclosed rounds · as of 30 Apr 2025 · PFN releases (summed by us) No lifetime total published. The 2024–25 round (¥24B) mixes equity and Mizuho loans with no split; early NTT investment undisclosed; Jun 2025 extension undisclosed. |
| Last round | 30 Apr 2025 · Extension · ¥5B 総額50億円（増資＋融資） |
| Latest public valuation | Not publicly disclosed (we do not estimate) |
| Key investors | Toyota, Fanuc, Kodansha, Mitsubishi UFJ Trust |
| Unicorn | Yes |
| Website | www.preferred.jp |
From Japanese sources
Findings from Japanese-language filings, government records, trade press, interviews and job posts, with our English analysis. Items tagged Japanese-only had no English coverage when we checked quickly. Most insightful first.
| 2025-11-05 | Regulatory filingJapanese-onlyPFN posts ¥7.84B net loss after switching its fiscal year-end to June Preferred Networks、25年6月期決算は最終損失78億円 1月決算から変更、変則決算に PFN's first June-year accounts show a ¥7.84B net loss, against a tiny ¥10M profit in FY1/2024, after it moved its year-end from January. Compalyze argues the loss likely covers 17 months. Either way, chip (MN-Core) and LLM (PLaMo) spending has pushed Japan's best-known AI company back into the red. Source: gamebiz (from PFN's statutory notice) |
|---|
Overview
Preferred Networks (PFN) is a Tokyo AI company that works across the whole stack: its own MN-Core AI processors (co-developed with Kobe University), supercomputers and a cloud service built on them, the Japanese-language PLaMo foundation models, and applied products. Its MN-3 supercomputer topped the Green500 energy-efficiency ranking three times in 2020–21. It earns from joint R&D with industrial partners, compute and model services (now also offered via Microsoft Foundry) and products such as the Matlantis materials simulator and PLaMo Translate. Its 2024–25 round, led by SBI, raised ¥24B including bank loans, and Mitsubishi Heavy Industries took a stake in September 2026.
Built from: Company info · 2024 funding release
Company timeline
Founded
| Date | 2014-03-26· PFN |
|---|---|
| Founders | Toru Nishikawa, Daisuke Okanohara |
Funding
| Date | Round | Amount | Lead / key investors | Source |
|---|---|---|---|---|
| 2024-12-23 | Growth round (first close) | ¥19B incl. bank loans (split not disclosed) 本ラウンドのファーストクローズとして総額190億円 | SBI Group | PFN |
| 2025-04-30 | Growth round (extension) | ¥5B incl. loans; round total ¥24B エクステンションとして…総額50億円…本ラウンドでの累計調達額は240億円 | PFN | |
| 2026-09-16 | Strategic | Undisclosed 三菱重工とPFN、資本業務提携契約を締結 | Mitsubishi Heavy Industries | PFN |
Key hires & board
None compiled yet.
Milestones
| 2024-10 | ProductLaunches PFCP AI cloud powered by MN-Core 2· PFN |
|---|---|
| 2024-12 | ProductStarts offering PLaMo LLM· PFN |
| 2026-09-16 | PartnershipCapital and business alliance with Mitsubishi Heavy Industries· PFN |
| 2026-09-18 | ProductMN-Core 2 compute to power Matlantis materials simulator· PFN |
| 2026-10-06 | PartnershipPLaMo 3.0 Prime and PLaMo Translate available on Microsoft Foundry· PFN |
| 2026-10-15 | RecognitionMiseMise wins a 2026 Good Design Gold Award· PFN |
News
| 2026-10-15 | MiseMise wins a 2026 Good Design Gold Award MiseMiseが2026年度「グッドデザイン賞・金賞」を受賞· PFN |
|---|---|
| 2026-10-14 | Sakura Internet, PFN and NICT start "active evaluation" platform for generative AI さくらインターネット、PFN、NICT、生成AIを継続的に評価・改善する「能動的評価基盤」の社会実装に向けた取り組みを開始· PFN |
| 2026-10-13 | Survey: 52% of job seekers find AI interviews fair vs 85% for face-to-face AI面接を「公平」と感じる就活生は52%、対面面接は85% 差を埋める鍵は「対話の質」· PFN |
| 2026-10-07 | Major update to PLaMo Translate PFN、AI翻訳サービス PLaMo翻訳を大型アップデート、 世界最先端AIモデルと同等の日本語翻訳精度を大幅に低いコストで実現· PFN |
| 2026-10-06 | PLaMo 3.0 Prime and PLaMo Translate launch on Microsoft Foundry PFN、Microsoft Foundryで国産生成AI基盤モデル PLaMo 3.0 PrimeおよびPLaMo翻訳を提供開始· PFN |
| 2026-09-30 | Survey of 308 graduates on AI interviews 内定式を前に27年卒の就活生308名に聞いた「就職活動におけるAI面接の実態調査」速報· PFN |
| 2026-09-29 | PFN exhibits and speaks at CEATEC 2026 PFN、CEATEC2026に出展・登壇· PFN |
| 2026-09-29 | PFN exhibits and speaks at CEATEC 2026 PFN、CEATEC2026に出展・登壇· PR TIMES |
| 2026-09-18 | MN-Core 2 compute to power Matlantis materials simulator PFNとPFCI、独自開発AI半導体MN-Core 2による高速な計算資源を材料探索用シミュレーターMatlantis向けに年内に提供開始· PFN |
| 2026-09-16 | Mitsubishi Heavy Industries and PFN sign capital and business alliance 三菱重工とPFN、資本業務提携契約を締結· PFN |
| 2026-09-15 | Taresuka AI Interview used in Okinawa Cellular internship selection Preferred Networksの「タレスカAI面接」が沖縄セルラー電話のインターンシップ選考において「書類では見えないポテンシャル」の可視化を確認· PFN |
| 2026-09-10 | PFN starts free trial of PLaMo Translate for online meetings PFN、日本語に特化したAI翻訳サービス PLaMo翻訳 のオンライン会議翻訳機能の無料トライアルを開始· PR TIMES |
| 2026-09-09 | CMIC Group adopts PFN's Talesca AI interview for graduate hiring Preferred NetworksのAI面接サービス「タレスカAI面接」、シミックグループ新卒採用選考に採用· PR TIMES |
| 2026-07-13 | PFN's Talesca AI interview wins prize at HR Award 2026 Preferred Networksが提供するAI面接サービス「タレスカAI面接」が厚生労働省後援 日本の人事部「HRアワード2026」に入賞· PR TIMES |
| 2026-06-19 | Talesca AI interview wins Japan HR Challenge Award Preferred Networksが提供するAI面接サービス「タレスカAI面接」が「第15回 日本HRチャレンジ大賞」にて「人材サービス優秀賞（採用部門）」を受賞· PR TIMES |
| 2026-06-17 | Okinawa Cellular adopts Talesca AI interview for graduate hiring 沖縄セルラー電話がPreferred NetworksのAI面接サービス「タレスカAI面接」を新卒採用に導入· PR TIMES |
| 2026-06-04 | PFN rebrands Talent Scouter AI interview as Talesca Preferred Networks、AI面接サービス「タレントスカウター」を「タレスカ」にブランドリニューアル· PR TIMES |
| 2026-05-21 | EDION and PFN start AI store-navigation trial on MiseMise platform エディオンとPreferred Networks、次世代店舗プラットフォーム「MiseMise」を活用したAI店舗ナビゲーションの実証実験を開始· PR TIMES |
| 2025-04-30 | ¥5B extension; round total ¥24B エクステンションラウンドで総額50億円の資金調達を実施· PFN |
| 2024-12-23 | PFN Raises Total of 19 Billion Yen in Latest Round· PFN |
Founder profiles: Daisuke Okanohara, Toru Nishikawa
Funding rounds in 2026
No 2026 rounds in our database yet. Earlier funding is summarised above from the cited source.