Matsuo Lab (UTokyo)

Matsuo Lab is the artificial intelligence research laboratory of Yutaka Matsuo at the University of Tokyo Graduate School of Engineering, in Tokyo, Japan. The lab, formally part of the Matsuo Iwasawa Lab (松尾岩澤研究室) since the 2024 promotion of Shixiang Iwasawa to Associate Professor, conducts research across deep learning, reinforcement learning, world models, robotics, and other application areas. Matsuo serves as Professor at the University of Tokyo and is one of the principal Japanese AI policy advisors, having chaired the Japanese government's AI Strategy Council (人工知能戦略会議) since its 2023 founding under the Cabinet Office. As of April 2026, Matsuo Lab is one of the principal Japanese academic AI research labs and has produced Japanese AI startup founding teams including PKSHA Technology, Naco-do, ELYZA, DeepX, Acculus, and other peers.

At a glance

• Founded: Roughly 2002 in Tokyo, Japan, when Yutaka Matsuo joined the University of Tokyo faculty.
• Status: Research laboratory within the University of Tokyo Graduate School of Engineering. Renamed Matsuo Iwasawa Lab in 2024 following Shixiang Iwasawa's Associate Professor promotion.
• Funding: University of Tokyo academic funding plus Japanese government research grants (JST CREST, JSPS Kakenhi, METI projects), Japanese industrial-cooperation funding, and selected international research grants.
• Director: Yutaka Matsuo, Professor at the University of Tokyo Graduate School of Engineering. Chair of the Japanese government's AI Strategy Council since 2023. Japanese AI research and policy leader.
• Other notable leadership and faculty: Shixiang Iwasawa, Associate Professor (since 2024). Alumni including the founders of PKSHA Technology, Naco-do, ELYZA, DeepX, and other Japanese AI startups.
• Open weights: Yes, partial. Selected research outputs released open-source through GitHub. Matsuo Lab alumni have produced open-research output through commercial spinouts including ELYZA's Japanese-language foundation models.
• Flagship outputs: Published research output across deep learning, reinforcement learning, world models, robotics, and other application areas. Alumni network producing principal Japanese AI startups. Japanese AI policy contributions through the AI Strategy Council and other government advisory roles.

Origins

Matsuo Lab was established roughly 2002 when Yutaka Matsuo joined the University of Tokyo faculty after completing his PhD at the University of Tokyo (2002) and serving in adjacent academic and research roles. Matsuo's early research focused on web mining, social network analysis, and other areas before transitioning to deep learning research in the 2010s alongside the broader deep-learning era.

The 2010s and 2020s saw Matsuo Lab's transition to deep-learning-era AI research. The lab anchored Japanese AI research talent recruitment through undergraduate and graduate research participation, with Matsuo Lab alumni founding multiple principal Japanese AI startups: PKSHA Technology (founded 2012, IPO 2017), Naco-do, ELYZA (founded 2018), DeepX (founded 2016), and other Japanese AI ventures.

The 2023 founding of the Japanese government's AI Strategy Council under Yutaka Matsuo's chairmanship anchored Japanese AI policy direction through 2023 to 2026. The 2024 transition to the Matsuo Iwasawa Lab structure, with Shixiang Iwasawa's Associate Professor promotion, has continued the lab's research-program continuity through senior faculty transition.

Mission and strategy

Matsuo Lab's stated mission is to advance Japanese AI research and to translate AI research into Japanese commercial applications. The lab's strategic premise reflects University of Tokyo's broader research-university positioning combined with Japanese-industrial-AI commercial-translation positioning.

The strategy combines three threads. First, faculty-led AI research across deep learning, reinforcement learning, world models, robotics, and other areas. Second, commercial-translation activity through alumni-founded Japanese AI startups. Third, Japanese AI policy contribution through the AI Strategy Council and other government advisory roles.

The competitive premise reflects Matsuo Lab's distinct positioning as the principal Japanese academic AI research lab with commercial-translation depth: faculty independence, Japanese government research-funding access, alumni-founded startup ecosystem, and Japanese AI policy influence.

Distribution channels include open-research publication through major academic venues, open-source code releases, the alumni-founded Japanese AI startup ecosystem, and Japanese AI policy contribution.

Models and products

• Published research output. Across deep learning, reinforcement learning, world models, robotics, and other application areas.
• Alumni-founded Japanese AI startups. PKSHA Technology, Naco-do, ELYZA, DeepX, Acculus, and other peers. Commercial-translation activity from Matsuo Lab research.
• Japanese AI Strategy Council. Yutaka Matsuo's chairmanship (since 2023). Principal Japanese government AI policy advisory body.
• Japanese AI policy contribution. Contribution through METI, MIC, the AI Strategy Council, and other Japanese government advisory roles.

Distribution channels include open-research publication, open-source code releases, alumni-founded startups, and Japanese AI policy contribution.

Benchmarks and standing

Matsuo Lab's evaluation framework is academic-research output (publication count, citation impact, faculty-led research-program quality), commercial-translation activity (alumni-founded startup count and commercial success), and Japanese AI policy influence rather than horizontal foundation-model leaderboards. Matsuo Lab has been consistently characterized in Japanese academic AI industry coverage as the principal Japanese academic AI research lab globally.

The alumni-founded Japanese AI startup ecosystem, including PKSHA Technology (publicly traded on the Tokyo Stock Exchange), ELYZA (substantially backed by KDDI and other Japanese strategic investors), and other peers, has anchored Matsuo Lab's commercial-translation credibility through 2010 to 2026.

The Japanese AI policy influence through Yutaka Matsuo's AI Strategy Council chairmanship and other government advisory roles has continued to anchor Japanese AI policy direction through 2023 to 2026.

Leadership

As of April 2026, Matsuo Lab's senior leadership includes:

• Yutaka Matsuo, Professor at the University of Tokyo Graduate School of Engineering. Chair of the Japanese government's AI Strategy Council since 2023.
• Shixiang Iwasawa, Associate Professor (since 2024).
• Senior research staff and graduate students across the lab's principal research areas.

Continued senior faculty recruitment and the 2024 transition to the Matsuo Iwasawa Lab structure have supported the lab's continued research output through 2002 to 2026.

Funding and backers

Matsuo Lab operates under University of Tokyo academic funding plus Japanese government research grants (JST CREST, JSPS Kakenhi, METI projects), Japanese industrial-cooperation funding from Japanese industrial partners, and selected international research grants. Specific lab-internal budget allocations are not separately disclosed.

Japanese government research-funding stability and University of Tokyo academic-research-funding access provide Matsuo Lab with financial-runway certainty. Open questions on near-term funding are limited compared to private labs, given the academic-research-funding base.

Industry position

Matsuo Lab occupies a distinctive position as the principal Japanese academic AI research lab, with faculty under Yutaka Matsuo's leadership, the alumni-founded Japanese AI startup ecosystem, the Japanese AI policy influence through the AI Strategy Council, and published research output. Industry coverage has consistently characterized Matsuo Lab as the principal Japanese academic AI research lab globally.

The 2023 to 2026 Japanese AI policy direction under Yutaka Matsuo's AI Strategy Council chairmanship has continued to anchor Japanese AI national strategy through Cabinet Office and METI direction.

Competitive landscape

• Sakana AI, Preferred Networks. Japanese AI commercial peers with overlap on Japanese AI ecosystem direction.
• Stanford AI Lab (SAIL), MIT CSAIL, CMU SCS, Berkeley BAIR, Princeton Language and Intelligence. US academic AI research peers.
• Tsinghua IIIS, Tsinghua KEG, BAAI, Shanghai AI Laboratory. Chinese academic AI research peers.
• ETH AI Center, EPFL, Tübingen AI Center, Max Planck Institute for Intelligent Systems, Mila. Non-Japanese non-US academic AI research peers.
• Alumni-founded Japanese AI startups. PKSHA Technology, Naco-do, ELYZA, DeepX, Acculus, and other peers.
• NTT, Fujitsu corporate AI research divisions. Japanese corporate AI research peers.

Outlook

• Continued faculty research output through 2026 to 2027.
• Continued commercial-translation activity through alumni-founded Japanese AI startups.
• The continued AI Strategy Council activity and Japanese AI policy direction through 2026 to 2027.
• Continued senior faculty recruitment under the Matsuo Iwasawa Lab structure.
• The Japanese government research-funding trajectory through 2026 and 2027 budget cycles.

Sources

• Matsuo Lab official site. Lab reference.
• Yutaka Matsuo Wikipedia. Founding professor reference.
• University of Tokyo. Parent university reference.
• Japanese Government AI Strategy Council. Japanese government AI policy advisory body.
• PKSHA Technology. Matsuo Lab alumni-founded Japanese AI startup.