Led · Seed · Jul 2026

Velpha provides an AI investment agent platform that helps individual investors discover stocks, balance portfolios, and monitor markets through a conversational interface. The company aims to make investing more accessible and engaging and focuses on improving AI-driven investment decision support. Founded in 2023 by Shusuke Takamizawa and based in Chuo-ku, Tokyo, Velpha recently raised approximately ¥60 million in a seed round to bolster its financial position. The raised funds are earmarked to strengthen securities account linkage security, integrate external data into product infrastructure, obtain necessary regulatory licenses, and launch new features. Velpha also plans to use capital to improve AI agent performance and to accelerate hiring for engineering and back-office functions.

Led · Seed · Jun 2026

dodoAI's flagship product is described as a “Sovereign Agentic OS,” a governance layer for AI agents that monitors, verifies, and audits AI-generated outputs and enforces human approval before final actions. The platform is model-agnostic, supporting providers such as OpenAI, Anthropic’s Claude, and Google Gemini, and can be deployed in public cloud, on-premises, or air-gapped environments. The company targets regulated industries including financial services, manufacturing, telecommunications, and energy. A disclosed pilot with a large property and casualty insurer applied dodoAI to one subsystem within a core platform of roughly 300 subsystems and reportedly reduced labor requirements by between 35% and 98% in targeted areas while maintaining quality standards. dodoAI was founded in May 2022 (originally as 58 Inc.), renamed on June 1, 2026, and has 34 employees across Tokyo, Ho Chi Minh City, and Hanoi.

Participated · Seed · Apr 2026

Almure builds a security infrastructure that lets enterprise AI Agents process confidential data inside hardware-isolated, encrypted execution environments using confidential computing and Trusted Execution Environments. Its architecture leverages technologies such as Confidential VMs and hardware enclaves and follows a purpose-built, top-down design intended to balance key-management level security with flexible AI processing. Target industries include finance, manufacturing, healthcare, and defense, where strict security and compliance requirements constrain AI adoption. The company aims to reduce the compliance-related cost barriers for enterprise AI Agent deployment and accelerate market adoption. To support these goals it is investing the newly raised seed capital into further R&D, product development, and hiring core team members.

Participated · Series A · Apr 2026

Ricerca develops RECERQA, described as an Agentic ERP that optimizes order placement and processing through autonomous data normalization and AI-driven decision support. The system uses a four-tier architecture called "Quattro" to normalize operational data and leverage accumulated data to support business decision-making and execution. Ricerca says that as more operational data is collected post-implementation, the AI's decision accuracy improves and the need for human verification declines. The company recently revamped its system to emphasize Agentic ERP characteristics distinct from conventional ERPs. Ricerca plans to expand its commercial rollout to major manufacturing and trading companies while scaling its team, prioritizing hires of AI engineers and forward-deployed engineers (FDEs). Financially, the company has raised ¥1.7 billion in Series A funding, with total financing exceeding ¥2 billion to date.

Led · Seed · Mar 2026

Boost Health develops and operates BOOST, a talent-success offering that pairs an AI tool—built on psychological science such as cognitive behavioral therapy—with monthly interventions from specialized coaches. The platform delivers continuous, individualized support intended to secure reproducible performance improvements and deeper behavioral change beyond traditional training and surveys. The company reports measurable impacts in client pilots: average productivity up 8 points after three months, stress improvements for 71% of high-stress users, a 13% engagement increase in an example case, and a 47% condition-score improvement for transferred employees at Dentsu. Boost Health is reinforcing its proprietary data foundation by integrating behavioral logs, stress metrics, and productivity indicators to quantify the ROI of human-capital investments. Planned uses of capital include advancing the AI product, strengthening the data platform, and scaling enterprise implementation and support. Founded in May 2022 and headquartered in Tokyo, the company is pursuing broader enterprise adoption, including a planned large-scale test at Hitachi in 2026.