Participated · Series A · Dec 2025

Saporo offers an identity security platform that lets enterprises see their hybrid identity fabric the way attackers do. Leveraging graph databases and graph theory, the software analyzes millions to billions of potential attack paths across on-prem Active Directory, Entra ID, AWS, Okta and other directories, then surfaces the smallest set of changes that delivers the biggest risk reduction. Continuous monitoring enforces segregation of duties and alignment with frameworks such as ANSSI and MITRE so environments don’t drift back into unsafe states. Large customers reportedly eliminate about 80 % of risky paths within the first year, and the startup posts customer retention above 140 %. Adoption is strong in regulated sectors like financial services, healthcare, government, manufacturing and technology. With new capital, Saporo will expand R&D, sales and marketing through 2026, strengthen one-click remediation, and add coverage for Google Workspace/Cloud and GitHub. The company also intends to deepen its footprint in France, Switzerland, Benelux, Germany and Italy while selectively entering the U.S. market.

Participated · Series A · Nov 2025

Cerrion provides manufacturers with AI-powered video agents that continuously analyze standard factory camera feeds to detect process deviations, quality issues, and safety risks in real time. When anomalies are spotted, the system can trigger automated interventions—such as slowing or stopping machines—and immediately notify relevant personnel, enabling problems to be resolved up to 50 percent faster while halving downtime and scrap losses. The platform is already deployed in live production at global manufacturers including Unilever, Riedel, Schott Zwiesel, Stölzle Lausitz, Sisecam, and Verallia, which serve major brands like Pepsi, Coca-Cola, Pfizer, and Novartis. Customers typically scale from single-site pilots to company-wide rollouts within months, highlighting rapid adoption. The company’s team draws talent from ETH Zurich, Google, and EPFL. Following its $18 million Series A raise, Cerrion plans to double headcount in Europe and the U.S. and broaden its platform beyond vision analytics to support additional manufacturing processes. No specific revenue or user figures were disclosed in the article.

Participated · Series B · Sep 2025

Founded in 2020 and headquartered in Schwyz, Switzerland, deskbird offers an all-in-one workplace management platform available via mobile app, web, Microsoft Teams, and Slack. The software lets employees book desks, meeting rooms, parking spots, and guest access and plan their weekly schedules, while providing employers analytics to right-size real estate, improve collaboration, and cut waste. deskbird is building AI-powered workplace intelligence that will deliver predictive space-usage insights and help create adaptive offices. The company already serves more than 10,000 offices and 250,000 users worldwide, with enterprise customers such as Samsung, Deloitte, Philips, and Airbus. Revenue figures were not disclosed, but the customer base and user metrics point to meaningful market traction. Fresh Series B capital will fund continued product innovation, deeper integrations, and geographic expansion across the UK, broader Europe, and additional international markets. The firm positions itself at the intersection of hybrid work enablement and real-estate efficiency, two sizable and growing themes.

Participated · Seed · Feb 2024

Jua offers the Earth Intelligence Platform, an AI decision engine powered by EPT-2.0, a proprietary large physics model trained on petabytes of multi-domain data. The platform simulates and predicts physical-world behavior to provide real-time forecasts of weather, energy supply-demand fluctuations, and grid-level dynamics. It is already in use by leading energy trading firms to simulate imbalances, anticipate pricing movements, and automate decision workflows. Jua plans to scale the commercial rollout across the energy sector, expand into adjacent markets such as agriculture and logistics, and double down on R&D. Financially, the company raised an $11M Series A, bringing total funding to $27M after a $16M seed round in 2024. Marvin Gabler is co-founder and was named the company’s new CEO. Jua is developing a foundational AI "physics" model that learns the natural world by ingesting diverse data sources to model and predict weather and climate. Its first application targets modeling and predicting weather and climate patterns for the energy industry, with that product due to launch in the coming weeks. The company plans to expand the model's use to agriculture, insurance, transportation and government. Jua says it trains on around 5 petabytes of data and brings together weather stations, satellite imagery, topography and other noisy, recent data into a single end-to-end system. Founders Andreas Brenner (CEO) and CTO Marvin Gabler bring weather research, deep learning and energy-sector experience to the project. The company claims its system uses 10,000 times less compute than legacy systems; Jua is still early-stage and has not yet launched its first products. Jua is an artificial intelligence climate startup developing technology to produce weather predictions for small, precise areas over short time windows. Its platform enables startups, companies, government institutions and researchers to create purpose-built weather models in a matter of hours. The company is currently offering the platform to a select number of customers on request and plans a wider release in early 2023. Jua intends to use its recent funding to expand its business reach. The company is led by co-founders Andreas Brenner (CEO) and Marvin Gabler (CTO). Jua maintains hubs in Zurich, Berlin and Cape Town.

Participated · Equity · Dec 2023

SuperDuperDB is an open-source Python package that lets developers integrate AI models, AI APIs, and vector search directly with existing databases to build and deploy AI applications. The framework (v0.1) supports streaming inference and scalable model training, can use in-database vector functionality or its own vector-index implementation, and is deployable standalone or at scale via Kubernetes. It supports a broad ecosystem of data stores (including MongoDB, PostgreSQL, MySQL, Snowflake, BigQuery, ClickHouse, DuckDB, and others) and AI models/APIs from PyTorch, Sklearn, Hugging Face, OpenAI, Anthropic, and Cohere. The company is San Francisco–based and an Intel Ignite portfolio company, and it emphasizes a Python-first approach to avoid forcing developers into SQL dialects. SuperDuperDB has received $1.75 million in early funding from Hetz.vc, Session.vc and MongoDB's venture arm, and released its framework on Product Hunt. The team is expanding partnerships, has run webinars and live coding with MongoDB and Cisco, is evaluating POCs with Intel and other SMEs, and plans a native app for Snowflake's marketplace.