Portfolio Category
AI portfolio companies
Updated
This collection is a public map of the companies tagged AI in Sahin Boydas’s portfolio data. It helps readers see the kinds of products represented—from models and compute to workplace software, healthcare tools, robotics and industrial systems—and find adjacent company pages. It is not investment advice, a ranking, or a statement that a company is a fit for every founder or reader.
This collection currently contains 80 public portfolio companies.
What this collection contains
The AI label is broad. Some companies are centered on models, data systems, chips or compute. Others apply AI inside a defined workflow, such as clinical chart review, legal work, construction, logistics, code creation or media production. A company can also appear in another public category when its work crosses a domain boundary. The AI collection consequently connects to software, hardware, health, climate, automation, aerospace and robotics pages.
The useful distinction is not whether a company uses AI as a label. It is where the technology sits in the product. Foundation-model and infrastructure companies focus on systems that make AI capabilities available. Developer and code tools help people create software. Enterprise products connect AI to a business workflow. Robotics and hardware companies extend it into physical settings. Health and scientific products apply it where evaluation, documentation and human responsibility carry particular weight.
The filters and tags support different research paths. Readers interested in technical infrastructure can follow AI Infrastructure, AI Chips, Compute and Photonics. Product researchers can move through AI Enterprise, AI Healthcare, AI Code, AI Agents, AI Video or AI Voice. People studying physical systems can use the related Robotics, Aerospace, Hardware and Automation pages. These labels are navigation tools, not judgments about quality or relative importance.
Representative paths through the collection
The following examples show the range of public category and subcategory labels. They are representative examples only, not a ranking.
Models, data and underlying systems
Anthropic is listed in AI with AI Security and Foundation Models subcategories. Together.ai has AI Infrastructure and Foundation Models tags. Lambda, Lightmatter, Piris Labs, Cerebras Systems and Extropic offer routes through compute, chips, photonics and infrastructure. These labels distinguish models from the systems used to develop or deploy them; they are not performance signals.
Software and workflow applications
The collection also contains products used inside a work process. Glean carries AI Enterprise and Productivity tags, while UpCodes, Ruli AI, Theo AI and CaseMark connect AI to public legal or operational descriptions. Replit, Inc., Warp, Rosebud AI, Flux.ai, Magic and Featherless AI connect AI to AI Code or Developer Tools. These routes clarify whether a product is aimed at developers, end users, a specialist team or another software system.
Health, science and physical systems
Brellium, Sully.ai, Biostate AI, Hologen and Slingshot AI connect AI with health-related tags. KoBold Metals, Terra AI and Terranox AI link it with climate or resource-focused work. Shield AI, Nomagic, Origami Robotics, Boost Robotics, Airtrek Robotics, Firestorm, Cytronic and Formulate span physical systems. Their adjacent aerospace, automation, hardware and robotics tags help readers consider operating environment as well as technical capability.
How founders and readers can use the page
Start with the company card closest to the product’s actual use case, then follow its category and subcategory links. Describe the product in plain language: who uses it, what workflow or system it touches, and what must work reliably. A more specific workflow description is often clearer than a fashionable label.
Then separate capability from deployment. NIST presents its AI Risk Management Framework as voluntary guidance for trustworthiness considerations in design, development, use and evaluation. Its materials organize work around governing, mapping, measuring and managing risk. A useful product explanation therefore covers the intended user, context of use, quality evaluation and responsibility for decisions.
Governance is also an organizational question. ISO describes an AI management system as policies, objectives and processes for responsible development, provision or use. The European Commission describes the AI Act as risk-based, with stricter obligations for certain uses and transparency duties in specific situations. These sources do not determine what applies to a particular product, but they show why deployment conditions belong beside technical capability.
Use the company descriptions, public categories, subcategories and status shown by the site for portfolio navigation. For a broader view, return to the angel investments hub, explore the public portfolio context, search the portfolio hub, or read the public portfolio hub.
Frequently asked questions
What does this AI portfolio collection include?
It brings together companies tagged AI in the public portfolio data. The collection spans foundation models, AI infrastructure, enterprise software, developer tools, healthcare, robotics, hardware, climate-related work and other applied products.
How should founders use this AI portfolio page?
Founders can use the taxonomy and company cards to understand how their product is described publicly, identify adjacent work, and prepare a concise explanation of the customer problem, deployment setting, data boundaries and evaluation approach. The page is not a promise of fit or investment advice.
Why do governance and deployment context matter for AI companies?
AI products affect people and organizations in different ways. NIST, ISO and the European Commission each describe approaches that connect AI design and use with risk management, transparency, documentation, oversight and accountability. The applicable approach depends on the product and where it is deployed.
