Three AI Startups Secure Seed Funding to Transform Legal, Data, and

Fearn, Upriver, and Niteshift each announced seed‑stage investments this week, collectively raising $26.5 million to accelerate AI tools that automate complex professional tasks. The funding underscores growing investor confidence in niche AI applications for legal, data, and software development domains [1].

Key takeaways

• Fearn secured $5.5 million, led by Kindred Ventures, to build AI‑powered patent drafting software for inventors [1].
• Upriver raised $14 million, with Valley Capital and Hetz Ventures heading the round, to deploy autonomous agents that clean and manage enterprise data pipelines [2].
• Niteshift obtained a $7 million seed round led by Greylock’s Jerry Chen, targeting AI coding agents that let companies avoid lock‑in with major model providers [3].

Automating Patent Drafting with Fearn

Fearn’s seed round, announced in June 2026, is anchored by Kindred Ventures and includes Designer Fund, Essence Venture Capital, and a16z Speedrun, Andreessen Horowitz’s startup boot camp [1]. Co‑founder and CEO Han Kim says the platform lets any inventor upload a technical document and generate a draft patent in minutes, scoring the draft’s completeness, producing drawings, and citing sources. The company charges a flat $2,000 per draft and claims the process reduces drafting time from weeks to minutes. Kim’s background in patent prosecution at Morrison & Foerster informs the product’s focus on structured, technically‑rich documents, a niche he believes is ripe for automation [1].

Data Engineering Agents from Upriver

Upriver, an Israeli startup founded by Ido Bronstein and Omri Lifschitz, closed a $14 million seed round led by Valley Capital and Hetz Ventures, with angel investors from the cybersecurity firm Cyera also participating [2]. The founders draw on Bronstein’s experience building data‑fusion tools for an elite Israeli military intelligence unit. Upriver’s software deploys autonomous agents that monitor data flows across platforms such as Fivetran, Snowflake, Airflow, and Tableau, identifying incomplete, duplicated, or unreliable data and helping maintain pipelines without extensive manual effort. The company reports early adoption by Unity, DMGT, and Nimble, and partnerships with Databricks and Snowflake [2].

Niteshift’s AI Coding Platform

Former Datadog engineers Sajid Mehmood and Conor Branagan launched Niteshift, raising $7 million in a seed round led by Greylock’s Jerry Chen, with angels including Reid Hoffman and Datadog’s Olivier Pomel joining [3]. The startup positions its AI coding cloud as a “vendor‑agnostic” layer that routes requests between models such as OpenAI’s GPT, Anthropic’s Claude, and open‑source alternatives, allowing customers to avoid dependence on any single AI provider. Mehmood argues that this approach mitigates the risk of “SaaSpocalypse” dynamics where large AI firms could undermine downstream users by launching competing products [3].