Elorian released its first model, Glimpse, as a research preview with API access and a waitlist-only chat interface. Glimpse uses a visual scratch space to crop.
Why it matters: Elorian's Glimpse shows a small team can train a visual reasoning model on a few hundred GPUs, offering a cheaper alternative for video-heavy tasks.
Why it matters: The round shows how a saliva-based hormone reader is being funded toward early-access devices in 2027, with the same sensing technology already tested on other markers.
Why it matters: Goodweek's seed round shows how SMBs are adopting AI without IT oversight, and how one startup plans to centralize model access and spending.
Why it matters: The round shows how AI-driven protein design is moving from lab research into agricultural biotech, with Y Combinator and Paul Graham backing a two-person Toronto team.
Why it matters: Evos's pre-seed shows outcome-priced AI operators landing inside existing TMS, WMS and ERP systems at small logistics firms, with early accuracy and ARR figures as evidence.
Why it matters: Solvera's outcome-priced engineering model, backed by Spacemaker founders, offers a concrete test of whether AI agents can replace hourly consulting in MEP design.
Swedish startup 120ft Factory AB launched with €576,000 ($644,000) from 22 angel investors to build autonomous microfactories run by humanoid robots. Customers pay per finished product.
Leela raised a $4.6M seed round co-led by Array Ventures and Female Founders Fund to build AI agents that execute compliance work for regulated financial businesses. Its RegLang language links each requirement to the underlying tasks and evidence, and agents already handle fraud and AML reviews, customer complaints, and draft suspicious activity reports for human review. One early customer's 40 policies became 262 actionable obligations.
Why it matters: Leela's seed round and RegLang approach show how AI agents are moving from tracking compliance work to executing it in regulated finance.
Why it matters: Forgepoint's seed round shows how AI quoting agents are entering factory sales desks, with Frank already at seven-figure ARR and rivals raising nearby.
London-based offensive security startup CodeWall launched with $4 million in pre-seed funding led by Osney Capital for autonomous AI agents that attack enterprises the way an outside hacker would. In an earlier engagement, an agent with no credentials gained read and write access to the production database behind McKinsey's Lilli AI platform within two hours. CodeWall pitches the ongoing testing as a replacement for the once- or twice-a-year penetration test.
Why it matters: Uvi's seed round and benchmark numbers show how much enterprises pay to rebuild agent context, and how a shared context layer changes that cost.
Why it matters: Forerunner's backing of a founder who ran a $2B seller network shows how commerce investors are betting on AI that decides what to promote, not just drafts the copy.
Miracly, an AI marketing platform from The Wild Ventures, launched with a €1.8 million pre-seed round led by Beenext Ventures and Lotus. It combines generative engine optimisation.
Why it matters: Stride's seed round shows how consumer preventive health is being built around lab diagnostics and an AI copilot, with retention as the open question.
Why it matters: Ambio's seed round shows how AI matchmaking is being applied to the demand side of recycled plastics, where weak buyer interest has stalled the market.
Why it matters: Causa Prima's two-sided agent network for invoice disputes shows how AI finance startups are moving from single-company tools to cross-company settlement.