Summarize with AI
Let's talk about a few of the hottest startup rounds from this week...
Trajectory
Trajectory raised a $45 million Series A at a $300 million valuation to automate continuous model training for nontechnical experts. Today, if a law firm or hospital wants an AI model to learn from its daily mistakes, they have to pay millions to hire armies of specialized machine learning engineers. Trajectory gives domain experts a clean interface that watches their daily feedback and automatically retrains open models in the background. With an 18-person team, they are already powering post-training for heavyweights like Harvey, Decagon, and Nvidia.
42AI
42AI raised a $62.5 million seed round at an undisclosed valuation to predict human decisions before people make them. Instead of looking at past transaction histories, their behavioral world model analyzes voice sentiment, response lag, and micro-reactions during customer service calls to calculate psychological traits. By turning those human tendencies into mathematical probabilities, they tell Wall Street brokerages exactly how individual traders will react when the market drops. Their model was trained on behavioral and biological datasets covering 30 million people and has already driven $30 million in dynamic hedging value for a single trading firm.
Tutor Intelligence
Tutor Intelligence raised a $75 million Series B at an undisclosed valuation to put adaptable robotic workers into everyday warehouses. Most warehouse robots freeze the second a package is tilted an inch off center, requiring expensive engineers on site to reset the line. Tutor solves this by pairing physical picking robots with remote human tutors who instantly steer the robot past tricky snags over the internet, feeding every recovery back into a fleet-wide training flywheel. Best of all, warehouses do not have to buy six-figure hardware upfront, because Tutor bills them strictly by the hour or per pick like temporary human labor.
Larridin
Larridin raised a $30 million Series A at an undisclosed valuation to measure the real financial return on enterprise AI. Companies everywhere are burning hundreds of millions on AI tokens, cloud compute, and software licenses without having any clue whether those tools are actually boosting employee output or just wasting money. Larridin connects directly to enterprise software and cloud bills to show executives exactly which AI tools and automated agents are delivering genuine productivity gains versus pure hype. Backed by both Andreessen Horowitz and Google's Gradient Ventures, they have already raised $47 million to become the standard auditor for the agentic economy.
The big takeaway across all four rounds is that investors are no longer funding flashy conversational wrappers. Capital is aggressively pivoting toward the unglamorous infrastructure that makes AI practical in the real world, whether that means physical robots in warehouses, continual model retraining, or measuring cold hard ROI.
Startup Breakdown
| Company | Round & Valuation | Core Problem / Solution | The Wild Stat |
|---|---|---|---|
| Trajectory | $45M Series A ($300M valuation) | Automates continual model learning so nontechnical experts can train domain AI without hiring forward-deployed engineers | Backed by Nvidia and Sequoia at an 18-person headcount, already post-training models for Harvey and Decagon |
| 42AI | $62.5M Seed round | Behavioral world model mapping psychological traits to predict trader decisions and human actions probabilistically | Trained on datasets spanning 30 million people, delivering an 85% F1 score and $30M in value for a single brokerage |
| Tutor Intelligence | $75M Series B ($117M total raised) | Deploys warehouse picking robots supported by remote human tele-operators to eliminate edge-case failures and continuously retrain fleets | Billed per pick or hourly like manual labor with zero upfront CapEx, backed by a16z and USV |
| Larridin | $30M Series A ($47M total raised) | Enterprise intelligence platform auditing AI spending and model adoption against measurable worker and agent productivity | Audits multi-cloud AI costs and workforce output, backed simultaneously by a16z and Google's Gradient Ventures |
Sources
- Trajectory: TechCrunch: Trajectory raises $45M · Trajectory Official Website
- 42AI: TechCrunch: 42AI comes out of stealth with $62.5M · 42AI Official Website
- Tutor Intelligence: The Robot Report: Tutor Intelligence Secures $75M for Warehouse Robotics · Business Wire: Tutor Intelligence Raises $75M Series B · Tutor Intelligence Official Website
- Larridin: Business Wire: Larridin Raises $30M Series A to Measure Enterprise AI Value · Larridin Official Website
Did this land? Leave a reaction.
Discussion
0 commentsMembers can join the discussion
Become a member to comment and reply to other members.
Become a MemberNo comments yet
Be the first to start the discussion.