Trajectory raises $45 million to automate artificial intelligence model training

Artificial intelligence startup Trajectory raised $45 million in a new funding round, TechCrunch reported on November 2, 2026. The investment values the one-year-old company at $300 million. Trajectory aims to automate how specialized artificial intelligence models learn from feedback and continuously improve. [1]

Current artificial intelligence models built by providers like OpenAI and Anthropic do not learn from their mistakes without additional training, according to TechCrunch. Businesses adapting these general models for specialized industries like law and finance often rely on internal research teams or expensive consultants. [1]

Trajectory was founded about a year ago by three Stanford classmates who previously worked at Google DeepMind and Apple, TechCrunch reported. The company provides tools that allow nontechnical industry experts to fine-tune models directly without hiring forward-deployed engineers. [1]

The development comes as artificial intelligence companies increasingly use automated systems for coding and research tasks. Artificial intelligence systems that iteratively design more capable successors undergo recursive self-improvement, though human oversight remains necessary to prevent operational bottlenecks. [1][2]