MIT Mobility Initiative

From DARPA to Robotaxis to Warehouses — The Full Arc of Autonomy

The MIT Mobility Forum featured Karl Iagnemma, a rare figure who has traversed the entire autonomy spectrum—from MIT research labs and DARPA challenges to founding Nutonomy, steering the Aptiv‑Hyundai joint venture Motional, and now leading Vecna Robotics. Iagnemma recounted the evolution of autonomous‑vehicle technology, describing the early “bad old days” of proof‑of‑concept demos, the pivotal impact of modern machine‑learning advances—especially convolutional networks and transformers—and the subsequent shift from pure technical feasibility to questions of business models and capital requirements. He emphasized that perception, once deemed the hardest problem, was dramatically accelerated by breakthroughs in computer vision that originated outside the AV community. With perception now reliable, the industry’s primary challenge is deploying a proven stack across diverse markets at minimal marginal cost. Iagnemma argued that technical architecture debates—such as end‑to‑end versus deterministic layers—are secondary to strategic considerations like partnership structures, regulatory approaches, and sustainable financing. Illustrating these points, Iagnemma described Motional’s joint venture with Hyundai as a response to the billions‑dollar capital needed for large‑scale deployment, and highlighted his current work at Vecna Robotics, where autonomous forklift technology mirrors passenger‑vehicle autonomy but offers a potentially more tractable business case. He also noted that social‑intelligence issues, such as interacting with first responders, can be addressed through machine‑learning models that recognize and respond to emergency cues. The discussion concluded that success in autonomy will be defined not by a single technical breakthrough but by the ability to integrate robust AI, navigate regulatory landscapes, secure deep‑pocketed partnerships, and build economically viable operations across multiple sectors.