Recordings of the ICRA 2026 keynote talks, plenaries, panels and tutorials are now available to watch, with IEEE hosting the video. The IEEE International Conference on Robotics & Automation took place in Vienna from 1 to 5 June, and the recordings put its main-stage programme online for people who did not attend.
The conference drew researchers and industry professionals whose expertise covers many areas of robotics. The release is organised into plenary talks, keynote sessions, industry keynote sessions, awards, keynote tutorials and panels.
Three plenary talks
Three plenary talks anchor the recorded programme. Ken Goldberg spoke under the title "A Tale of Two Cultures: Can Agentic Coding Close the Gap?" The title poses a question about whether AI systems that write code can narrow a divide between two cultures. The recording is where that question gets its answer.
Barbara Mazzolai gave "Natural Revolution: Biological Principles for Frugal and Sustainable Robotics." The title points to biology as a source of design principles, with frugality and sustainability as the stated goals.
Roland Siegwart's talk was "Aerial Robots – From Omnidirectional Flight to Physical Interaction at Height." It covers a progression from flight in any direction to aerial machines that make physical contact with their surroundings at elevation.
Keynote sessions and tutorials
The keynote sessions are grouped by theme. There are six: Autonomous Vehicles & Navigation, Medical & Healthcare Robotics, Robot Perception & Spatial AI, Manipulation, Humanoids and Embodied Design, Robot Learning, Planning & Foundation Models, and Human-Robot Interaction.
Separate industry keynote sessions sit alongside them, as do recordings of the awards and the keynote tutorials. The listing does not name the speakers for these sessions or give their running times, so this article does not either.
Six panel discussions
The panels take up policy, money and research culture as much as engineering. Panel 1 looks at moving from humanoid robotics research to startup creation, and the role public funding plays in that step. Panel 2 addresses sustainability in robotics, from green design to real-world impact. Panel 3 covers building sustainable and trustworthy AI for automation.
Panel 4 is titled "Publish or Perish: Surviving the Paper Deluge – Is AI the solution? If not, what is the Solution?" Panel 5 takes on "Robots for All" in a fragmented world, with competing global visions and shared futures from Europe, Asia and the United States. Panel 6 examines impact and financial return on humanoid investment.
What the programme shows, and what it does not
The first thing the list shows is where the conference put its attention. Humanoids appear in a keynote track and in two of the six panels, one on turning research into startups with public funding and one on financial return. Neither panel title treats humanoids only as a technical problem. Both frame them as something that needs capital and has to show a return. The recordings do not say what the panelists concluded, so the titles are the only evidence for that framing.
Sustainability is the second thread. It appears in Mazzolai's plenary, in Panel 2 and in Panel 3. Three separate slots on one theme suggest it is treated as a design requirement and not a side topic. The listing gives no detail on what was proposed, such as specific methods, metrics or commitments, so how concrete the discussion was cannot be judged from the titles alone.
The third thread is the field's own workload. Panel 4 asks whether AI can help the research community cope with a flood of papers, and then asks what the solution is if it cannot. Goldberg's plenary asks a related question from another direction, about whether agentic coding can close a gap between two cultures. Both are framed as open questions. Neither title claims an answer, and anyone citing these talks should check the recordings before attributing a position to a speaker.
Panel 5 stands out because it names three regions, Europe, Asia and the United States, and describes the world as fragmented. A conference held in Vienna gave a full panel to how those regions' visions compete and overlap. That is a sign that geography and policy are now part of the technical agenda. The source does not say who sat on the panel or what it recommended.
Public recordings matter here for a practical reason. Conference papers record results. Plenaries, keynotes and panels record how senior people frame problems and priorities, which rarely appears in a paper. Putting that material online lets students, funders and smaller labs see it without paying to travel to Vienna. The release does not say whether the videos are free to watch, how long they will stay available, or whether they carry captions, so access terms remain unconfirmed.
Several things are unproven. Titles are not content, and the list does not show whether the humanoid funding panels produced evidence on returns or only opinion. It does not show whether the sustainability sessions went beyond principles. Attendance figures and the number of papers presented are not given, so the scale of this ICRA cannot be compared with earlier editions from this material. What can be said is narrower. The 2026 programme paired technical keynotes with panels on money, trust, sustainability, publishing pressure and international competition, and the full ICRA 2026 keynote talks are now there to check against that reading.