Silicon Valley Space Week (SVSW) opened in Mountain View on October 28 with the ‘hottest’ of hot topics under examination for its Satellite Innovation segment. Those topics covered the usual space, rocketry and launch suspects, each brought bang up to date by highly skilled panellists but topics such as AI, D2D and LEO-based telecom featured throughout the day. Most sessions touched on how automation and AI would impact the space economy. Chris Forrester reports.
The opening panel addressed Next-Generation Satellite System for 2026 and Beyond and comprised Dr Tom Cwik (Chief Technologist at NASA’s Jet Propulsion Lab), Simone D’Amico (Chief Science Officer and Associate Professor of Aeronautics and Astronautics (at Stanford), Dr Ward Hanson (Lecturer in Economics at Stanford), Sita Sonty (Partner & MD at Alix Partners), Brooke Stokes (Partner at McKinsey & Co) and moderated by Janna Lewis (Senior VP of Policy & General Counsel, Astroscale US).
The panel’s key topics included – of course – AI and Quantum as well as the space economy generally and the technologies that were moving the investment needle in the space economy. ‘Space-based telecom’ saw the audience select as a major topic for examination.
McKinsey’s Brooke Stokes said that civilian-based funding remained a challenge, and some areas – not least Earth Observation – had not performed as well as some had anticipated. Nevertheless, she saw AI helping many aspects of the industry overall.
Dr Hanson asked his fellow-panellists whether there would ever be a competitor to SpaceX, and stressed that he had never seen such a variety of opinions on AI, ranging from a ‘Dot-Com-type bubble burst’ expectation and ranging up to concepts that were AI-based but had yet to be invented.
Sita Sonty said 2026 would see some differences, and where Launch was increasingly commoditised, but innovation would come from integration and would draw investment into the space sector and with value creation coming in 2026. “Cost reduction would draw investment, and AI could drive that,” she added.
Simone D’Amico had no doubt that AI would be a major catalyst and embracing scalability and intelligent infrastructure and these would attract capital inputs. Remember, today’s satellite model where a satellite had a few Megabits of data on board but this was no longer practical, and where the complexity of mega-constellations demanded satellites to be able to reason, and to collaborate with other satellites, and AI and where a single satellite could manage multiple tasks and missions. “We are in the transition phase and moving from static to intelligent platforms,” he advised.
JPL’s Dr Tom Cwik told delegates that JPL was continually looking at value creation with an emphasis on speed and better formulation, and where an AI assistant helped improve systems and achieve speedier solutions. However, he stressed that these developments might still be a year or two away before AI could be incorporated into constellations.
The majority of SVSW’s delegates – in an audience poll – saw AI being a “fundamental game changer” (although some 13% saw AI as “over-hyped”) although the industry adoption of AI would benefit as “a significant efficiency enhancer for existing applications” said others.
Quantum Communications also featured highly in the opening panel’s discussion. For example, Simone D’Amico told delegates that there was already real-world usage of Quantum in space. However, AI and Quantum when added together presented a huge challenge and the industry needed computer scientists to handle the challenges that such a combination would create. “We need a new model, so that we do not work in silos.”
Dr Cwik said he expected Quantum networking to make considerable progress, although in his view it would not see Quantum computers being installed on satellites, but would be ground and operator based.
The overall conclusion was that AI was going to make a difference although the new technology was all very well in terms of space logistics but it still depended on human creativity – at the moment – but the day when AI would control perhaps thousands of satellites was definitely coming.
AI was also extensively discussed in SVSW’s session: Building the modern space stack: Software-Driven systems and AI, and where Drew Svor (Partner, Sheppard Mullin Richter and Hampton) moderated a panel comprising Alan Campbell (Principal Space Specialist Solutions Architect at Amazon Web Services), Dr Lucy Hoag (Founder/CEO at Violet Labs), Dr Owen Brown (VP/Solutions Development at Scientific Systems), Guy de Carufel (CEO at Cognitive Space) and Lauren Perry (AI Principal Engineer/Scientist at The Aerospace Corporation).
Lauren Perry summed up the dilemma facing many of her clients and the space sector in general and who wanted AI in their business but perhaps didn’t know where to start! She said that she was seeing greater progress in AI and the tools available to clients. “We are moving in a positive direction.”
The panelists looked at the software stacks needed, whether close to the task or perhaps at the ‘back seat’ and where the focus was perhaps on observation and whether it was on board the craft or at the edge. Automation was critical in particular for anomaly detection, machine learning and the degree of human involvement.
Alan Campbell gave a simple explanation where the question was ‘how many ships were in port at Log Beach this week compared with last? He said the first challenge was for AI to identify where Long Beach was, and what a ship was and the hundred and one other identifiers which first had to be established. But AI was getting there, especially in changes and context within an environment. “If you can do the job with cost-efficient basic machine learning and predictable, and were repeatable, then spot on”.
Dr Owen Brown explained that speed, scalability and where large numbers of satellites were involved and where ‘baby steps’ might first be needed before large-scale adoption was needed. He added that AI in a multi-domain environment, perhaps in maritime, and certainly in military and which demanded speed and scale [was cool] but for commercial applications and those use cases were more difficult to identify.
Nevertheless, Dr Brown told delegates that AI and its impact was nothing less than a ‘Supersonic Tsunami’ and in years to come AI would auto-generate changes within the software stack. AI would eliminate the need for today’s mental models.
Dr Lucy Hoag added that space was increasingly democratised but software tools and the stack in general were needed to achieve interoperability and the interstitial glue was needed as a proxy for agnosticism. AI will fundamentally change the stack, but Government is far too slow in looking at these problems, and our enemies will do the work needed.
Cambell praised satellite operator SES and what he described as its transformational workload plans to modify its Sat-Com responsibilities and to flexibly adjust to missions including moving workloads to different ground stations.
However, Lauren Perry said she had visited many satellite operator’s control rooms and many of them had not changed in 20 years! She saw AI and Machine Learning making significant inroads in these areas over the next 5 years. She added that the push towards AI adoption and deployment had also come from the commercial sector and that Government needed to do more.
Guy de Carufel argued that there would be a need to be careful about putting everything in one giant stack and black box. There will need to be guard rails which understand the different functions for these pockets of data.
The overall message from SVSW was that AI was very definitely present, and available, and its sophistication would be of a huge benefit to most segments of the space and satellite industry and especially in terms of mass examination and comparison of data.
