Overview: On January 8, according to Archer Aviation’s investor relations website and GlobalAir, Archer announced at CES 2026 that it will collaborate with NVIDIA to develop and deploy next-generation aviation AI technology based on the IGX Thor platform. According to Joby’s official website, Joby announced that it will be a launch partner for IGX Thor, used to accelerate the development of next-generation autonomous flight. At the same exhibition, the two eVTOL companies successively introduced automotive-grade AI computing power into their respective flight projects.
According to a press release published on Archer’s investor relations website, the company announced this partnership at CES 2026 in Las Vegas, with the aim of “developing and deploying next-generation aviation AI technology.” According to its statements on social media and in industry publications, the collaboration involves areas such as pilot safety and airspace integration. Joby’s statement is more direct: according to Joby’s official website, the company is launching as a partner using the IGX Thor, with the goal of “accelerating the development of next-generation autonomous flight.” Both companies are using the same computing platform, but Archer is initially focusing on pilot assistance and airspace integration, while Joby is directly targeting autonomous flight, indicating a difference in their respective timelines.
According to Autonomy Global, Archer’s Midnight is a piloted eVTOL aircraft designed to carry four passengers and one pilot, optimized for short-range, high-frequency routes between cities and between cities and airports. As of early 2026, the Midnight is in the final stages of the FAA type certification process, having passed the final airworthiness standards and moved into the compliance verification and test flight phases. This means that the AI capabilities brought by IGX Thor will be integrated into a model that is close to obtaining certification and will still be piloted by a human pilot, initially serving the pilot before potentially being used for autonomous operation.
Fully autonomous operation presents a much greater challenge. According to droneintelligence.ai, certifying a fully autonomous eVTOL aircraft is extremely difficult because the autonomous system itself is the subject of safety validation. Regulators are not only reviewing the aircraft's structure but also the entire software system that enables the aircraft to make autonomous decisions. The media outlet mentioned that relevant companies are expanding their test fleets and accumulating flight data to meet the FAA's review requirements. Moving from pilot-assisted to fully autonomous operation, while computing power is an important starting point, the progress in gathering evidence is the universally recognized key hurdle in the industry.
Aviation companies are making an appearance at CES, demonstrating that AI computing power providers are becoming a new player in the aviation supply chain. For companies like Archer and Joby, adopting mature, automotive-grade computing platforms offers the benefits of a well-established ecosystem and rapid iteration. However, the trade-off is the need to prove to regulators that this system is equally reliable under aviation-level safety standards. In the short term, the race for autonomous flight hinges on securing computing power; in the long term, it depends on the progress of certification. Beyond the applause at CES, the FAA's review process is the real proving ground.