HeartCore Enterprises' subsidiary HeartCore Financial signed a non-exclusive sales agency agreement on November 12 with ONE AI, a Korean physical AI developer led by former Intel Asia-Pacific executives, to bring ONE AI's Axis.ONE platform to Japanese manufacturers. Axis.ONE is pitched as a retrofit: it adds vision, perception, decision-making and continuous learning to robots that are already on factory floors, running on standard Intel-architecture industrial PCs and edge devices rather than dedicated GPU clusters or cloud servers. HeartCore will earn commissions on referred customer revenue starting at 15% in year one, stepping down to 10% and then 8% over the deal's three-year initial term.
The pitch is squarely aimed at the gap between the AI robotics industry's demos and the roughly 4.6 million industrial robots already installed worldwide. Most of those machines are precise but blind: they repeat a fixed motion and cannot adapt when a part shifts slightly or a bin runs low. Replacing them with AI-native robots is expensive and slow. Axis.ONE's bet is that operators would rather bolt intelligence onto what they already own.
Why On-Premise AI Is the Real Fight in Factory Automation
ONE AI claims Axis.ONE can run on-site, including in air-gapped networks with no cloud dependency at all, and that it cuts deployment cost by up to 90% compared with GPU-based systems. It also claims to shrink the time robots need to learn a new task from roughly 100 days down to 10, or in some configurations from 30 days to a single hour. Those are the kind of numbers that, if they hold up in production rather than pilot settings, would undercut the GPU-heavy approach favored by most physical AI startups chasing humanoid and warehouse robot contracts.
HeartCore CEO Sumitaka Kanno put the gap bluntly: industrial robots have been precise for decades but "cannot see, adapt, or learn." ONE AI CEO In Chan Baek's framing — that robots "should not have to be replaced in order to become intelligent" — is effectively a pitch against the capital expenditure cycle that companies like Figure, Agility and Apptronik are asking manufacturers to commit to.
A Different Bet Than the Humanoid Boom
While most of the robotics funding story in 2026 has gone toward new humanoid hardware, this deal bets that the faster, cheaper win is software layered onto the installed base. Japan is a logical testing ground: it is one of the world's largest industrial robot markets, with an aging workforce and manufacturers under pressure to automate without large capital outlays. HeartCore, better known as an IPO consulting firm than a robotics player, is essentially acting as a distribution arm for ONE AI's edge-AI technology rather than a competitor to robot makers like Fanuc or Yaskawa, both headquartered in Japan.
The deal's real test will be whether Axis.ONE's claimed cost and speed advantages survive contact with messy factory floors, where lighting, part variation and legacy PLCs tend to break clean lab numbers. If it works at scale, it offers manufacturers a cheaper on-ramp to physical AI than buying new robots outright — and a reason for the current wave of GPU-dependent competitors to worry about pricing.
Source: GlobeNewswire