Platform Bundling Threatens Skild

The real threat is not that bigger platforms build better robot models, it is that they can make robot intelligence feel free inside deals customers are already signing. Skild sells a standalone brain layer through APIs and cloud services, but NVIDIA can fold Isaac into chips, simulation tools, and OEM partnerships, while Google linked DeepMind robotics models to Apptronik and a broad tester network. That makes distribution and pricing as important as model quality.

• Skild is horizontal software. A robot maker maps its joints, sensors, and cameras into Skild Cloud, then calls high level behaviors instead of writing control code from scratch. That only works if OEMs and enterprises choose a separate intelligence vendor, rather than taking one bundled with hardware, cloud, or developer tools.
  1 sacra 2 sacra
• NVIDIA already has the natural bundle. Isaac spans simulation, training, perception, manipulation, and deployment, and it is being adopted by robot makers and industrial partners. If a robot OEM is already buying Jetson, Omniverse, or NVIDIA training infrastructure, adding Isaac is much easier than adding a new independent stack.
  3 nvidia 4 nvidia
• Google is pushing the same pattern from the software side. DeepMind says Gemini Robotics works across multiple robot forms and is partnered with Apptronik plus more than 60 testers. Intrinsic is also now being brought closer into Google, which strengthens the case for a full stack combining foundation models, cloud, and robotics workflow software.
  5 deepmind 6 deepmind 7 intrinsic

Going forward, the winners in robotics AI will be the companies that become the default layer inside existing buying channels. For Skild, that means moving beyond a raw model sale and becoming hard to replace through vertical modules, private infrastructure, and deployment data that bundled platform offerings cannot easily copy.