With Computex 2026, Asia's largest IT and computing trade show being held in Taipei, Taiwan, the prospects for the Robotics market, including humanoids, are becoming more concrete. The industry sees the potential for the robot market to extend the artificial intelligence (AI) semiconductor boom centered on data centers into a second supercycle.
For Samsung Electronics, the Robotics market is seen as a chance to broaden the customer base of its foundry (contract chip manufacturing) business. Until now, results at the Samsung Electronics foundry business have swung depending on orders from a few large customers such as Nvidia, Qualcomm, and AMD. As the humanoid and industrial robot markets grow, demand could rise not only for AI chips but also for various semiconductors such as sensors, power semiconductors, and microcontroller units (MCUs).
According to the industry on the 2nd, the Samsung Electronics foundry division is widening its order strategy beyond existing mobile application processors (APs) and AI chips in step with the expansion of the Robotics market. Led by Han Jin-man, head of the Samsung Electronics foundry division (president), the unit is said to be focusing on expanding contract manufacturing opportunities for advanced sensors, power chips, and MCUs needed for Robotics.
A source familiar with the Samsung Electronics foundry division said, "From a sales and marketing perspective, the scope of activity at the foundry division has expanded two to three times compared with the past, to the point you could say it is pursuing extensive customer acquisition." Inside Samsung Electronics, the current moment—when rival TSMC's production capacity has effectively reached its limit—and the next one to two years are seen as a golden time to secure new customers.
A defining feature of the Robotics market is that the types of semiconductors needed vary greatly depending on a robot's performance and use case. Simple industrial robots are built around MCUs, sensors, and motor drivers, but the closer they get to humanoids, the more important high-performance AI chips that process camera, voice, and tactile data in real time and low-power, high-capacity memory become. That is why there is an assessment that robots are evolving from simple automation equipment into "moving Edge AI servers."
Humanoid semiconductor demand largely splits between the robot body and the training infrastructure. The robot body needs an onboard AI system-on-chip (SoC) that quickly processes multiple sensor data while reducing power consumption, along with low-power memory such as the LPDDR family. By contrast, data centers that train humanoids could see rising demand for high bandwidth memory (HBM) and high-performance solid-state drives (SSDs) to run robot foundation models, simulations, and vision, language, and action models.
An industry source said, "Humanoids could be a new opportunity for Samsung Foundry to secure large advanced customers," adding, "AI chips for humanoids, like those for Autonomous Driving, require high-performance inference, power efficiency, and long-duration operational stability, making it highly likely they will translate into demand for advanced nodes such as 4-nanometer and 2-nanometer." The source said, "Chips aimed at both self-driving cars and humanoids, like Tesla's next-generation AI6 chip, are representative cases Samsung Foundry can target."
Samsung Electronics' strength is that it can propose memory and foundry together within one company. For AI chips for humanoids, not only compute performance but also memory bandwidth, packaging, and power efficiency all grow in importance. Samsung Electronics has high bandwidth memory (HBM), DRAM, advanced packaging, and advanced foundry at the same time, allowing it to bundle and propose customized AI chips and memory for customers. However, stabilizing 2-nanometer production yield, closing the customer ecosystem gap with TSMC, and obtaining reliability certifications required for long-duration operation are cited as challenges Samsung Foundry must solve.