Artificial intelligence chip companies are transforming from component suppliers into platform builders, as demand shifts from standalone hardware to integrated computing systems, a senior executive at Chinese semiconductor firm Black Sesame Technologies said on Thursday.

Speaking at the “Talk to the World” forum hosted by TMTPost on the sidelines of CES2026 in Las Vegas, Black Sesame’s chief marketing officer Yang Yuxin said that while traditional chipmakers could succeed by delivering chips and documentation, AI chip companies must now provide a full-stack technical foundation.

“AI chip companies can no longer just sell chips and manuals,” Yang said. “They need to build the entire technology base — including the chip, the foundational software, the toolchain and even model research teams.”

Yang said the defining feature of AI-native hardware is its computing capability rather than its connectivity, adding that future devices must be able to solve problems autonomously rather than merely transmit data to the cloud.

“The core of AI-native hardware is computation, not connection,” he said. “Devices must be capable of thinking and acting on their own.”

Yang also outlined his view of how demand for AI computing power will evolve across the electronics industry.

“In the future, every electronic device will require computing power,” he said. “Shipment volumes will form a pyramid: the smaller the computing power, the larger the shipment volume.”

His comments reflect a broader shift in the AI hardware industry, where companies are increasingly competing on ecosystems, developer support and software integration, rather than on raw chip performance alone.

As AI models become more embedded in consumer electronics, industrial equipment and vehicles, chipmakers are under growing pressure to offer end-to-end solutions that reduce deployment complexity and speed up adoption for customers.