Beijing Tashan Technology Co., Ltd., a leading AI tactile-sensing company, has completed its shareholding restructuring and converted to a joint-stock company, formally positioning itself to become China’s first listed “tactile-intelligence” firm.
Founded in 2017 in Beijing, Tashan is an AI tactile-sensing chip and application-solution developer founded by a transnational R&D team with backgrounds at Tsinghua University and the University of Manchester. The company builds chips, sensors, algorithms and full-stack solutions serving humanoid robotics, smart automotive, home appliances and consumer electronics, and it positions tactile perception as an independent learning and control pathway rather than a mere add-on to vision.
Its technology centers on the world’s first analog-digital hybrid AI tactile chip with a spiking-neural-network (SNN) architecture, decoding multi-dimensional tactile signals — 3-axis force, material, proximity and temperature. Newer products include the Emerald E10A dynamic tactile chip (2.9 μs time resolution, 300 kHz measurement frequency), fingertip sensors with 0.01 N force resolution, full-body multimodal e-skin, a TS-ECHO data-collection glove, and a visual-tactile fusion inspection workstation.
Commercially, monthly tactile-sensor deliveries have reached tens of thousands, with H1 2026 orders exceeding four times full-year 2025. The company has partnerships with over 180 robotics-chain clients and dozens of global manufacturers, and has closed six funding rounds. Its July 2026 Series B, backed by Taiping Innovation, Joyson Electronics, AUX and others, valued the company at roughly RMB 7 billion.
In September, Tashan and 2024 Turing laureate Prof. Richard Sutton launched the “Robot Kindergarten,” a reinforcement-learning platform that lets robots accumulate tactile experience through real-world interaction.
“Tactile sensing is the only channel through which machines can truly touch and alter the physical world,” said CEO Ma Yang. “We invite more partners to join us in building the native data systems and model paradigms that unlock physical intelligence.”
