AI can now run on devices. No cloud, no delays, less energy and lower cost. See how this change could affect everyday technology.
Artificial Intelligence (AI) is no longer just a buzz word. The demand for AI is rising with each passing year. Since the demand for AI is rising, so is the demand for cloud data centers. Companies running large language models (LLMs) face high inference costs, slower responses, energy-hungry operations, and security risks. By 2035, global energy use for AI could exceed 175GW, making cloud-only AI increasingly expensive and inefficient.
The KL1140 chip from Kneron claims to tackle these challenges by moving powerful AI processing from the cloud to portable edge devices. It allows full LLMs to run locally, reducing energy use by up to three times and hardware costs by 10x. Multiple chips can match GPU-level performance for models with up to 120 billion parameters, while consuming far less power. Independent testing has confirmed its efficiency.
Designed for real-time AI tasks, the chip supports natural language processing, voice interfaces, intelligent vision, and robotics directly on devices. This eliminates cloud delays and keeps sensitive data secure, letting developers and enterprises build AI applications that work offline.
Practical uses of the chip include security robots that understand commands and work without WiFi. Cars can handle voice commands and AI tasks entirely on-board, even without cell service. Enterprise AI assistants can run on small local servers, keeping data private. Manufacturing machines can watch video, process voice, and make decisions right on the factory floor. By enabling advanced AI at the edge, the KL1140 opens the door to portable, secure, and energy-efficient AI applications that were previously impractical.
“The twin threat of high costs and vast energy consumption means the status quo of AI computing is fundamentally unsustainable,” said Albert Liu, Founder and CEO of Kneron. “The KL1140 is our response to the challenges of scaling LLMs in the cloud alone. By running advanced models at the edge, we’re achieving a technical milestone that opens up entirely new applications for everyday devices, putting the power of LLMs directly into the hands of users.”
