Vikram Gupta on edge AI
From Unpacking the Strategy Behind Synaptics Astra SL2610 Launch · · Synaptics
“The edge is evolving rapidly, and we see a natural osmosis of AI models moving from the cloud to the very far edge, driven by latency, privacy, and energy consumption considerations.”
On , Vikram Gupta, Senior Vice President, GM of IoT Processors & Chief Product Officer at SYNAPTICS INC, spoke about edge AI during Unpacking the Strategy Behind Synaptics Astra SL2610 Launch on Synaptics.
About Vikram Gupta
Vikram Gupta, Senior Vice President, GM of IoT Processors and Chief Product Officer at Synaptics, has been active in discussing the company's edge AI strategy throughout 2024 and 2025. In November 2025, he announced the launch of the Astra SL2610 series of Linux processors, which he described as featuring a dual-core Cortex-A class SoC with a 3D GPU and an NPU developed in collaboration with Google. Gupta stated that the company decided to develop its own AI hardware accelerator IP rather than licensing off-the-shelf IP, and that it is incorporating Google's Coral NPU into all future Synaptics processors. He said the company plans to open-source its AI hardware and software to enable ecosystem development. At Embedded World 2025, Gupta introduced the Astra SR100 series of AI-enabled microcontrollers, describing them as a new line of MCUs designed with a "geared" approach to power consumption, with different levels of intelligence and power usage. He stated that the SR100 series consumes about 100 milliwatts in its highest performance state)Skip Gupta has also spoken about the broader industry trends driving edge AI adoption, citing latency, privacy, and energy consumption as key factors. He has emphasized the importance of security in IoT devices, noting that the SL2600 series includes PSA Level 3 security. In a panel discussion at Synaptics Tech Day 2025, Gupta said that "edge AI is the new frontier" and that the company's SL2610 family of products represents rapid progress in making AI work efficiently at the edge. He has also discussed the company's approach to making AI "frictionless" for customers by offering models, helping with model optimization, and providing a framework for customers to make their own modifications.