Uploaded June 2026 | Updated September 2026, 1 week ago
It’s not about maximizing compute; it’s about optimizing it.
At Computex 2026 in Taipei, NXP Semiconductors made the case for "right-sized AI," with Ajith Mekkoth, executive VP of AI and chip engineering, arguing that AI deployment should be driven by application requirements rather than the pursuit of ever-larger models and greater compute power.
“What I meant by right-sized AI is that different types of models have their place,” Mekkoth said, emphasizing that many applications can be served effectively by smaller AI models rather than generative AI models. It's a fit-for-purpose approach centered on selecting “the right model that's applicable for the application at hand.”
While cloud-based AI remains relevant for some workloads, Mekkoth noted that “the token costs are too expensive, the latency is too big” for many use cases. Advances in silicon technology now make it possible to perform significant compute locally, reducing dependence on cloud infrastructure.
NXP believes the transition toward edge AI will accelerate, particularly in consumer applications. However, Mekkoth said adoption challenges are often linked less to AI itself than to overall system design. “We believe it's not just an AI problem; it's a problem of system design.”
To illustrate NXP's strategy, Mekkoth highlighted industrial motor control, voice-enabled human-machine interaction, and robotic control as examples where AI can operate efficiently within tight power and latency budgets. He also said the acquisition of Kinara has strengthened NXP's portfolio by adding generative AI capabilities and discrete NPUs that can be combined with the company's processors to scale edge AI according to customer requirements.
Looking ahead, Mekkoth expects concerns around data ownership, privacy, and security to further strengthen the case for edge AI, as organizations seek greater control over their data and real-time decision-making capabilities.
#NXP
#RightSizedAI
#PurposeBuilt
#HMI
#EfficientEdgeDeployment
#industrialmotorcontrol
#voiceenabledhumanmachineinteraction
#roboticcontrol
It’s not about maximizing compute; it’s about optimizing it.
At Computex 2026 in Taipei, NXP Semiconductors made the case for "right-sized AI," with Ajith Mekkoth, executive VP of AI and chip engineering, arguing that AI deployment should be driven by application requirements rather than the pursuit of ever-larger models and greater compute power.
“What I meant by right-sized AI is that different types of models have their place,” Mekkoth said, emphasizing that many applications can be served effectively by smaller AI models rather than generative AI models. It's a fit-for-purpose approach centered on selecting “the right model that's applicable for the application at hand.”
While cloud-based AI remains relevant for some workloads, Mekkoth noted that “the token costs are too expensive, the latency is too big” for many use cases. Advances in silicon technology now make it possible to perform significant compute locally, reducing dependence on cloud infrastructure.
NXP believes the transition toward edge AI will accelerate, particularly in consumer applications. However, Mekkoth said adoption challenges are often linked less to AI itself than to overall system design. “We believe it's not just an AI problem; it's a problem of system design.”
To illustrate NXP's strategy, Mekkoth highlighted industrial motor control, voice-enabled human-machine interaction, and robotic control as examples where AI can operate efficiently within tight power and latency budgets. He also said the acquisition of Kinara has strengthened NXP's portfolio by adding generative AI capabilities and discrete NPUs that can be combined with the company's processors to scale edge AI according to customer requirements.
Looking ahead, Mekkoth expects concerns around data ownership, privacy, and security to further strengthen the case for edge AI, as organizations seek greater control over their data and real-time decision-making capabilities.
#NXP
#RightSizedAI
#PurposeBuilt
#HMI
#EfficientEdgeDeployment
#industrialmotorcontrol
#voiceenabledhumanmachineinteraction
#roboticcontrol
