Is neuromorphic computing finally ready to challenge the GPU's dominance in AI inference? Many startups are bringing these innovations to market. One is BrainChip, whose latest neuromorphic chip design promises milliwatt-level AI inference for edge applications. This is in stark contrast to the...
www.linkedin.com
My passion is turning AI research into practical, commercial solutions with global reach.… · Experience: Analog Devices · Education: Boston University · Location: Boston · 500+ connections on LinkedIn. View Massimiliano Versace’s profile on LinkedIn, a professional community of 1 billion members.
www.linkedin.com
Alf Kuchenbuch used to work for Analog Devices, by the way…
Comment by Neuromorphicism (
https://www.linkedin.com/company/neuromorphicism/posts/)
aka Michał Antropik (
https://www.linkedin.com/in/michał-antropik-a97458165/):
Repost by Aaron Goldberg, SVP at Quantum Ventura:
Analog Devices’ top AI predictions for 2026, from decentralized humanoid robotics and analog AI compute to agentic systems.
emag.directindustry.com
1/ Decentralized AI in Humanoid Robotics
Massimiliano Versace, VP of Emergent AI at
Analog Devices, predicts that by the end of 2026, decentralized AI architectures combining sensing with neuromorphic and in-memory compute will move from pilot programs to early commercial deployment in
humanoid robotics:
“We’ll see humanoid robotics systems getting a bit closer to biological systems, where local circuits in sensory organs and spinal pathways handle reflexes and balance, allowing smoother, more adaptive movement, drastically reduced power consumption, and freeing the central brain to think and plan.”
Versace also emphasizes the role of intelligent sensors embedding novel AI compute directly within the sensor:
“By enabling real-time, low-latency AI at the edge, robots will become more efficient, responsive, and capable of near-biological sensory-motor skills,” he explains.
This fusion of decentralized AI and advanced compute architecture will allow humanoid robots to operate fluidly in dynamic environments, paving the way for practical and pervasive applications.
2/ The Rise of Analog AI Compute
Analog AI compute, once sidelined due to precision and scalability limitations, is expected to make a comeback in 2026. Unlike conventional digital processors that separate sensing and computation, analog AI performs inference using the physics of the sensing substrate itself, collapsing these layers into a unified framework.
Versace predicts that:
“By the end of 2026 we’ll see initial deployments and adoption of this technology, particularly in robotics, wearables, and autonomous applications, where analog AI enables real-time responsiveness, smoother interactions, longer battery life, and more natural behavior in the devices they power.”
3/ AI Steps Into the Physical World
For Paul Golding, VP of Edge AI and Robotics at Analog Device, AI will be moving off screens and into the physical world as the next frontier of AI will be
physical intelligence:
“The scaling laws that powered the success of large language and
vision models will continue through 2026 but will extend into models that learn from vibration, sound, magnetics, and motion (stubborn attributes of the physical world). I predict these physical reasoning models will migrate from the datacenter to the edge, powering a new type of fluid autonomy that thinks and acts locally, sensitive to local physics and without recourse to centralized servers.”
Golding highlights hybrid “world models” that integrate mathematical and physical reasoning with data-driven sensor-fused dynamics.
“Think of a
mobile factory robot that can reason for itself and determine what to do when faced with an unexpected obstacle.
These systems will not only describe their environments but actively engage with them, learning from their own experience.
4/ Audio as a Dominant AI Interface
By 2026, audio is expected to become a key reasoning channel in consumer electronics. Spatial sound, sensor fusion, and on-device reasoning will transform earbuds, AR glasses, and in-vehicle sound systems into context-aware companions. According to Golding,
“These technological leaps will lead to significantly better noise cancellation in our hearable devices, improved battery life, and new form factors that haven’t yet been imagined. The always-in-ear hearable experience, already on the rise among Gen Z, will become increasingly prevalent due to the “super-human” hearing of context-aware AI.”
5/ Agentic AI and Physically Intelligent Models
Another prediction refers to
agentic AI, the next wave of edge AI.
AI agents are AI systems that can autonomously decide and act in the world. Golding predicts the mainstream adoption of digital twins to give AI models physical-system awareness:
“2026 will see the mainstream arrival of digital twins to imbue large models with physical-system awareness. Imagine AI models learning to predict forces instead of text, but within the safety of a scalable simulated environment. Physically intelligent foundation models will merge reasoning with sensor intelligence to orchestrate machines, simulations, and data.”
Today, many factories have the technology to do predictive maintenance. But Golding expects a future where an agent on the factory floor acts on that prediction:
“It autonomously reroutes the production line to a healthier machine, adjusts the strained machine to 70% to extend its life, and coordinates with supply chain agents to adjust inventory—all without human intervention.”
6/ Micro-Intelligence: AI’s Agentic “Inception”
A final prediction for 2026 is the emergence of micro-intelligences. These tiny, recursive models with deep reasoning in narrow domains are capable of running at the edge. These systems will bridge the gap between today’s rigid programmed AI and sprawling foundation models like GPT-5, orchestrating specialized agents to solve practical engineering problems.
For Golding,
“I predict the rise of new types of AI benchmarks designed to measure and encourage a new kind of engineering intelligence—multiagent micro-intelligences that can collaborate to solve complex engineering problems, moving from the world of abstract mathematical challenges (like Math Olympiads) to practical problem-solving systems.”
