Edge AI Is Becoming a Catalog Feature, and Acquirers Are Pricing It That Way

October 27, 2026  •   David Pring-Mill

Texas Instruments now sells microcontrollers with a built-in neural processing unit at commodity prices, and it says it intends to carry that block across its entire microcontroller portfolio. Edge AI, in other words, is turning into a feature of the chip catalog rather than a product of its own.

That shift arrived in the middle of a buying spree. Over 2025 and 2026, onsemi agreed to acquire Synaptics, Microchip bought Hailo, Analog Devices bought Alif Semiconductor and NXP bought Kinara. Each acquirer describes buying the compute together with what surrounds it: wireless connectivity, software and an installed base of customers.

Where the price can be read, my analysis finds no premium in it that can be traced to the edge AI compute. onsemi’s revised all-cash offer valued Synaptics only slightly above the peers its own board used for comparison, at a time when Synaptics’ edge AI processor line was not yet a meaningful source of revenue. Hailo, once richly valued in private funding, was sold for a small amount after struggling to produce significant revenue, according to Globes.

The market itself is growing. The report sizes edge AI embedded processors, the microcontrollers, embedded processors and vision chips that run AI inference inside devices, at $2.1 billion in 2025, growing about 20% a year to $5.3 billion by 2030. AI-capable microcontrollers are the largest and fastest-growing of its three device tiers. Incumbent microcontroller and embedded-processor houses hold most of the market, and Ambarella leads the independents that remain.

For an acquirer or an investor, I read the price record as a lesson about where the value sits: in what surrounds the neural processing block, not in the block itself. Where the report does find a higher multiple, it went to a fast-growing robotic perception business that already earns revenue: Cognex’s purchase of RealSense. The thesis concerns compute, not every asset with edge AI in it. The same test applies to the next deal: ask what the buyer is paying for besides the neural processing unit.

The full analysis, with the price record deal by deal, the market sized by device tier, supplier type and application, and three scenarios to 2030, is in the report Edge AI Embedded Processors Market, 2025–2030, available on the report’s page.

  1. ← Previous Article Pending the CNA, the Network Carrying Most of Mexico’s Challenger Lines Changes Hands

    Next Article → FERC Wants Faster LNG Capacity Increases. Plaquemines Shows Where They Pay Off