embedded-ai
(8)
Phantom AI Stopped Chasing the Robotaxi - and Sold the Safety Layer Every Truck Was Missing
A bruising highway demo exposed the gap between autonomy theater and dependable safety. Phantom AI spent the next eight years making the less glamorous bet - affordable software that can see a pedestrian, fuse a radar return and help an ordinary vehicle stop.
Gigacatalyst
Gigacatalyst is a Y Combinator-backed (P26) startup that embeds AI agents inside B2B SaaS products so non-technical sales and customer success teams can build the missing features a customer needs just by describing them in plain English. Its agentic builder reads the host product's own APIs and design system, then generates working microapps in about a minute, closing the feature gaps that stall enterprise deals and drive churn.
XMOS
XMOS is a fabless semiconductor company based in Bristol, UK, that designs the XCORE family of multicore, deterministic processors for embedded voice, audio, AI and control at the edge. Founded in 2005 by a team rooted in the University of Bristol and Inmos transputer heritage, XMOS built its name on far-field voice processing (VocalFusion, the first far-field solution qualified for Amazon Alexa) and now positions XCORE as a 'Generative System-on-Chip' (GenSoC) - silicon whose behaviour can be described in natural language while guaranteeing real-time, deterministic performance.
ModelCat AI
ModelCat AI, formerly the low-power chip pioneer Eta Compute, builds an agentic, hardware-aware platform that automatically designs, trains, optimizes and validates machine-learning models for embedded, edge and IoT devices. Its patented AI-in-the-Loop workflow, backed by a physical hardware farm that calibrates results against real silicon, compresses a model-development cycle that once took 12 to 24 months down to a matter of days. The Sunnyvale company works with chipmakers including NXP and Alif Semiconductor to hand developers ready-to-run models tuned to the exact constraints of the hardware they will run on.
Applied Brain Research
Applied Brain Research (ABR) is a Waterloo, Ontario AI hardware company spun out of the University of Waterloo's Centre for Theoretical Neuroscience. It builds brain-inspired chips and software that run real-time AI - full-vocabulary speech recognition, text-to-speech and sensor processing - directly on edge devices at power levels measured in milliwatts. Its patented state-space models and the Legendre Memory Unit power the TSP1 Time Series Processor, which ABR calls the world's first single-chip solution for full-vocabulary speech recognition, doing the work of cloud voice AI while consuming 10 to 100 times less power.
Inventive
Inventive is a San Francisco startup building customer-facing AI that B2B SaaS companies embed inside their own products. Its first product, an embedded AI Analyst, lets a software vendor's customers use plain language to explore data, build custom reports, and act on insights. What sets Inventive apart is the layer around the agent: built-in quality and learning systems that test across accounts, catch mistakes, and propose fixes so the AI measurably improves week over week instead of degrading after a good demo. Founded in 2022 by ex-Google, Meta, and Microsoft product and platform veterans, the company launched out of stealth in June 2024 with a $6.5M seed round led by Wing VC.
Sensory, Inc.
Sensory is a Santa Clara-based company that builds on-device AI for voice, sound, and biometrics. Since 1994 its embedded software has shipped in more than three billion consumer electronics, from toys and wearables to cars and smart speakers, giving devices the ability to listen, see, and verify without sending data to the cloud.
MIPS
MIPS is a San Jose-based semiconductor IP company with a 40-year heritage in RISC computing. After abandoning its own legacy architecture, MIPS reinvented itself around the open RISC-V instruction set and now builds compute subsystems for autonomous platforms across automotive, industrial, robotics, and edge AI markets. Its Atlas portfolio - spanning Sense, Think, Act, and Communicate building blocks - targets the low-latency, safety-critical demands of what the company calls Physical AI. In July 2025, GlobalFoundries announced a deal to acquire MIPS.