Neuron Industries publicly launched its Cortex AIC on August 24, 2026, a $2,950-to-$6,450 industrial controller that programs in conversational English and promises to displace the programmable logic controllers that have dominated factory automation since the Reagan administration. The Y Combinator–backed startup targets Q4 2026 for general availability and is already running paid pilot projects, according to an Aug. 24 press release.
The timing isn't coincidental. A generation of controls engineers who cut their teeth on ladder logic in the 1980s and '90s is nearing retirement, and the factories they've kept running face an uncomfortable question: Who's going to maintain all that hardware once they're gone?
Dennis Ren, Neuron's CEO, puts it bluntly. The equipment is "frozen in the '90s," he said, and the expertise to keep it alive is walking out the door.
What Sets Cortex Apart
The device runs deterministic, real-time Python at scan rates reaching 4,000 Hz with 0.5-microsecond P99.9 jitter on its flagship model, the El Segundo, California–based company says. Those specs matter in industrial settings where timing errors of even a few milliseconds can ruin a product batch or damage equipment.
An AI assistant does the heavy lifting on the programming side, translating plain-English instructions into controller code, generating unit tests, and verifying behavior in simulation before anything touches the factory floor. The system ships with Synapse, an on-board integrated development environment; an HMI that overlays live sensor readings on process diagrams; and Hippocampus, an on-device historian that sidesteps the per-tag licensing fees that often inflate costs in traditional setups.
Protocol support includes EtherCAT, PROFINET, and Modbus TCP for fieldbus communications. Operators can run the controller air-gapped or stream data via MQTT, OPC UA, and REST, Neuron said.
"Neuron's mission is to drop the barrier to entry for industrial automation to the floor," Ren said in the launch statement.
The Team Behind It
Ren brings experience from Freeform, where he built control stacks for autonomous metal 3D printing, and stints at Apple's Vision Pro system engineering team, Amazon Go, and Tesla's Fremont plant. His co-founder, CTO Kenneth Rhee, previously worked on Amazon's EBS and Silk browser teams before founding WorkIPO, an ERP and manufacturing execution system provider for industrial clients.
Rhee frames the product as an attempt to import software-industry best practices into a world that's been slow to adopt them. "Software engineering got 30 years of continuous innovation: version control, testing, code review, and fast feedback," he said. "The goal is to enable engineers to develop industrial hardware at the speed of software."
Whether factory operators, often risk-averse by necessity, will embrace that speed is another matter.
First Deployments
CureWood, an engineered-wood manufacturer in Surrey, British Columbia, replaced a 1991 Siemens S5 controller with Cortex in August. Neuron decoded the legacy machine code, scanned ladder-logic printouts, then ported the logic to Python and verified all 300 control statements in simulation. The actual cutover took eight hours.
The project ran $2,950 for the controller, roughly $3,000 for third-party I/O modules, and $4,000 for integration and travel, according to CureWood President Ben McKinnon in a company case study published in August. "I honestly feel like we just went from being in 1991 to being in 2026 over the course of like four days," he said.
Neuron also installed controllers at Rise Reforming, a Y Combinator S26 cleantech startup piloting biogas-to-methanol conversion at a Chicago wastewater plant, based on LinkedIn posts from both companies in August.
Industry Response and Competitors
Established vendors aren't standing still. Rockwell Automation's FactoryTalk Design Studio Copilot generates ladder logic and structured text. Siemens introduced the TM-NPU AI coprocessor module for its S7-1500 line. CODESYS launched an MCP Server in July that lets large language models generate IEC 61131-3 code. Vention commercialized its MachineMotion AI controller in May 2025, also targeting buyers looking beyond traditional PLCs.
But the incumbents are adding AI features atop their existing PLC runtimes rather than rethinking the architecture from scratch, which is Neuron's bet.
OEM Magazine offered a note of caution in an Aug. 6 analysis, arguing that PLCs won't vanish quickly. Safety certification timelines, the need to prove return on investment, and the industrial sector's preference for equipment that lasts decades all slow adoption of new platforms.
Ren, speaking to Atoms Not Bits in July, said the PLC market moves 30 million units annually. He compared the industry to Nokia before the smartphone revolution. "It smells like Nokia before they died," he said.
The company hasn't disclosed funding amounts beyond its Y Combinator participation. Whether Cortex can crack an entrenched market dominated by brands that have spent decades earning factory operators' trust remains an open question. But the looming retirement wave suggests the window for new entrants may be opening wider than it has in years.
