A two-person Y Combinator startup says its software can replan production lines in seconds when machines fail or orders change, taking aim at an entrenched problem in manufacturing software.
Ethan Phillips and Kolya Warwick Schwinge built Orvixo to solve a frustration they believe still plagues factory floors: production schedules that break the moment reality intrudes. Their pitch is straightforward. When a machine goes down or a rush order arrives, most planning software requires manual reworking by specialists who know how to coax the system back into alignment. Orvixo's platform, they say, does the replanning automatically.
The London-based company emerged from Y Combinator's Summer 2026 batch with a product aimed squarely at manufacturers in precision machining, metal fabrication, complex assembly, and similar industries where production planning remains as much art as science. Phillips and Schwinge claim their system can go live in weeks, not the months or years that legacy implementations often require.
Whether that promise holds in practice remains to be seen. No public customer logos, case studies, or named pilots have been published on the company's site, YC profile, or LinkedIn as of August 11, 2026. What it does have is a clear target: an advanced planning and scheduling market where failure rates run high and frustration runs higher.
A Market Ripe for Disruption, or Just Difficult
The startup enters a space littered with disappointed buyers. McKinsey reported on April 7, 2026, that around 65 percent of advanced planning and scheduling programs fail due to poor data management. The consulting firm pointed to poor data management as a primary culprit, though anyone who has watched a factory planner wrestle with scheduling software might offer additional theories.
Orvixo's approach hinges on building what Phillips and Schwinge call a "live model" of factory operations. The system connects to enterprise resource planning and manufacturing resource planning platforms, pulling in data about machines, staffing, capacity constraints, and existing orders. It then generates schedules designed to balance competing priorities: delivery dates, throughput, inventory levels, cost.
The company's website includes illustrative UI captions showing a schedule maintaining 97.1 percent on-time delivery even after a machine breakdown, with the system adding 1.5 hours of overtime to compensate. Another caption claims jobs were rescheduled in 0.8 seconds. Third-party validation for these figures has not been published, and the specifics of the test scenarios remain unclear.
Planners still make the final call. They review the AI-generated schedules, adjust them if needed, and can run hypothetical scenarios before approving overtime or expediting orders. That human-in-the-loop design may prove critical, given the stakes involved when factory schedules go sideways.
Founders Who Showed Up With Donuts
Phillips brings a background in growth work at Attio, a customer relationship management startup, along with a physics master's from Cambridge where he studied quantum computing. Schwinge appears in UK Companies House filings as a Person with Significant Control in a March 2026 document, though his professional background before Orvixo is less detailed in public records.
The two have embraced a notably hands-on sales strategy. "We rented a GMC, bought a few hundred donuts and visited manufacturers across the state to meet the people actually running their operations," Phillips wrote on LinkedIn in early August 2026, recounting a Michigan roadshow during the Advanced Manufacturing Expo. That kind of field work suggests founders who understand that selling to factory planners requires more than slick demos.
"We're building the AI-native operating layer for modern factories," Phillips announced on LinkedIn when the company joined Y Combinator. The phrasing is ambitious, positioning Orvixo not as incremental improvement but foundational infrastructure.
UK corporate records show some administrative reshuffling. The company incorporated ORVIXO LTD on March 16, 2026, then filed to dissolve that entity in early August. That pattern is common among YC-backed startups, which often reincorporate as U.S. C-corporations after going through the accelerator. Orvixo has not confirmed whether it completed that reincorporation or what structure it currently operates under.
Competing in a Crowded Field
Orvixo faces well-established players with deep pockets and long customer lists. Siemens Opcenter APS, PlanetTogether, Kinaxis, o9 Solutions, Blue Yonder, SAP's Production Planning and Detailed Scheduling module, and Dassault Systèmes' DELMIA Ortems all compete for manufacturing planning budgets. Each has built integrations, accumulated implementation knowledge, and weathered the inevitable complications that arise when software meets factory floor realities.
Newer entrants are circling the same opportunity. Flexciton focuses specifically on semiconductor fabrication scheduling, where the complexity and capital costs make optimization particularly valuable. Dime markets itself as a broader factory operating system. Both emphasize AI capabilities, making this a space where the technology hype cycle intersects with genuine operational pain points.
Demand for better scheduling tools appears robust, at least according to industry surveys. A Deloitte study from May 2025 found 35 percent of manufacturers ranked advanced production scheduling among their top two investment priorities for the following two years. The same survey revealed that 46 percent faced moderate to significant challenges hiring qualified planning and scheduling staff. If those trends hold, the market opportunity is real.
The Adaptability Argument
Schwinge laid out the company's core thesis in a blog post shared in late July 2026. Traditional advanced planning systems, he argued, become outdated almost as soon as they go live. Factories change constantly: new machines arrive, old ones get decommissioned, product mixes shift, staffing fluctuates. Most scheduling software requires experts to manually update the underlying models when those changes occur.
"An optimiser does not schedule a factory. It schedules a model of one," Schwinge wrote. The distinction matters. If the model diverges from reality, the schedules it produces become progressively less useful. Orvixo's answer is a system where planners and engineers can describe operational changes in natural terms, with AI proposing and testing model updates rather than requiring deep technical reconfiguration.
Whether that approach works in practice will depend partly on how well the AI handles the messy details of real manufacturing environments. Factories are full of informal constraints, tribal knowledge, and exceptions that defy easy modeling. The planners who currently keep production moving often know things that never make it into formal systems.
The company declined to share pricing details, funding amounts beyond the Y Combinator backing, or specifics about current customer deployments. With a team of two, Orvixo is betting that a small, focused group can move faster than larger competitors weighed down by legacy architectures and enterprise sales cycles.
That bet has worked before in enterprise software, though the graveyard of manufacturing tech startups suggests it is far from guaranteed. For now, Orvixo remains a promise: faster deployment, better adaptation, schedules that stay useful when reality refuses to cooperate. Proving those promises will require more than screenshots and LinkedIn posts.
