Nicolas Yeh likes to tell people he grew up in warehouses—which, in a sense, he did. His family's third-generation supply chain business moved consumer goods into U.S. retail stores, and young Yeh spent enough time around forklifts and pallet racks to develop what he calls an informed impatience with how slowly the industry modernizes.
When warehouse labor costs surged in the pandemic's wake, Yeh had an idea that felt either obvious or naïve, depending on your perspective: Maybe the robotics industry had been solving the wrong problem. Warehouses didn't resist automation because they doubted its value. They resisted it because robots cost too much and took too long to deploy.
Now Yeh and co-founder Joshua Ibrahim are testing that theory with Manifold Industries, a Los Angeles-based startup that emerged this summer from Y Combinator's latest batch. Their pitch? Forget six-figure capital outlays and eighteen-month integration timelines. Pay by the pick instead.
Manifold says its autonomous robots can roll into a warehouse and start working quickly, handling up to 150 cases per hour at a per-pick rate the company claims beats human labor costs—all with zero money down.
It's an audacious model in an industry littered with ambitious promises. Whether it actually pencils out is another question entirely.
The Usage Bet
The economics here hinge on flipping the traditional automation playbook. Instead of purchasing robots outright or locking into multi-year leases, operators pay only for completed picks. Manifold hasn't disclosed the actual per-pick rate, which makes evaluating the value proposition difficult, but the approach mirrors a broader shift toward Robotics-as-a-Service that's been gaining momentum across the logistics sector.
The hardware itself targets familiar warehouse workflows: case picking, trailer loading and unloading, palletizing. Each robot features a six-axis arm with a suction gripper mounted on a mobile base equipped with LiDAR, depth sensors, and camera arrays. The system handles boxes up to 65 pounds and promises 16 hours of runtime per charge.
What Manifold emphasizes—perhaps more than the technology itself—is speed of deployment. Traditional warehouse automation projects often require months of planning, facility retrofits, and painful integration with legacy systems. Manifold claims its robots can go live quickly, building real-time maps of racks, pallets, and pick zones while slotting into existing warehouse management systems without major infrastructure overhauls.
That's the theory, anyway. The company says it has systems in operation, though it hasn't identified customers or provided independently verified performance data from deployments.
Chasing the Long Tail
Yeh frames the opportunity in stark terms: The U.S. warehouse industry spends roughly $75 billion annually on labor, yet most facilities remain stubbornly manual. In company materials, Manifold cites research suggesting only 10% of American warehouses have deployed any automation, though Yeh himself wrote on LinkedIn in July that "over 80% of US warehouses still have ZERO automation whatsoever."
The exact figures vary depending on how you define automation, but the broader reality is hard to dispute. Despite high-profile deployments by Amazon, Walmart, and major third-party logistics providers, the vast majority of mid-sized distribution centers still operate much as they did two decades ago. Capital requirements and integration complexity have effectively locked out smaller operators who might want automation but can't justify the upfront investment or risk.
That's the market Manifold is targeting. Not the massive fulfillment centers with dedicated automation teams, but the 500,000-square-foot regional warehouses and third-party logistics operators who've been priced out of the robotics revolution.
In his launch post this summer, Yeh claimed the robots work "up to 5x faster than humans," noting that human pickers in B2B operations typically handle 30 to 50 cases per hour, while Manifold's robots target 150 cases per hour when mobile. Whether those numbers hold under varied warehouse conditions—different product mixes, seasonal surges, facility layouts—remains untested at scale. The company hasn't released third-party validation of its performance claims.
When Robots Get Stuck
The technical architecture reveals how Manifold plans to scale without requiring an army of on-site technicians. The robots handle autonomous navigation, object recognition, grasp planning, and movement using what the company describes as AI-based systems. When they encounter something they can't resolve—a crushed box, an unusual SKU, a blocked aisle—remote operators take over via teleoperation.
This human-in-the-loop design has become standard practice in commercial robotics, allowing systems to maintain high uptime without full autonomy. The orchestration layer integrates with customers' existing ERP, WMS, and warehouse execution platforms, positioning the robots as augmentation rather than replacement.
Still, the model raises questions Manifold will need to answer as it scales: How many remote operators does a fleet require? What happens when multiple robots hit edge cases simultaneously? How do margins hold up when every pick needs to be profitable and some percentage require human intervention?
Manifold is currently accepting reservations for what it's calling a "Fall 2026 Fleet Rollout," though specific deployment timelines and customer commitments haven't been disclosed.
Unconventional Backgrounds
Yeh brings an unusual résumé for a robotics founder. Before launching Manifold, he worked as an analyst at Balyasny Asset Management covering analog semiconductors, then moved to Starwood Capital Group handling hotel acquisitions. He holds a degree from the University of Pennsylvania and lists experience with a $60 million hotel sale and asset management responsibilities for an Extended Stay America joint venture.
Co-founder Joshua Ibrahim's background, according to Y Combinator's directory, includes software engineering internships at Meta—working on Messenger peer-to-peer systems and Instagram misinformation detection—and degrees in electrical and systems engineering from Penn. The YC page also lists him as holding a PhD in applied mathematics from Caltech, though Caltech's current directory shows Ibrahim as a non-degree student—a discrepancy that would require clarification directly from the founders.
The company emerged publicly in mid-2026 as part of Y Combinator's summer cohort and currently lists a team of two. Interested operators can reach Yeh directly through contact information on the company's website.
A Crowded Field
Manifold is entering a warehouse robotics market that's both hot and increasingly saturated, with well-funded competitors already making aggressive moves.
Boston Dynamics' Stretch robot claims handling speeds of 600 to 800 cases per hour for truck unloading—significantly faster than Manifold's stated target. Pickle Robot raised a $50 million Series B in February and already has orders for more than 30 units. Nomagic announced in March it would deploy up to 50 AI-powered robots across Zalando's European fulfillment network.
Those competitors generally target larger operators with capital budgets for automation pilots and the scale to justify traditional pricing models. Manifold's per-pick approach positions it toward smaller facilities that might never hit the volume thresholds or ROI calculations required for conventional robotics investments.
Whether that strategy succeeds depends on execution challenges the company hasn't yet faced at meaningful scale: managing distributed fleet logistics, providing support across geographically dispersed sites, maintaining margins when pricing is tied to individual transactions rather than capital sales.
Yeh has been active in founder outreach, posting in Reddit communities for third-party logistics providers and supply chain professionals throughout June and July, pitching the robots and seeking early conversations with operators. Response on LinkedIn has been mixed—enthusiasm for the pricing model, but also skepticism about practical limitations. One commenter questioned whether suction grippers dropping cases into bulk containers would work for fragile products like glassware.
Fair question.
Timing the Window
Industry forecasts suggest Manifold's timing may be right, even if the business model remains unproven. Gartner predicted in April that by 2030, half of new warehouses in developed markets will be designed as robot-centric facilities where human workers become optional rather than central. Prologis forecasts that up to 50% of modern European warehouses could incorporate some form of automation by 2035.
Labor constraints, safety concerns, and relentless throughput demands continue pushing operators toward mechanization. The question isn't whether warehouse automation will expand—it's who captures the long tail of facilities that traditional robotics companies can't profitably serve.
For now, Manifold remains a two-person operation with target specifications and deployment claims awaiting real-world validation. The company hasn't disclosed specific per-pick pricing, minimum commitments, or service-level agreements. Operators interested in the Fall 2026 rollout can schedule calls through the company's website.
The bigger question—the one that will determine whether Manifold becomes a case study in innovative business models or cautionary tale about hardware economics—is whether pay-per-pick pricing can actually work at warehouse scale. Not just for customers, but for Manifold itself.
If it does, the model could unlock a massive market that's been waiting for someone to remove the upfront risk. If it doesn't, Manifold will join the long procession of robotics startups that discovered hardware is hard and warehouses are unforgiving.
Either way, it won't take long to find out.
