Maven Robotics emerges with $100M to challenge warehouse automation rivals

Updated September 10, 2026 5:58 pm

Maven Robotics has come out of stealth with $100 million in new funding and a plan to scale its warehouse automation systems, after winning early business from a major consumer goods company despite having little more than an idea, a small team and a robot illustration in 2024. The startup is now betting that practical, end-to-end industrial robotics can beat flashier competitors in one of the toughest corners of physical AI.

The Santa Clara, California-based company says its third-generation robots are already operating in customer sites, and it intends to build 250 more while starting work on a fourth-generation platform. Maven’s pitch is straightforward: instead of solving a narrow robotics task, it wants to automate entire warehouse workflows, from receiving goods to building pallets for outbound shipping.

What Maven Robotics is building

Maven Robotics is developing wheeled industrial robots with two articulated arms that can move through warehouse environments, handle boxed goods and reorganize inventory for shipping. The company’s first commercial focus is mixed palletizing, a process that requires assembling pallets with a combination of different products based on changing demand from retail stores.

That may sound narrow, but in warehouse operations it is a valuable and repetitive job that still depends heavily on human labor. Customers often need to reconfigure shipments within a day or two of goods hitting store shelves, forcing warehouse teams to pull and repack products manually. Maven wants to automate that chain end to end.

The robots are designed to travel at speeds of up to 10 miles per hour and can lift loads of as much as 30 kilograms, according to the company. In a training area at Maven’s facility, the machines use suction-based grippers to pick, place and stack cartons in a controlled environment before being deployed in live operations.

Why mixed palletizing matters

Mixed palletizing sits at the intersection of logistics efficiency, labor pressure and inventory volatility. Distribution centers serving retailers frequently have to assemble pallets containing different goods destined for the same store, and the mix can change quickly as demand shifts.

That makes the task well suited to robotics, but only if the system can work reliably in messy, real-world conditions. Maven’s argument is that warehouse operators do not need a robot that can do everything in theory; they need a machine that can work long hours, integrate with existing software and reduce the cost of moving goods.

Hamza Derbas, Maven’s co-founder and chief executive, said the company is focused on building robots that connect directly into warehouse systems and take on the full job rather than just a single step.

He described the company’s approach as one that links into a warehouse management system on one side and produces truck-ready pallets on the other. That full-process framing is central to Maven’s sales pitch and, in the company’s view, the reason it won an early deal against competitors that already had robots in market.

How a startup with almost nothing won an early customer

Maven’s origin story is unusual even by startup standards. In 2024, the company had not yet built a product and was operating with little more than a team and a visual mock-up of a robot. Still, Derbas learned that a large consumer goods company was meeting several robotics vendors about warehouse automation and pushed to get in front of the customer.

Rather than leading with technical theory, he asked to visit the customer’s facilities. Once there, he and his team watched workers perform the job and identified the process bottlenecks where automation could create immediate value.

That field-first approach helped Maven beat established rivals in the deal, according to the company. The startup’s case was not that its robots were already the most advanced, but that it understood the industrial workflow better than companies coming from research or narrowly engineered robotics backgrounds.

From warehouse observation to deployment

The company says that after roughly two years working with that initial customer and a handful of other partners, it now has as many as eight robots operating for 16 hours a day with uptime of 99% or higher. In industrial robotics, those figures are meaningful because they suggest a system is moving beyond demos and into production-like reliability.

At Maven’s facility in Santa Clara, a live video feed shows robots working in a customer environment while people walk nearby. That combination of remote monitoring and on-site human supervision is typical of early warehouse automation deployments, where operators still need a way to intervene if a machine gets stuck or a workflow changes unexpectedly.

Who backed the new funding round?

Maven’s $100 million round was led or joined by RoboStrategy, LocalGlobe, Vine Ventures and XTX Markets Ventures. The funding will support production of the company’s third-generation robots and the start of design work on a fourth-generation platform.

The size of the raise is notable because it arrives at a time when investors are still sorting winners from a crowded field of physical AI and robotics startups. Many firms have strong technical demos. Far fewer have durable customer deployments, repeatable revenue and proof that their systems can survive the complexity of real warehouses.

Investor Jack Pearson of RoboStrategy said the company stands out because of its industrial systems experience, not because it comes from a research lab culture optimized for a single machine architecture.

That distinction reflects a larger debate in robotics. Some startups are trying to build general-purpose systems from the top down, while others focus on solving one operational problem at a time and expanding only after they have data and customer confidence. Maven clearly sits in the second camp, at least for now.

Why warehouse robotics is so hard

Warehouse robots must deal with changing layouts, inconsistent lighting, damaged packaging, human workers, software integration and a constant stream of edge cases. A polished demo can look impressive while still failing under the pressure of a real logistics center running around the clock.

Maven’s leadership argues that robotics companies often underestimate those realities. The company says its advantage comes from designing around operations rather than around a single technical benchmark. In practice, that means building systems that can tolerate interruptions, use live data and improve continuously once they are deployed.

The role of self-driving car veterans

Maven is one of several physical AI startups staffed by former self-driving car engineers. That background is useful because autonomous vehicles produced some of the most advanced tooling for collecting, labeling and learning from real-world data at scale.

According to the company, the same basic loop applies to warehouse robots: gather data from machines in operation, send it back quickly, retrain models, evaluate performance, test variants, adjust weights and then redeploy the improved system. The pace of that cycle can determine whether a robotics company gets better over time or stalls after its initial launch.

Derbas spent years at Apple before starting Maven, including work in the company’s special projects organization, which was widely believed to be connected to the abandoned self-driving car effort. Before that, he worked in automotive engineering with a focus on electric vehicles. He co-founded the company with his brother, Khalid Derbas, who is now chief financial officer and previously worked in private equity.

What makes Maven different from humanoid robot rivals?

Maven is deliberately rejecting the idea that humanoid form factors are necessary for warehouse automation. While some robotics companies are chasing bipedal designs that can theoretically operate in human-like environments, Maven is focused on wheeled robots that are cheaper, simpler and more practical for the tasks it is targeting.

Derbas has praised some competitors but argued that two-legged machines introduce unnecessary complexity and cost for the jobs his company wants to solve. In his view, industrial buyers care less about futuristic design and more about return on investment, uptime and predictable performance.

That is an important strategic fork in the robotics market. Humanoids may eventually prove useful in some settings, but warehouse operators often need a dependable machine that can be deployed without redesigning an entire facility. Maven is staking its business on that more incremental path.

Agility and the market for industrial robots

Agility Robotics is the closest public-market comparator often mentioned in this space. The company is headed for a public debut through a $2.4 billion SPAC transaction and has positioned itself around safety and specific industrial workflows. But its bipedal robot design reflects a different thesis about how automation will evolve.

Maven’s view is more skeptical. The startup says the economics of warehousing reward systems that can work immediately and scale efficiently, not necessarily machines that mimic human anatomy. Whether that becomes a durable advantage will depend on execution as much as on product philosophy.

How does Maven plan to expand beyond palletizing?

Maven plans to use palletizing as the foundation for a broader move into material handling, automation and eventually fabrication. The company says it needs more data to teach robots to manipulate a wider set of materials and to cope with more complex tasks than stacking packaged goods.

That next phase is where the technical challenge grows sharply. Once a robot moves beyond boxed consumer goods, it has to handle irregular objects, softer materials and less controlled environments. Those tasks demand better perception, better grasp planning and more adaptable control systems.

To move in that direction, Maven says it is combining data from its own robots with third-party sources and has even created special pincer-style gloves that let people mimic the intended gripper shape during training or research. The goal is to accelerate the collection of useful manipulation data before the robots are asked to take on more difficult work.

• Current focus: mixed palletizing in warehouse and distribution settings
• Next phase: broader material handling and automation
• Longer-term ambition: fabrication-related tasks
• Training strategy: combine real deployment data with additional sources
• Product strategy: solve one customer problem at a time before broadening scope

Why data is the real moat in physical AI

For robotics startups, the hardware is only part of the product. The harder problem is building a data engine that improves the machine every time it works in the field. Maven appears to understand that the most valuable asset may be the operational feedback loop, not the robot chassis itself.

That is why the company is investing in fast data pipelines and repeated retraining. Each deployment can reveal new failure modes, from package misalignment to unusual pallet loads or temporary changes in warehouse traffic. Capturing and learning from those moments can turn a narrow system into a more robust one.

Still, the strategy has limitations. Focusing on one task at a time can build a business, but it can also leave a startup vulnerable if another company or frontier AI lab quickly produces a more capable model that generalizes across a much wider set of tasks. Maven acknowledges that risk implicitly by framing itself as a company trying to solve industrial labor rather than win a pure model race.

Derbas said the company is not competing to build the most famous robotics model; it is competing to make industrial labor easier to automate at scale.

What comes next for Maven Robotics?

The immediate challenge is execution. Maven must turn promising deployments into a repeatable manufacturing and sales engine, deliver the 250 robots it now plans to build and prove that its fourth-generation design can outperform its current system.

It also has to decide how quickly to expand beyond palletizing without stretching its engineering team too thin. Every new task increases the demand for data, testing, safety validation and customer support. In robotics, premature expansion can be as dangerous as moving too slowly.

For now, Maven is presenting itself as a company built around industrial pragmatism rather than futuristic spectacle. That message may resonate with warehouse operators looking for reliability over novelty and with investors who want signs that physical AI can produce businesses, not just prototypes.

The startup’s emergence underscores a broader shift in the robotics sector. The biggest opportunities may no longer belong only to the companies with the flashiest demos, but to those that can quietly solve expensive, repetitive labor problems in a way that pays back fast enough for industrial customers to care.

If Maven’s early deployments hold up, the company could become one of the stronger examples of how to build a robotics business from the warehouse floor up. If not, it will join a long list of startups that discovered that real-world automation is far harder than it looks from a pitch deck or a demo room.