Physical AI Is Here. The Midwest Is Ready.

TechNexus Venture Collaborative and Wells Fargo bring together tech leaders to discuss the real-world deployment of robotics and Physical AI across Chicago’s heavy industries.

For years, physical AI has lived mostly in labs and pitch decks. That's changing — and the reason comes down to a simple shift: the foundational technology largely exists. The missing piece is deployment. Across warehouses, construction sites, manufacturing floors and logistics networks, the question is no longer whether robots and intelligent machines can do the work. It’s whether the industries that need them most are ready to commit.

That is where Chicago enters the conversation with a distinct advantage. This is a city — and a broader Midwest corridor — built around the exact sectors physical AI is designed to transform. The companies here are not passive observers waiting for Silicon Valley to hand them a solution. They are operators running fleets of robots, managing millions of square feet of industrial real estate, and building the heavy equipment that moves the physical world. They understand the problem from the inside. That combination of industrial density, domain expertise and institutional scale is the foundation for making physical AI actually work.

At Chicago at Work: Robotics & Physical AI in Industry , an event hosted by TechNexus Venture Collaborative and Wells Fargo on June 2, moderators Fred Hoch and John Huber guided two separate conversations with eight panelists who detailed their perspectives on this emerging space.

The first panel focused on bringing physical AI technologies from the lab environment to the capital markets. Panelists included Hyde Park Angels Managing Partner and OMAXN co-founder Pete Wilkins, Nine Four Ventures General Partner Kurt Ramirez, Jump Capital Partner Jason Felger, and Northwestern University Professor Ryan Truby.

The second panel centered around the physical AI buyers who are operating the future. Panelists included CAT Digital Director of Artificial Intelligence Digital Product Management Charlie Wood, Locus Robotics Chief Strategy Officer Gina Chung, Prologis Vice President of Property Management Vince Zuppa, and GS Futures Managing Partner Aaron Toppston.

Others presenting included John Ramirez, the founding partner of ForgeX at TechNexus, which builds commercialization pathways for frontier technologies at the intersection of emerging industry demand and frontier research, and Ed Briganti, Managing Director, Wells Fargo Corporate & Investment Bank.

The conversations all had one clear message: physical AI is no longer just a frontier technology. It is beginning to show up in the real environments where industrial work happens — warehouses, construction sites, manufacturing floors, logistics networks and heavy equipment operations.

Practicality is Beating the Hype

Commercial success will depend more on practicality than technology alone. The companies most likely to win are not simply pitching robotics to their clients, but solving specific pain points with clear ROI.

At Locus Robotics, Gina Chung is seeing 17,000 robots moving swiftly through warehouses every day across more than 360 warehouses in 20 countries. At Prologis, an increasing number of customers are looking for buildings ready to support automation and future technology integration, said Vince Zuppa. At Nine-Four Ventures, Kurt Ramirez saw robotics adoption in narrow, repeatable tasks such as drilling holes and tracking jobsite progress, where speed, safety and schedule compression can deliver measurable ROI.

Real-World Data Becomes the Advantage

These deployments also generate something even more valuable: real-world operating data. Physical AI demands more than digital AI at the orchestration level because companies are managing physical fleets, said HPA’s Pete Wilkins. That process depends on movement, interaction and machine behavior in live environments.

CAT Digital’s Charlie Wood emphasized that a lot of hard foundational work is required before companies even get to build usable models. That includes locating the right data, making sure it is high quality, assigning clear ownership and cleaning it so it is ready for operational systems.

That data is not just operationally useful — it may become the defining competitive asset in physical AI. Unlike generative AI, which is trained on text and images, physical AI models are trained on movement, interaction, and real-world task completion data. Companies that deploy at scale accumulate proprietary datasets that are difficult to replicate. The result is a compounding flywheel: more deployments produce better models, better models drive better performance, and better performance enables more deployments. The companies that move first and accumulate that data at scale may be the hardest to displace.

Trust Comes First

But strong data foundations alone do not guarantee adoption. For founders and investors, customer curiosity is not enough. They want buyers that are ready to commit and put the technology into operation. However, there is still “a huge gap” between “I want automation” and “I’ve got a signed contract,” said Aaron Toppston of GS Futures.

That gap is why early adoption is still happening in lower-risk environments where companies can test and build confidence. As deployments mature, the challenge will shift toward scaling into more complex use cases while solving practical obstacles such as integration, component availability and battery life.

Many of the systems and technologies needed to advance robotics in the near term have already been invented; what comes next is “operationalizing them and actually seeing the deployments,” said Jason Felger, partner at Jump Capital.

Capital Is Moving, But Founders Must Earn It

The financing environment surrounding physical AI is active. Panelists and investors described an AI-driven capital supercycle, with significant funding flowing into the segment and large deals returning after a period of caution. For companies building the enabling infrastructure — the picks and shovels of the physical AI stack — opportunity exists alongside the robotics companies themselves.

But access to that capital is selective. Investors are not writing checks on vision alone. The signals they look for include contracted backlog and signed customer commitments, a team with real hardware experience, and a clear, defensible hypothesis on ROI. Founders were also reminded that hardware companies operate under a fundamentally different clock than software: products cannot be iterated and shipped overnight. Investors expect founders to acknowledge that constraint explicitly and demonstrate a pipeline matched to their manufacturing capacity.

Non-dilutive capital — from customers, strategic partners, and grant programs — was raised as an underutilized path, particularly for companies at early scale. Founders who think carefully about the right capital source at each stage of development, rather than defaulting to venture, are likely to find more options available than they expect.

The machines are ready. The data is accumulating. The capital is moving. What remains is the will to commit — and that, more than any technical breakthrough, is what will separate the regions and companies that lead in physical AI from those that watch from the sidelines. The Midwest has spent decades building, operating and refining the exact industries physical AI is designed to transform. That history is a head start — and this region intends to use it.

By Alex Chen at TechNexus Venture Collaborative