Deep Dive: Metropolis

How a computer vision startup became America’s largest parking operator and a test case for a new model of venture-backed growth

Deep Dive: Metropolis

As we wrote last week, more software companies are choosing not just to sell software, but to vertically integrate: building or buying the operating businesses their software powers, rather than staying content as a vendor selling into someone else’s operation.

Metropolis is perhaps the best example of this trend. Metropolis acquired its first business, a parking operator, the ninth largest in the US, in 2022. In 2024, it raised $1.8 billion, the majority of which was used to acquire the publicly traded, 100-plus-year-old parking operator SP+. Metropolis is now the largest parking operator in the US.

Metropolis ties into a lot of trends we cover at Thesis Driven. Its value-add is powered by computer vision technology, and it’s emerging as an operator improving outcomes in an oft-ignored real estate niche: the $131 billion parking industry. But parking as a category faces a unique set of challenges with the rise of electric and fully autonomous vehicles, putting the future of these physical assets themselves into question.

We’ll explore all of that through the rise of Metropolis in today’s deep dive.

Having the (Computer) Vision

A few months ago, we wrote about computer vision as an underexplored but high-potential domain of AI. Computer vision gets a lot less love than large language models, mostly because it’s harder to play with: you can’t organize your shopping list or summarize a business book with frontier computer vision tech. And unlike isolated LLM point solutions, computer vision-powered tools tend to get packaged into vertical software rather than sold as standalone products.

But computer vision promises to be just as impactful on real estate as any LLM, if not more, precisely because real estate is a business built on looking at physical stuff and making judgment calls about it. That’s the kind of pattern recognition computer vision is built for, and we’re already seeing it show up in development (OpenSpace) and in acquisitions like FoxyAI and Restb.ai, both of which are gaining real traction and delivering measurable value.

Metropolis co-founder and CEO Alex Israel saw the opportunity to apply computer vision to mobility. “We looked at companies like Amazon Go, and how they were implementing just-walk-out technology,” said Israel. “And we realized that there was a real opportunity to look at cities, and mobility specifically, about how people transact in a vehicle.” In other words: if a customer could just walk out of an Amazon Go store without stopping by the register to pay, why can’t someone simply leave a parking garage without futzing with passes, tickets, and pay stations?

“We leverage artificial intelligence and computer vision to capture a vehicle, to create a fingerprint of a vehicle, and then, when a user engages in our experiences, to send them a text message when they arrive and to seamlessly charge them when they leave,” Israel explains.

The technology goes beyond simple license plate recognition. “You can take your license plate off, and we still know it’s you. The unique qualities of a vehicle are that of the unique qualities of a fingerprint. The scratches, the bumper sticker. The license plate plays a part, but that’s just one heuristic component of the model.”

Metropolis claims a 20% NOI bump in parking revenues. At Westwood Marketplace in Los Angeles, for instance, that uplift held. But in contrast to many real estate technology companies promising on-site staffing reductions through automation, Metropolis focuses on revenue gains rather than expense savings.

Owners work with Metropolis one of two ways. In a management contract, Metropolis takes over operations entirely, staffing, technology, customer experience, and shares revenue with the owner, tying Metropolis's upside directly to the asset's performance. In a lease deal, Metropolis pays the owner fixed rent in exchange for full operational control, giving the owner predictable income while Metropolis captures whatever upside its technology generates.

“We have fans, we have loyalists,” explains Israel. “People look for that Metropolis logo. They look for where a Metropolis location is, as juxtaposed to the static parking operator across the street.” The company’s technology has improved the parking so much that people seek out Metropolis-branded garages, letting those garages push occupancy and rates higher than the lot next door.

Today, Metropolis boasts 30 million members. Central business districts often have multiple parking garages within a few blocks of each other, and it’s easy to imagine a driver picking one over another for reasons that have nothing to do with price, particularly if they’re new to a city or driving a rental car with unfamiliar dimensions and would rather not angle up to a pay station and fumble with a credit card. Find a Metropolis garage instead, and you just drive in and out. For a lot of people, that convenience is worth another few dollars.

But the best technology in the world doesn’t matter if real estate operators aren’t willing to adopt it. And parking has some peculiar traits that make novel technology adoption a particularly tough sell.

A Concrete Problem

Parking is unique among real estate assets. For one, there’s a lot of it: 24,000 square miles by some estimates, enough to cover the state of West Virginia. More than ten times as many American square feet are dedicated to parking as to office and multifamily combined.

Yet relatively few investors look at parking as an asset in its own right. While there are tens of thousands of parking garages, they’re typically viewed as appendages to other asset classes (office or multifamily, say), not as assets that stand alone. Not many institutional asset managers have dedicated parking sleeves.

As an asset management afterthought, parking has been relatively untouched by technological innovation, even as AI has played a growing role in real estate operations in recent years. “[Parking] hasn’t evolved in a hundred years,” explained Israel. “It was the last bastion of non-institutionalized real estate.”

“You go to every major real estate investor and they have a Class A office fund. They have a residential fund. They have an industrial fund. None of them have a parking fund. It’s forgotten. It’s this dirty underbelly that hasn’t really evolved.”

But the same neglect that made garage owners sluggish adopters of technology also made them poor targets for a typical SaaS sales funnel. Parking managers on their own were uninterested in new technology, and the asset managers who usually push efficiency initiatives through their portfolios treated parking as an appendix bolted onto whatever office or multifamily asset it happened to sit beneath.

“The original thesis was that we were going to grow just like any other SaaS AI company in the space,” said Israel. “But we quickly realized that the market was fragmented, and the operators didn’t have the technical know-how or the willingness to deploy technology at scale. So we realized we had to vertically integrate. We had to become a parking operator. We had to eat our own dog food, if you will.”

Rather than selling SaaS into existing parking management companies, Metropolis would buy them outright and use its own technology to boost owner returns and win deals.

“At first investors were uncomfortable. There was trepidation, there was hesitation, that we were going to engage in M&A risk, that there was going to be technological risk. This idea of marrying an AI company with an old-world operating business that had been around for a hundred years was absolutely unheard of.”

But buying a parking operator in 2022 gave Metropolis the best path to prove out its technology: more control over the customer experience, and a sandbox to validate the model without negotiating every operational change with a third party. That let the company demonstrate the relationship between technology and demand, and track a reduction in leakage, meaning people leaving a garage without paying, which can run above 10% at some properties.

If Metropolis was dabbling in M&A in 2022, it put rocket fuel under the strategy in 2024 after its $1.8 billion raise in fresh capital from investors including Eldridge and BTD/MSD Partners. It used that capital, in part, to acquire parking operator SP+. Today, the combined company operates more than 4,500 locations.

But Metropolis is looking beyond purely inorganic growth, investing in third-party parking management services offered directly to owners, as well as potentially launching its own PropCo. And the company has also launched plans to expand use of its recognition technology into hospitality and quick-service restaurant contexts.

“How do we scale from 6% to 12% of the market?” asked Israel. “Independent of just our organic growth, we are starting to look at what it would look like to spin up a PropCo to actually own the underlying real estate itself.”

Parking’s Future

Many real estate investors are skeptical of parking’s long-term potential given the rise of autonomous vehicles. In a world where a sizable share of the vehicle fleet drives itself, how much parking is still required, and where?

Israel doesn’t deny technology’s potential to change the nature of parking; after all, betting on technological transformation is a centerpiece of Metropolis’s business. Rather, the company is looking to help owners shepherd their garages through the changes to come.

“Parking right now represents 15 to 30% of the surface area of our cities,” said Israel. “We want to future-proof these assets. How do they convert from parking lot today to mobility hub in the future?”

“These garages should enable the cleaning, servicing, charging, and deploying of all future forms of mobility, whether that’s a vertical takeoff and landing vehicle, whether that’s a drone, whether it’s a droid, or whether it’s an autonomous vehicle.”

Israel noted that Metropolis has signed partnerships with “all the major AV and eVTOL companies” to provide technology, operations and infrastructure. This is a deliberate decision given where parking is likely headed. The rise of electric vehicles means charging infrastructure is a real need right now, but the requirements of autonomous fleets will likely go well beyond that, folding cleaning and basic maintenance into parking infrastructure too.

Macro shifts aside, Metropolis’s model reflects where venture seems to be headed, with companies ranging from tech-forward architecture firms like Cove to permitting software company Pulley taking a similarly vertically integrated approach. Some, like owner’s rep firm Zero RFI, are prioritizing growth through acquisitions of their own.

“This idea of marrying artificial intelligence and SaaS companies with old-world services businesses is definitely the future of venture,” noted Israel. “You’re going to see over the next five, 10, 15 years, the growth buyout with a technology company buying an old-world business become a dominant thesis within venture.”

“These businesses are sticky, they’re unique, and they’re very difficult to disrupt.”

Whether that thesis holds for parking specifically depends on a bet most institutional owners have never had to make: that a century-old asset class nobody bothered to operate well is actually worth operating well. Metropolis is the biggest live test of that bet, and its results over the next few years will tell the rest of the industry whether the dirty underbelly of real estate was hiding a real business all along.

—Brad Hargreaves

 

Great! You’ve successfully signed up.

Welcome back! You've successfully signed in.

You've successfully subscribed to Thesis Driven.

Success! Check your email for magic link to sign-in.

Success! Your billing info has been updated.

Your billing was not updated.