Intro
Sometime between 2013 and 2017, a small gaggle of startups attempted to do for valet parking what Uber had done for rides: tap on your phone, summon a uniformed attendant to any street curb, and walk away from your car without having to think about where it was parked. By the end of 2017, that entire category was effectively dead as a consumer business.
The names are hard to remember now, but they raised real dollars and worked in real cities. Luxe Valet, founded in San Francisco in 2013, operated in six metros at its peak. Zirx came next, in 2014. It raised about $36M before shuttering its consumer service in April 2016. Vatler, ValetAnywhere, Carbon, FlightCar were all peers in this second wave, each chasing variations on the same concept. Each defunct or pivoted within a few years.
This is for the people who have mourned the loss of those apps and wondered what happened and for anyone trying to make sense of how a category that made so much sense in a pitch deck can crash up against a wall in operations. The short answer is not “they ran out of money” which is what gets reported when a startup dies but almost never is the proximate cause of death. The deeper answer is unit economics and a category mismatch with the rideshare playbook the founders were cribbing from. A valet attendant has to park a car and come back later to retrieve it which is structurally different work than a driver doing one complete trip. That difference doesn’t show up in a pitch deck and it almost guarantees the consumer-app model never finds positive margins at the price the consumer will pay.
Below is a deep dive on that math, the ride-hail analogy that broke against the force of reality, the brand-by-brand trajectory of the cohort, and what took root after the apps left the building.
Why the on-demand valet model never made the math work
A rideshare driver completes one trip per fare: one pickup, one drop-off, paid for one motion. An on-demand valet attendant performs two: take the car when the customer arrives, retrieve and return it when the customer leaves. That doubling is the entire problem in one sentence, and it shows up nowhere in the pitch deck the founders were pitching from.
Think about the per session cost floor. Say a customer drops off the car at 7pm and picks it up at 10pm. The attendant does the park motion at 7pm and the retrieve motion at 10pm and is on the clock for some fraction of those three hours no matter how many other customers are served in between. At $20 to $25 per hour fully loaded for labor (very modest for an urban market) the per session labor cost ends up somewhere in the $15 to $30 range before the company has paid for the parking lot, the insurance, the app or the dispatcher. Consumers who had no problem paying $40 for an Uber ride home from dinner balked at paying $30 to "park their own car for them" because in the consumer's mind that is what the service is.
The metric the founders focused on was sessions per day per attendant. That metric seemed OK: an attendant could have a reasonable number of sessions, say 4-6, in an evening shift. The metric hid the fact that the labor was doubled. The right unit is labor-hours per session, and it was structurally more onerous than rideshare because of the retrieve leg.
The second unseen expense: dead time. Between the 7pm drop and the 10pm pickup, the attendant was either taking care of a customer (good) or loitering at a corner waiting for the next ping (paid, but not productive). A rideshare driver between fares can roam toward anticipated demand and almost always has the next trip ready within minutes. Valet attendants between drop-offs are fixed to a small geographic radius, defined by where the customer's car was parked because they need to be able to get back to it. That radius makes the next-customer search orders of magnitude harder.
The rideshare analogy that broke the founders
Luxe and Zirx both sold to investors with some variant of the same line: Uber for parking. The framing was elegant, it got the rounds closed, and it survived almost zero contact with operations. Three breaks killed the analogy.
The first was supply density. A rideshare driver can route to any pickup within a few miles in a dense city, because the driver brings the car and the car defines the service radius. An on-demand valet attendant needs to be within minutes of the customer AND within minutes of a parking lot that has open spaces, AND the lot has to be cheap enough that the company can still make money paying for the spot. Stack those constraints in a single neighborhood and the geography that works is much smaller than what rideshare needs. Density requirements in any city under SF or Manhattan were brutal.
The second break was capital and liability. Rideshare drivers provide their own car as the working capital of the business. The platform owed the driver a cut of the fare and that was the end of the platform's capital exposure. On-demand valet flipped that. The platform supplied the labor and supplied the lot. The customer supplied the car, which became the platform's liability the moment the attendant slid behind the wheel. A scratched bumper or a dented door was a claim the platform owned. Rideshare never carried that risk.
The third break, the one founders seldom discuss, was utilization. Rideshare drivers between rides have an unlimited geography to wander through; the next pickup will pop up somewhere within convenient distance. Valet attendants between requests are anchored to a small radius by the customer car they are charged with fetching. The implied utilization rate the model needed to be profitable was much higher than the geography could support. Founders designed rideshare-style utilization and got operations that looked like rideshare on paper but which created rideshare-shaped losses without rideshare-shaped scale paths.
The arc, 2013 to 2017
Five inflection points define the consumer on-demand valet category from launch to category death.
2013, the founding. Curtis Lee and Craig Martin launch Luxe Valet in San Francisco. The thesis was simple: valet doesn't have to be confined to a restaurant or hotel. Anyone with a phone should be able to hail an attendant to any curb. The product launched in SF as a first market and hired W-2 attendants in uniform vests to give the operation a sheen of professionalism that gig-economy parking gigs had not.
**2014 and 2015, the wave.** Sean Behr launches Zirx out of San Francisco in 2014, raising at premium valuations on the same on-demand thesis. Vatler, ValetAnywhere, Carbon and FlightCar all show up in the same space in about the same 18 month window. By the end of 2015 Luxe had expanded into Los Angeles, Seattle, Chicago, Austin and Boston. Zirx had a multi-city footprint of their own. The category looked like a real race.
April 2016, the pivot signal. Zirx abandons the consumer on-demand valet model in public and repositions to B2B parking and mobility service for office buildings and corporate campuses. The signal was unmissable to anyone paying attention: a well funded player in the category had just told the market the consumer unit economics did not work. The other players kept going, but the pivot reframed what was possible.
August 2017, the acquisition. Volvo Cars announces the acquisition of Luxe. Price not disclosed. The team rolled into Volvo's mobility services group, instead of continuing to operate the Luxe consumer service as a standalone. The consumer service was shut down within months.
**Late 2017, category death.** By the end of 2017, there were no consumer-facing players left in the consumer on-demand valet category. Founders of some companies went on to found other companies. Assets and trademarks from other companies were acquired and quietly retired. The window had been, roughly, four years, peak to trough.
The cohort, brand by brand
It wasn't one startup story. Six names defined the wave. The two best-funded went furthest into the consumer market before the math caught up.
Luxe Valet
Founded in 2013 in San Francisco by Curtis Lee and Craig Martin. Raised approximately $25M total, with a Series A led by Venrock and Hummer Winblad also participating. At peak, the consumer service was present in SF, Los Angeles, Seattle, Chicago, Austin, and Boston. Volvo Cars announced the acquisition in August 2017, the team joined Volvo's mobility services group, and the consumer facing service was gradually wound down later that year. The acquisition price was not publicly disclosed, nor were exact revenue or employee numbers at the time of the deal.
Zirx
Founded in San Francisco in 2014. CEO Sean Behr. Raised around $36M in funding including a Series B from Bessemer Venture Partners. The company produced the same consumer-app driven service model as Luxe: tap to summon an attendant to drive to any curb. Zirx publicly abandoned the consumer product model in April 2016, repositioning the product as a B2B enterprise parking and mobility service for office buildings and corporate campuses. The company changed its name to STRATIM in 2017, to better match the brand with the B2B product pivot. STRATIM was acquired by KAR Auction Services in 2018. Zirx's consumer product (the one that was Luxe's direct competitor), was shut down in April 2016.
Vatler, ValetAnywhere, Carbon, FlightCar
Four other related on-demand-parking startups, also from the same 2014 to 2018 window that either never launched or could not sustain themselves as independent consumer businesses. Less reporting on each than on Luxe and Zirx, for reasons stated. Each did something very similar (summon a service to your curb via an app) and each encountered a variant of the same fundamental unit-economics wall that broke both the better-funded ones and the lesser. Some changed their approach, some quietly shut down, some were acquired into adjacent companies. FlightCar in particular had more to do with peer-to-peer airport-parking rentals than summoning a valet via app, but was in the same investor mind-space at the time and had the same outcome. The pattern repeated across this cohort, more than any individual tale, is what makes this a category death, not a string of unrelated failures.
What the apps were actually trying to replace
Valet, prior to 2013 in the United States, was one of three things, none of them very consumer-friendly. It was hotel-bound, the porte cochère, the bell stand, the guy in a vest who took your car and gave you a stub, existing because the hotel needed it, rarely chosen by the consumer. It was restaurant-contracted, an operator hired by a single restaurant, or a grouping of restaurants, for a busy stretch of the week, again chosen by the venue and inherited by the guest. It was event-specific: weddings, galas, fundraisers, holiday parties, with a planner or host phoning a local operator and booking a crew for the night.
Booking valet outside of one of those three contexts (say, a private dinner party at a home, or a small business celebrating an anniversary) meant cold-calling local operators, asking about minimums, comparing hourly rates, and trusting that whoever showed up would be reliable. The friction was real. The apps were attacking that friction. The pitch to consumers (open the app, get a uniformed attendant in minutes) and the pitch to investors (formalize an informal gig market) were both reasonable on their face.
What the apps got wrong was to assume that the friction was the product. The friction was actually a price signal. Phone-based, contract-based valet was friction-laden because the underlying labor was expensive, and demand was spiky, so operators rationed access to it. Wrap an app around the same labor and the same demand and the labor cost does not go away; it just gets paid by a venture round instead of the consumer. The pre-2013 system was less elegant, but it was honest about the math.
What filled the gap
The consumer-app category died. The underlying demand (people who wanted to book valet for a one-off event, a private dinner, a small business gathering, a wedding) did not. What replaced the apps was older and slower and structurally honest: a directory model and the local operators who had been there the whole time.
Directories take the friction the apps were chasing and solve it without trying to absorb the labor cost. A national list of operators, sorted by city, with quote routing for one-off events, allows a host to compare crews and rates the way a 2013 host would have done by phone, just faster and with more options visible at once. The labor cost still gets paid by the customer, because that is where it has to live for the operator to stay in business. The friction comes down because discovery comes up, not because anyone is hiding the labor.
Operator-side aggregation did the other half of the work. Decades-old local valet companies absorbed the demand the apps had been after, in most cases without changing how they staffed, dispatched, or priced their service. The thing the apps were trying to formalize wasn't a new market. It was an existing one that needed a better front door.
If you're here by searching for how to book a valet on demand with a broken app, the real world follow-up post in this series is the next one: On-demand valet is gone, how to book valet in 2026, which describes the live options and how to compare them.
Frequently asked
Why did Luxe Valet shut down
Luxe Valet was sold to Volvo Cars in August 2017 and staff joined Volvo's mobility services group. The consumer-facing offering was shut down later that year. The fundamental problem keeping the consumer business from operating on its own was unit economics: a valet attendant has to both park and later retrieve every car, so the labor is twice the amount per session compared to a rideshare driver who only does one trip per fare. That math busted the consumer price point the app was aiming for.
Is Zirx still around
Zirx is not the name of a consumer valet app. Zirx turned its focus from a consumer on-demand product to B2B enterprise parking and mobility services for office buildings and corporate campuses in April 2016. The company rebranded itself STRATIM in 2017 to more closely match its B2B focus. KAR Auction Services purchased STRATIM in 2018. The Zirx product that was a direct consumer competitor to Luxe was eliminated at the April 2016 pivot.
Is there a current on-demand valet app
No consumer on-demand valet app has since achieved the scale that Luxe and Zirx had in 2015-2016. By late 2017, the category effectively died as a consumer business. Booking valet today is done via a different model: directory listings of local operators with quote routing for one-off events, plus the long-standing contracted-operator relationships that venues and event planners have always used.
What replaced Luxe and Zirx
Consumer apps got supplanted by a directory plus local-operator model. Years- or decades-old local valet companies soaked up the demand the apps were chasing. Directories accelerated discovery by listing operators by city, and routing one-off event quotes to many providers simultaneously. The labor cost (paid by the customer) remained where it needed to be. But friction in finding an operator plummeted.
Why didn't on-demand valet scale like Uber did
Rideshare drivers supplied their own car as business' working capital, and did one trip per fare. Valet attendants parked and later retrieved each car, double labor per session. The platform also assumed liability for customer's car at moment attendant slid behind wheel, something rideshare never carries. Density reqs were tighter too, attendants needed both customer within minutes and cheap parking lot within minutes. Analog looked clean in pitch deck and broke in operations.
Did any on-demand valet startup actually succeed
No consumer on-demand valet startup became a sustainable standalone business. Zirx went B2B in 2016, and was eventually acquired after its pivot. Luxe was acquired by Volvo Cars in August 2017, and its consumer service was shut down. Vatler, ValetAnywhere, Carbon and FlightCar did not survive as independent consumer businesses. The trajectory was the same for the entire cohort: strong early traction, structural inability to make the unit economics work at consumer price points, exit/shutdown by late 2017.
Can I still book a valet on demand today
Yes, but not through a consumer-facing app. The only realistic option is a directory of local operators by city that routes one-off quote requests to multiple providers, exactly the way a host or planner would have booked valet by phone before the apps came along (just with less friction). For a step-by-step walkthrough of the live options, check out On-demand valet is gone, how to book valet in 2026
What cities did Luxe Valet operate in
Luxe Valet grew to San Francisco, Los Angeles, Seattle, Chicago, Austin, and Boston at its height. San Francisco was the first market entered, and remained the deepest deployment until acquisition. Adding each additional metro (more attendants, more parking-lot contracts, more dispatch coordination) provided only diminishing revenue lift relative to the deepening unit-economics problem that had never been solved for the category.
*Editorial review: Roy Nickolai of All About Parking (https://allaboutparking.com).*