Silicon Valley’s $445 million robot pizza revolution that wasn’t

“What if there was a better way to make and deliver pizza to you?” asked CNBC’s Jim Cramer—shirtsleeves rolled up, wagging his finger at the camera—as he passionately set up a segment of his popular Mad Money show in June 2018. “Turns out, there is, which brings me to Zume, Inc., a Silicon Valley–based startup trying to bring this industry to the modern era.” Cramer, on location in Mountain View, California, and standing in front of a shiny red food truck, quickly introduced Zume’s robots, cuing up footage of a robotic arm balletically moving pizzas in and out of ovens inside a warehouse. This, he intoned, was a revolution. “We automate everything else, why not pizza?”

When Zume cofounder Alex Garden stepped into the frame, the frenetic energy spiked further. “If you think about the huge explosion in scenarios that food companies are trying to address today,” Garden said, meaning improving margins, finding workers, and meeting consumer demand for the convenience of delivery and healthier food, “this”—Garden pointed at the vehicle—“is the key to making it all work.”

The truck behind him was tricked out with custom refrigerators, ovens, and other kitchenware that let it cook food on the go. “This is one of the most complicated things anyone’s ever attempted before,” Garden proudly explained, ticking off on his fingers all of the tech that Zume was deploying. “We’ve got artificial intelligence, we’ve got logistics networks, we’ve got positioning, robotics, automation, I could go on.”

He went on. Zume made pizza with robots in a warehouse and then finished it in these kitchens on wheels. It served three markets near San Francisco, but Garden had good news for everyone else: Zume had become a platform. “All the technology we built for Zume Pizza is now available” to any dining chain, convenience store, or grocer in the country that wanted its own roving restaurant.

“Wow!” Cramer exclaimed, going in for a hearty handshake just before commercial. “You have got one great idea.” He stopped to correct himself. “It’s not an idea, it’s a business.”

In June 2023, five years after the CNBC segment almost to the day, Zume was liquidated. The company, which raised approximately $445 million in its lifetime, is one of Silicon Valley’s most embarrassing failures.

Cause of death? Well, for one, too many ideas and not enough business.

Zume did not go under because of a huge scandal. It never committed the kind of fraud that sent Theranos founder Elizabeth Holmes to prison. Zume is a sadder kind of failure: Hundreds of accomplished people spent years trying as hard as possible to realize Zume’s world-changing vision. The company’s flameout was so ignominious for those who experienced it that neither Garden, cofounder Julia Collins, nor other senior executives agreed to speak with Fast Company for this story. Those employees who did asked to do so anonymously.

Being Silicon Valley’s overfunded robot pizza startup is a terrible legacy. Bot jokes aside, many of Zume’s ideas were good: There is a role for tech in routinizing drudge work and serving customers better. Most of its technology worked, and there are tech companies and restaurant chains today pursuing the ideas that Zume glommed together in its effort to present itself as the “Amazon of food.” The analogy deliberately evoked a business so complex that it touched every part of food production and distribution. But Zume’s baroqueness—in its engineering, its financing, and its founder’s hype—doomed it.

It’s a story perfectly tailored for its time, one that begs the question: Why is the tech industry so obsessed with pizza?

Garden first got interested in what eventually became Zume after sitting next to a Pizza Hut franchisee on an airplane. “I thought, This person isn’t particularly competitive, and maybe I should learn more about pizza.”

Step one: Stake a claim. In 2013, Garden, who started a video-game business at 15 years old and was later an Xbox executive, filed a patent for cooking food as it was being delivered. As he developed what would become Zume’s “bake-on-the-way” process, he came to believe that automating pizza-making itself—using robots to construct the pies—would cut production costs. Better margins meant more money for high-quality ingredients and “better jobs”—not the sort of dull, repetitive tasks common in food service. By reinventing the entire process, Garden reasoned, he could beat the corporate pizza chains at their game. Garden, though, had no experience in food, restaurants, or robotics.

He met Collins, who would become his cofounder, through a mutual friend. The Harvard grad and Stanford MBA was working in New York City, where she had helped launch and run Mexicue, a Mexican–barbecue mashup restaurant that started as a food truck and grew to multiple locations in two states. But Collins, the daughter of prominent San Francisco real estate developers credited with helping to revive the city’s historically Black Fillmore District, was looking to move home.

Garden and Collins reputedly bonded over two things that would later become the company’s founding principles: Serving food to people is a sacred trust and Every American has a right to delicious, affordable, healthy food. When the pair launched Zume Pizza in the summer of 2015, they printed the words onto a giant poster and hung it on the kitchen wall. (There’s no indication that a third value, “no assholes,” which Collins shared in a 2016 interview, ended up on any wall.) But while Collins would often refer to these noble precepts to explain Zume’s existence, Garden, by contrast, liked to push what he believed was their flashy and groundbreaking technology.

Starting around 2015, both tech and finance types loved noting that a $1,000 investment in Domino’s at the time of its 2004 IPO would then have made an investor more cash than the same one in Google, which went public a few weeks later. Zume Pizza, with its promise that its tech could make healthier, cheaper pizza than Domino’s and deliver it better, too, made an appealing bet.

Both Collins and Garden spoke about Zume Pizza’s promise in the clichéd techno-optimism of the mid-2010s. “I want to feed the world and save the planet, a vision that big,” Collins told one interviewer. “That got people’s attention more so than saying, ‘I want to launch a new logistical platform for delivering food to customers.’”

So did putting the words pizza and robots in the same sentence. The novelty of food-service bots delighted VCs and the media. The investor who led Zume’s seed round praised Garden as an “autodidact,” adding that Collins’s restaurant experience sealed the deal. When TechCrunch toured Zume’s offices in 2016, it dubbed Collins’s approach to pizza as “Elon Musk–esque” and declared the company a “quintessential Silicon Valley startup” because it had a Juicero—the $700 juice-packet press—“front and center.”

Inside the kitchen, most of the robot pizzaiolos were given Italian first names. They worked an assembly line next to human colleagues who stepped in to do tasks the robots couldn’t—at least not yet. A “doughbot” flattened the dough into circles; a conveyor belt then carried it to Pepe and Giorgio, stainless steel sauce dispensers outfitted with lasers that dropped a precise amount of sauce onto dough. The conveyor belt passed under Marta, a robotic arm equipped with a tool that mimicked a ladle and spread the sauce evenly. Humans added cheese and other toppings, and then Bruno, a $100,000 giant industrial arm with a tiny conveyor on the end built by Swedish robotics company ABB, moved pies from the assembly line to an 800-degree oven. Finally, Leonardo cut the pizza into slices.

Collins hired a chef from the trendy Brooklyn pizzeria Roberta’s to lead recipe development, and the first pizzas Zume served to customers rolled off the line in April 2016. Based on early Yelp reviews, people either loved the experience—or they absolutely hated it. Fans celebrated the robots and stunningly fast 15- or 20-minute deliveries. Detractors complained about bland toppings and soggy crusts. Some balked at downloading an app or creating an account just to order dinner, a practice that DoorDash and others later normalized.

By the fall, Zume was ready to deploy its bake-on-the-way delivery trucks, dubbed “mobile kitchens.” Each $200,000 “restaurant on a rubber foundation,” as Garden once described them, was filled with equipment intended to optimize every second of a pizza delivery. Zume’s software worked to predict pizza demand and deploy trucks to strategic locations. The vehicles were loaded with partially baked pies that had been prepared on the assembly line back in Mountain View. Inside the mobile kitchen, any of 56 ovens could fire those pizzas when the truck was four and a half minutes from a customer’s home.

After clearing a number of regulatory hurdles, Zume was the first in the world with a truck capable of cooking food while in transit. In Garden’s view, everything anyone knew about pizza delivery had to be rethought, which led to Zume’s operations multiplying in complexity. Consider the mobile kitchen’s custom slicer. “I wanted to avoid having a three-compartment sink because it takes up a lot of room,” Garden said in 2017. That meant no small utensils. “In order to do that, you had to make the pizzas eject out of the oven into a package, and you had to cut them in the box with a special pizza cutter that we designed, that cleans itself after every cut.”

The traditional cardboard pizza box also wouldn’t do. “We’re a startup, no one can tell us what the rules are,” Garden said after being dissatisfied with existing offerings. “Why don’t we just get a whiteboard and draw what a cool pizza box would be?” Zume designed its own fully compostable, round packaging made from sugarcane fiber.

The trucks could serve a jaw-dropping 120 pizzas per hour, though they didn’t necessarily get the chance. At the time, Zume delivered around 250 pizzas per day.

The mobile kitchens quickly ran into an immutable reality: gravity. Hot sauce, melting cheese, and the Bay Area’s hills didn’t mix; cheese would slide off Zume’s pizzas as they baked en route. It was a small but laughable hiccup. Instead, Zume parked its mobile kitchens in central locations and a fleet of branded Fiats and scooters ferried pizzas to customers.

This wasn’t exactly the disruptive, planet-saving future Zume had promised. As one former employee notes acidly, “Being such a sustainable company, the trucks were running off gas, parked on the side of the street, which does not make sense at all, right?”

Whatever Zume’s vehicles lacked in function, they made up for in marketing. Garden never passed up an opportunity to create buzz by strategically deploying a mobile kitchen to wherever Valley cognoscenti might be. Once, Garden sent a truck to impress a prominent VC during their poker game in tony Atherton, California, a wealthy Silicon Valley enclave 30 miles south of San Francisco. “This is one of those infamous Silicon Valley stories,” Garden told the investor and podcaster Jason Calacanis on stage in 2017 with two other food robotics startups.

The story actually didn’t appear to be all that notorious. But the industry loved its bad boys, from Musk to Uber’s Travis Kalanick, so Garden was working hard to vault Zume Pizza into infamy.

Pizza, Zume’s founders insisted, was just its proof of concept. “Pizza is to Zume as books were to Amazon,” Collins explained. After all, pizza had long been associated in Silicon Valley with being a great way to mainstream a novel technology. Pizza was the first physical good ordered on the internet in 1994 and the first commercial transaction paid in Bitcoin in 2010.

After Zume raised $48 million in late 2017, the startup’s ambition rose with its valuation ($170 million). Collins and Garden officially shortened the company name from Zume Pizza to Zume, befitting their desire to sell their technology—and trucks—to any food business.

Only one investor—Masayoshi Son, CEO of SoftBank—could now give Zume the capital it needed to become the Amazon of food. Son had launched his almost $100 billion Vision Fund in 2016, pumping unprecedented sums into Uber, WeWork, DoorDash, and dozens more tech companies that he believed could accelerate AI, robotics, and the transformation of everyday life. Zume thought it was a perfect fit.

According to a former employee, Zume tailored a pitch deck for Son, laying out a six-part business model that would link everything from agriculture to transportation to electric vehicle charging to mobile kitchens to ordering to packaging. The vision: “one elegant system that combines everything in the food chain together,” they recall. Garden fixated on the idea of limitless opportunity but left the interpretation and execution to everyone else.

To deliver the pitch, Garden again deployed a pizza truck—with him aboard—to a strategic destination: Son’s front door.

Garden arrived at Son’s $117 million, nine-acre estate in Woodside (20 minutes from Zume HQ), and incredulously, the gate was open. More remarkably, Son was home and agreed to a demo. Whether impressed by the truck, Garden’s tenacity, or both, Son committed $375 million, promising another $375 million later. If Zume received the full sum, it would be valued at $2.5 billion.

This made Collins the first Black woman to have cofounded a unicorn company. “I cracked that ceiling because I have no interest in being the ‘only,’” she would later say. But Collins left Zume in November 2018, the same month the money arrived. She described it to Fast Company in 2022 as “a heartbreaking experience deciding to leave something that was so tied to my sense of purpose and identity.”

Collins’s language was gracious, but a former Zume Pizza employee explained the split in harsher terms: “It did not end particularly well between Julia and Alex.”

Garden, now CEO, had won the big check that legitimized his ideas. Or, as the employee put it, “Alex saw the SoftBank funding as a mandate to go nuts.”

NO BAD IDEAS: ZUME’S FOOD-TECH CONCEPTS WERE MOSTLY LEGIT. A HYPE-FREE GUIDE TO WHO’S PURSUING THEM NOW

As much as Garden wanted Zume to break free from pizza, it couldn’t. Within a week of the cash infusion, Garden had dinner with the leadership team of Pivot Packaging, a young Southern California company that Zume had been working with to manufacture its distinctive pizza boxes. He floated buying Pivot. Pizza Hut, which had approximately 18,000 global stores, wanted to test Zume’s molded-fiber packaging, and Garden needed to scale up production, fast. In February 2019, Zume bought Pivot for $20 million, half cash, half stock. It would run its packaging operation out of Pivot’s plant in Camarillo, California, a town of 70,000 in Ventura County, east of Los Angeles.

Garden, obsessed with finding partners he believed could operate at “Silicon Valley speed,” expected to move quickly into food service and restaurants. But emails between Zume and Pivot executives days after the acquisition closed underscored an already strained relationship. “If we whiff on March, the board will have my ass,” Garden wrote Scott Lilley, Pivot’s president and brand new Zume employee, just before boarding a plane in Tokyo. He had dinner plans with Pizza Hut’s CEO for the evening he returned. But manufacturing couldn’t scale like software.

No idea was off the table. If it seemed exciting, Zume hired a team to pursue it, before anyone vetted the notion or figured out how the company might profit from it. “Everyone was told to do everything,” says a former employee. As a result, random things kept getting inserted into the grand vision. Zume pitched a vertical and sustainable farming business, named the “Gigaranch” after the Tesla Gigafactory, to the Saudi government. (Saudi Arabia’s sovereign wealth fund was SoftBank Vision Fund’s primary backer.) It explored a sensors business that would track how delivery food was handled. (SoftBank had acquired the chipmaker Arm in 2016 for $32 billion with the goal of powering an internet of things revolution.) “We did not make one cent for any of that,” says another employee.

As Zume made hundreds of new hires, its only revenue came from its original pizza business. Zume Pizza was bringing in “something like $1 million per year, ridiculously low,” says this source. At the same time, it was losing up to $150 per pizza it delivered.

Garden had actually decided in mid-2018 to spin out Zume Pizza and perhaps sell it. The robots—Pepe and the gang—were barely functioning, and the gear in the trucks kept breaking because everything was custom-built. When well-meaning employees suggested off-the-shelf replacements as a quick and affordable fix, Garden lost his temper.

Multiple employees recall the difficulty of approaching Garden with bad news. They tried to suggest that Zume narrow its focus; any one of the company’s business units might have found its legs and generated cash. But Garden would say, “SoftBank won’t let me.” That connected vision netted the company its millions, and that was the technology it had to stick with.

Restaurants, for example, loved the idea of using Zume’s high-tech food trucks to grow but complained that they were too expensive. Rather than retooling them, Zume concluded that it needed a cool brand to embrace its mobile kitchens and inspire other businesses to buy in. Executives (again) took an interest in a pizza business: &pizza, an upstart D.C.-based chain known for its distinctive oblong-shaped pies with cheeky, of-the-moment names like Giant Ass Breadstick (G.A.B.S.) and its founder Michael Lastoria’s advocacy for better-paying, more equitable food-service jobs. &pizza had about 30 stores on the East Coast and a shine that Zume Pizza had failed to develop out West.

Lastoria and Garden were introduced around the time SoftBank invested in Zume, and months later, &pizza executives flew to Mountain View to tour Zume’s kitchen. “We walked through and we’re like, holy shit, this is so much extra work,” says Andy Hooper, then &pizza’s president and chief operating officer. “We can do this faster with 12 people at $17 an hour.”

And yet, Hooper believed Zume could help &pizza. It wanted to expand to Miami, and Zume’s trucks could help it test the market. Zume’s software might help &pizza in its kitchens, organizing orders for pickup and delivery, he thought.

As they neared a deal, things got weird: Zume didn’t want &pizza’s money; it wanted to invest in &pizza. Hooper was invited to Zume’s Camarillo plant to present to the company’s board of directors. Rather than meet in a sterile board room, they did so aboard a tricked-out vintage double-decker bus in the parking lot. Zume had bought the bus and worked with a customization company to build a modular kitchen on the bottom floor. The top was a party space, the roof raised and sides extended to reveal a horseshoe-shaped bar rimmed by red leather barstools. The bus’s AV setup alone—the screens and sound system that displayed Hooper’s presentation—cost more than $100,000.

A few of Zume’s mobile kitchens were parked nearby, outfitted with kitchen equipment to demonstrate their versatility beyond pizza. Hooper remembers an Asian bowl concept, a sandwich concept, and an Indian curry being spooned into Zume’s custom packaging to show how well it held up to hot, saucy food.

“How Alex convinced their [investors] to take capital they had put into Zume and instead put it into &pizza is beyond me,” Hooper says. “It’s the most bizarre thing I’ve ever been a part of.” In addition to the cash, &pizza received six of Zume’s mobile kitchens, which &pizza would brand as its own, a stamp of approval on Zume’s vision.

Fresh off this deal in September 2019, Zume was nearly ready to unveil its Pizza Hut relationship. But the startup could only manage to send enough boxes that October for a one-day promotion at a single location in Phoenix. The test was well covered by the press, and Pizza Hut wanted more boxes. When Zume admitted it couldn’t deliver them in late 2019, it lost the contract.

At that moment, the turmoil within Zume was invisible to outsiders. Zume appeared to be—and acted like—a successful, high-growth company. It threw parties, and it opened offices in San Francisco’s startup-heavy SoMa neighborhood, Seattle, and Greenwich, Connecticut. (Employees were befuddled by that one too.) LinkedIn named it one of “the hottest [startup] companies to work for now.” Zume ranked No.25, listing an employee head count of 510 with 109 open roles across product, engineering, and sales.

Then SoftBank’s signature Vision Fund investment, WeWork, blew up.

After the failure of the coworking company’s IPO, Son promised investors that moving forward his other portfolio companies must be “self-financing.”

For Zume, which had continued to use its windfall to hire aggressively into December, the first priority in this new climate was finally shedding Zume Pizza. In her LinkedIn bio, Zume’s former CFO, Meredith Whitney, touts how she “built and executed a large-scale corporate reorganization/transformation plan based on selling Zume Pizza, an unprofitable piece of Zume’s original business.” But no one was buying an ailing pizza brand, and in January 2020, Zume killed its pizza robots and laid off 360 people.

The media, which had mostly accepted Zume’s claims, now lumped the company into the larger narrative about SoftBank’s excesses. A Bloomberg story—“SoftBank’s $375 Million Bet on Pizza Went Really Bad Really Fast”—exposed Zume’s earlier problems keeping the cheese on its mobile pizzas, among other juicy bits. The Verge kicked dirt on it with the epitaph “Zume out.”

With Zume Pizza out of the way, some employees saw an opportunity to finally build a business. During those early COVID days, there were paradoxically signs of hope. Zume Forward, the division selling the trucks and software, could help shuttered restaurants build delivery programs as they scrambled to adapt. But the company couldn’t yet let go of that larger vision or its cost structure. The trucks were still too pricey; interested restaurants told Zume they couldn’t justify the cost during the most uncertain operating conditions they’d ever experienced.

Garden was barely visible to the remaining employees working to hone Zume’s focus and win over customers. Starting in fall 2019, every all-hands meeting was canceled the day before, according to an employee. When April 2020’s all-hands predictably disappeared from the calendar on the eve of its scheduled time, Garden followed up the next day with an email laying off another 200 people, effectively eliminating Zume’s mobile kitchens division.

Zume’s last viable business was its compostable-packaging unit, which employed about 100 people. But that, too, seemed to be falling apart. Zume’s containers couldn’t legally hold food in some jurisdictions, including its hometown of San Francisco, because they contained PFAS, or “forever chemicals.” Days after the April layoffs, the original owners of Pivot, the packaging company Zume had acquired a year earlier, sued Zume for breach of contract and fraud. It alleged that the company had failed to provide the support it promised, which prevented Pivot’s leaders from achieving performance-based milestones tied to the cash payout.

Zume countersued in August alleging that it had been defrauded by Pivot misrepresenting its packaging expertise. Eventually, the lawsuits went into private arbitration. “I certainly learned a lot during this experience,” Pivot’s Lilley writes in an email. “The company ceased operations, which I believe speaks to how the company was setup [sic] and operated.”

During the early days of the pandemic, the company briefly made masks from its packaging facility, but soon Garden and Zume seemed to get serious about making the packaging business work. It licensed its designs and “advanced fiber thermo-forming equipment” (the machines that created food containers from biodegradable fibers) to a global network of distributors and manufacturers. An early employee managed to expand Zume’s packaging business to India, and in July 2021, Garden smiled broadly from the cover of The Pulp and Paper Times, an Indian trade publication that was likely not the sort of press he had in mind during Zume’s heyday.

The strain of trying to make Zume work weighed on Garden. “I had hair when I started this,” he’d say after he started shaving his head. He also, for the first time publicly, expressed some regret. “I didn’t realize—I should have, it’s my mistake—that robots making pizza was effectively the perfect storm for a venture-backed company,” Garden said on a podcast in July 2022. “If I had known that, I probably would’ve kept a much lower profile, because we’ve spent the last seven years explaining that we’re not a robot pizza company.”

Garden’s last-ditch hope was to make Zume a coffee-lid company. In November 2022, Zume’s machines did their first test run of molded-fiber lids. According to a former employee, most went into the trash. “It was way worse than it should’ve been,” they recall, while acknowledging that they expected some problems to work out. In a meeting the next day, Garden praised the production run. Potential investors had watched, according to Garden, and had never seen such technology before. “I was like . . . were you somewhere else? That’s not what I saw,” the employee says.

Five days later, Garden stepped down as CEO (but remained chair). The new CEO laid off close to half of the business’s Southern California staff, reducing head count from 140 to 78. In a somewhat cruel irony, that same month, Zume’s revamped packaging—now formulated without forever chemicals—received a Time Best Inventions of 2022 award. There’s still a link to “coffee lids” on Zume’s website but clicking it serves up an error message.

Last June, a decade after Garden filed the patent application that eventually became Zume, news broke that the company was being liquidated in an assignment for the benefit of creditors. The process, an alternative to bankruptcy, allows a company’s investors and fiduciaries to wind down a business quickly, with little publicity or disclosure of its remaining assets. In September, a private-equity-backed company based in eastern Tennessee—metaphorically about as far away from Silicon Valley as you can get—acquired Zume’s machinery to further its own growth.

Perhaps unsurprisingly, all of Zume’s main characters are doing just fine. In 2019, Collins launched Planet Fwd, software to make it easier to launch climate-centric consumer businesses. She created her own brand of crackers made with regeneratively grown grains to prove the concept—a remarkably similar setup as Zume, though Collins has disavowed any parallel. She sold Moonshot Snacks to Patagonia Provisions in March 2023.

Garden joined Fortress Investment Group in March 2023 as a managing director. At the time, Fortress was majority owned by SoftBank, but in May, SoftBank agreed to sell its stake to the sovereign wealth fund of the United Arab Emirates and Fortress management for $3 billion. The deal is currently facing scrutiny from the Committee on Foreign Investment in the United States.

SoftBank and Son have had a tumultuous several years, but the generative AI boom and a recent windfall in taking Arm public has Son back aggressively pursuing AI investments. As he wrote in 2019, “We want to be the conductor of the AI Revolution—a maestro of the orchestra comprising AI entrepreneurs.”

The coterie of former employees I interviewed sound truly disappointed when they talk about Zume. “I wish I had more time working with these people,” one tells me. Another calls her coworkers the smartest people she’d ever met.

They believed Zume had a viable business somewhere inside itself—if those divisions had been given a chance to thrive apart from Garden’s epic vision. Indeed, today there’s ample evidence of companies exploring similar ideas. Chipotle, which has its own investment arm, has backed a startup called Hyphen, a robotic system that the burrito slinger is testing to assemble its highly customizable orders. Euromonitor has predicted that ghost kitchens—restaurants that serve only delivery and takeout customers, such as CloudKitchens, the $15 billion startup founded by former Uber CEO Travis Kalanick—are capable of becoming a $1 trillion industry in less than a decade. Every delivery marketplace has predictive algorithms to forecast customer supply, restaurant availability, and driver demand, and in October, Domino’s partnered with Microsoft to figure out how to use AI to make ordering “smarter.”

Uber Eats has vowed to eliminate plastic waste from its deliveries by 2030 and has offered grants to restaurants to hasten the transition. DeliverZero, a reusable packaging provider that raised a $3.3 million seed round in July, is already building its own logistics network on top of Uber’s restaurant delivery business in three states.

And in Los Angeles, a startup named Stellar Pizza uses automation to bake pizzas in the back of vans. It’s staffed by a team of former SpaceX engineers. “Running a restaurant is not rocket science,” cofounder and CEO Benson Tsai tells me. “It’s much, much, much harder.”