Inside the collapse of Daqri's $300M bet on AR

This article was originally published on Protocol.com.

Everything went wrong for Daqri. The startup rode a wave of augmented reality hype and about $300 million in funding to a series of half-baked products before failing spectacularly and shutting down last year.

One of Daqri's last remnants was recently acquired by Snap: The company confirmed to Protocol that late last year it took on certain Daqri assets and about two-dozen employees, who now work in the company's newly opened Vienna office under the leadership of former Daqri CTO Daniel Wagner. Snap didn't disclose the purchase price, but the timing lines up with a $34 million acquisition disclosed in its annual report to shareholders.

In interviews, more than 10 former employees detailed Daqri's demise. In its decade of existence, the company built a retro-futuristic augmented reality helmet with a hefty $15,000 price tag, used fancy videos to sell the world on visual computing, and acquired a series of other startups in offices around the globe, only to see it all slip away when its glossy AR dreams collided with reality. It's a story that anyone working in the AR and VR space should take to heart, if only to avoid a similar downfall.

The early days, and a big pivot

Daqri launched in 2010 as an early mobile AR company that used QR codes and real-world objects to create interactive experiences on smartphones. Its first projects included a multiyear partnership with Crayola, which used Daqri's technology to turn coloring book pages into digital 3D objects on your phone.

The company also worked on a platform to let others develop their own AR content and experiences, but executives quickly realized that they were way too early for consumer AR. In 2014, a Kickstarter campaign for Daqri-developed interactive blocks attracted fewer than 200 backers, which was enough to fund the project but hardly a critical mass.

That's when company executives decided to change course and focus on businesses instead of consumers. It's also when Daqri started to get massive cash infusions from Tarsadia Investments, a Southern California-based private equity company. And, according to a number of its former employees, it's when a lot of Daqri's troubles started.

Former employees who talked to Protocol all did so on the condition of anonymity for fear of retaliation, including from the company's main funder, which still controls legal entities connected to Daqri. Some also expressed concerns about getting singled out at their current place of work for talking to the media.

As one might expect, any company that ends up laying off most of its employees will leave some with ill feelings. It's also worth noting that not everyone who Protocol talked with was critical of Daqri. Depending on their time of employment, position and location, some praised Daqri, its leadership, and job perks that included unlimited vacation, yoga classes and nap rooms, as well as frequent offsite events and parties.

Daqri's founding CEO Brian Mullins, his successor Roy Ashok and its main investor Tarsadia did not reply to requests for comment.

From a $5 million valuation to $275M in funding, with strings attached

There aren't a lot of public disclosures about Daqri's funding, but a lawsuit of a co-founder, who alleged that he got cheated out of his fair share, reveals that the company was valued at $5 million in an asset transfer in May 2013. The company announced a $15 million fundraise later that year, and Tarsadia went on to prop up Daqri with a steady flow of cash that had amounted to a reported $275 million by 2017. Sources told Protocol the total amount put into the company was in the $300 million range.

Tarsadia had made several bets in the pharmaceutical space and saw in Daqri a chance to invest early in a nascent technology that could be transformative to the workplace. Daqri employees with knowledge of the relationship told Protocol that Tarsadia was anything but a typical startup investor. Instead of taking a minority stake, the private equity company bought out Daqri. "None of the employees really owned any part of the company," said a long-time senior employee.

The ownership structure was problematic, multiple former employees told Protocol. "The incentives were wrong," said one former employee. "Everyone had already cashed out."

Daqri's executives were left without power, effectively turning them into subordinates. "It was a group of middle management that was reporting to the investors," the former senior employee said. Tarsadia executives are even said to have taken part in Daqri's executive off-sites, which resulted in Daqri's leadership avoiding frank discussions about problems at the company. "The investor was always in the room," the former senior employee said.

A futuristic helmet that remained a prototype

As part of its focus on businesses, Daqri announced the development of a futuristic-looking hard hat with integrated AR glasses in 2014. It was an audacious project that was pitched to construction companies, manufacturers of heavy machinery, and even the U.S. Navy.

The project required a lot of technical expertise so Daqri embarked on a spending spree, acquiring AR software specialist ARToolworks, EEG headband maker Melon, manufacturing startup 1066 Labs and holographic display developer Two Trees Photonics. The company quickly ballooned to around 400 employees from different corporate cultures in six offices in Los Angeles, the Bay Area, Ireland, the U.K. and Austria.

And the technical challenges remained immense. Daqri initially bet on Android to power the helmet, only to change course and switch to Linux soon before the product was supposed to ship in 2016. "A lot of people were unhappy" about that switch, said one employee who worked on the project for multiple years.

Daqri also learned the hard way that you can't develop workplace products exclusively in the lab. The helmet, which cost $15,000, was a highly specialized piece of equipment with sensors like a thermal camera that made for a good sales pitch. Workers in the field on the other hand often cared about much simpler features, like the ability to see what's behind them. Or they didn't want to be bothered with technology at all and found all the bells and whistles alienating. "We received pushback for the same exact thing that gave us orders," said the former employee who had been working on the project.

On top of that, Daqri had trouble obtaining the necessary certifications needed to use the helmet in industrial workplaces. "It was a product that was never finished for scale deployment," said the former senior employee. "I wouldn't say that it was functional," said the former employee who worked on the project.

Another employee, who was with the company about four years, had an even franker assessment: "None of it was real."

Marketing as a core part of the business

You wouldn't have known about any of the issues from looking at Daqri's marketing. The company received favorable coverage from a number of publications. It also churned out a steady stream of polished videos on its own, showing Daqri's helmet in a variety of industrial settings, on a Navy ship, and, inexplicably, worn by a man donning a jet suit, Iron Man-style.

"There were people making videos all the time," said the former senior employee. "It absolutely was showmanship. That was part of the company DNA."

"That became a core part of the business, that marketing arm," said another employee.

In conversations with Protocol, multiple employees attributed some of Daqri's heavy focus on marketing and promotion to its proximity to Hollywood. The company's Los Angeles headquarters was in the Los Angeles Center Studios facility, where it occupied two floors, including one that used to house the set for AMC's "Mad Men" and where decorations included movie props.

Daqri's relentless marketing wasn't just about selling AR to the world. It was also meant to woo the investor-owner, and ultimately convince staffers of the company's potential as well. In one instance, Daqri staff even produced a video for an internal holiday party that portrayed senior executives in the future, talking about how big of a success the company had become.

It worked: While some were skeptical about aspects of Daqri's technology, many bought into the vision of its founders. "We didn't even see the demise coming," said one former employee.

Layoffs, spin-offs, 'Hunger Games' jokes

That demise started around the beginning of 2017, when it became increasingly clear that Daqri's smart helmet wasn't going anywhere. In March 2017, Business Insider reported that 80 employees worldwide would be laid off. It resulted in the closing of two offices, according to a former employee.

Additional layoffs followed every couple of months. In Los Angeles, one of the two office floors quickly emptied and was subleased to another company. Former employees said that the atmosphere at the company quickly changed, with some criticizing the executive leadership for not handling the layoffs well.

Staffers who were let go were forced to leave the office as quickly as possible, at times without having a chance to say their goodbyes. One employee who stayed on well into 2018 said that it had turned into the most toxic workplace of their career. Another employee recalled that the team started to make "Hunger Games" jokes. "It was very stressful," that employee said. "The company was just hemorrhaging people."

At about the same time, the company shifted its efforts from the smart helmet to a simpler set of AR glasses that retailed for $5,000 — more than Microsoft's HoloLens AR headset, which first went on sale for $3,000 in 2016. Daqri's leadership was swapped out, with Chief Product Officer Roy Ashok taking over as CEO, and founding CEO Brian Mullins being forced to leave.

By this point, investor-owner Tarsadia was already looking for ways to minimize its losses. Instead of pouring in more money, Tarsadia proceeded to spin off some of the company's acquisitions into separate entities. The Two Tree Photonics team, which had started to develop AR for car windshields while at Daqri, incorporated as Envisics. A team headed by early Xbox architect Seamus Blackley, who had joined Daqri in March 2017, got spun off into an R&D company called Pacific Light & Hologram.

Daqri's staff dwindled in subsequent months, and the company ultimately shut down operations and let go its final staff in September 2019. Its office furniture was auctioned off soon after, and its patent portfolio went up for sale earlier this month. The lawsuit of Daqri's co-founder against the company is still ongoing and is expected to go before a jury soon.

Lessons from Daqri's failure

Asked why Daqri failed, former employees offered a range of opinions. Many of them cited excessive spending fueled by big checks, a lack of direction, and the inability to pivot from projects in time. They also cited the role of the owner-investor and failures of leadership.

But ultimately, the demise also has a lot to do with factors that aren't all that unique to Daqri: Making hardware is difficult, as is inventing products for a use case nobody yet wants. That's a problem that AR companies like Magic Leap face as well. The Florida-based startup, which raised over $2.6 billion in funding to date and is currently looking for more, is also known for blue-sky marketing that aims to not just sell its product, but AR in general.

Like Daqri, Magic Leap and Microsoft have also more recently pegged their hopes for AR's commercial success to enterprise use cases, betting that companies will buy hardware at price points out of reach to the average consumer. But as Daqri shows, selling to big companies, and getting AR to work in industrial contexts, can be extremely challenging as well.

And while Daqri may have failed, it may still play a significant part in the future of AR. "Daqri was also the beginning of a lot of people's careers," said an employee who went on to work for a different AR company. If anything, alumni could help other companies avoid the same mistakes.

Lessons learned from Daqri's failure may be valuable to any company looking to build AR products, including Snap. The Vienna team formed from Daqi's wreckage is said to be working on AR software for the company.

Software, one might add, that could ultimately also power a future AR version of the company's Spectacles wearable — just don't expect it to look like a helmet.

Janko Roettgers, the author of this article, is now writing Lowpass, a weekly newsletter about AR, VR, streaming and more. Sign up here for free.