Jamie McCarthy/Getty Images; Justin Sullivan/Getty Images; Loon/Alphabet; Emmanuel Dunand/AFP/Getty Images; Samantha Lee/InsiderSummary List Placement
For years, images of Loon's stratospheric balloons could be spotted throughout Google's campuses. Employees pitching to advertising clients would sometimes slip in a picture or mention of the project in their otherwise dull, statistic-heavy decks. Occasionally, a balloon could be spotted drifting across the presentation welcoming new employees to the company.
Wherever the balloon was seen, it was there to say the same thing: Google was way more than an advertising machine.
But last month, after nearly a decade of development, parent company Alphabet grounded the Loon project. The company's desire for a global, balloon-driven internet dwarfed to financial woes, a failure to attract more outside capital, and a sinkhole of skyrocketing costs.
Loon, which sat among Alphabet's "moonshots" – a hodge-podge of spin-off projects focused on longer-term ideas – was especially close to the heart of Google cofounder and former CEO Larry Page.
Now, it joins a growing graveyard of the company's boldest ideas.
In interviews with half a dozen employees, as well as analysts, it became clear that while Loon's executives started to worry about the project's funds, they were also planning costlier upgrades in the hopes of luring in investment.
"My memory of Loon was the word 'hubris' was used a lot," said one former employee who, like the others in this story, spoke to Insider on the condition of anonymity because they were not authorized to speak to the press. "They talked about taking this science fiction experiment and turning it into a commercial product."
"The message was, to do something bold, you have to have a degree of hubris."
As Loon burned through $100 million a year, tensions grew between employees who wanted to spend more on improving the technology and those who wanted to push ahead with what Loon had already built. As the company struggled to make headway on key deals, it explored new revenue lines such as delivering internet to cruise ships and offshore oil rigs.
This is the story of how Loon became one of Google's most ambitious and zaniest ideas – and how it came crashing back down to Earth.
A Loon spokesperson declined to comment on the details of this story.Enter the A-team
In 2015, Larry Page was asked at the Fortune Global Forum to name the Google moonshot project that he was most excited about. Google had recently blown up its corporate structure into Alphabet – a new parent holding company that allowed it to keep its stable of more cash-intensive experiments at a distance from Google, its core financial powerhouse.
Page quickly named "Project Loon," which lived inside the company's hush-hush moonshots division, known as X. Loon's idea was to spread internet access across the globe using a network of helium balloons floating 60,000 feet in the sky, partnering with telecoms companies to beam their LTE services to hard-to-reach places.
"There's very many places you go in the world where you still don't have a cell signal. I think Loon could change that," said Page, "in a way I think most people don't appreciate."
Others had caught on. Facebook was also racing to control the stratosphere for its own solar-powered drones, which brought their own challenges.
Then there was Elon Musk's SpaceX, which Google had just invested $900 million in, and which wanted to bridge the digital divide using a constellation of satellites hovering much higher up in the low Earth orbit.
As was later revealed, this would also require the user on the ground to install a special antenna. For Loon, all you'd need was a cell phone.
In 2018, Loon was spun out into its own Alphabet company, giving it its own corporate structure, payroll, and incentives scheme – a process sometimes referred to internally as "Alphabetization." While these companies enjoy the backing of Google's deep pockets, they are also pushed to become commercially viable and seek outside investment.
Loon kept its headquarters inside X, giving it a direct line to some of Google's brightest minds and zaniest scientists whenever the team ran into engineering problems. Among them was the "captain" of moonshots himself, Astro Teller.
Teller, who often arrives to meetings on roller blades, was perhaps an even bigger champion for the project than Page, though insiders say his involvement in the day-to-day running of Loon after the spin-out was minimal beyond sitting on the company's board.
Running the show was left to CEO Alastair Wingarth and his squad of executives, known internally as "A-Staff."Loon flexed with early experiments
Loon tried to demonstrate its early potential through a number of short experiments.
In 2017, the Federal Communications Commission granted Loon an experimental license to get Puerto Rico residents back online after Hurricane Maria destroyed the island's infrastructure. In 2019, Loon deployed balloons to Peru just 48 hours after the country was hit by a magnitude 8.0 earthquake.
But bigger plans – the ones that would actually make Loon money – were struggling to take off, including a project to bring full LTE coverage to Sri Lanka.
In early 2019, Loon raised $125 million from SoftBank subsidiary HAPSMobile, which was working on beaming internet access from high-altitude drones. The SoftBank backing gave Loon an important runway and a strategic partner in the stratosphere, from which Facebook was now absent.
"The team had been accustomed over the years to be an R&D project," said one former employee. "SoftBank coming in gave us a sense of governance and accountability."
At the same time, Loon was pouring millions into research and development and employee salaries, without reaping significant revenue.
Loon's tech gradually improved, but by the time is shut its doors its balloons could only stay afloat for an average of 150 days, after which time the balloon film and helium needed to be replaced. Some of the team thought that getting the balloons up faster and making them last longer would not only make them more efficient, but it would make Loon more attractive to outside investors.
To get there, Loon built an expensive new launch system at its Winnnemucca, Nevada site that was much taller than the existing one, making it capable of launching larger balloons. It was ultimately never used, two former employees said.
This all created internal tensions between leaders who believed that Loon needed more time to perfect its technology and some who wanted the company to push ahead on deals using its current technology and get money on the table, those same two former employees said.
"They just kept redesigning the wheel," said one. "I kept thinking, at some point someone is going to walk through that door and start dropping pink slips. It couldn't go on."
The cost to keep Loon afloat kept growing. By the end of its lifespan, the company was spending around $100 million a year, according to a report by The Information last November. A separate former employee familiar with Loon's financials confirmed that figure was accurate, describing an average monthly spend of $10 million-$15 million.Loon explored other forms of revenue as it tried to stay afloat
One former employee said there was an understanding inside Loon that, as the world came increasingly online, the business was not going to make much money from underserved communities alone. Instead, it saw more value in building an infrastructure from which it could form lucrative licensing deals. "The end result wasn't supposed to be the balloon," they said. "The balloon was supposed to help us build out a hardware and software structure in the atmosphere.
"There was kind of an acknowledgement from the team that trying to make money off the backs of people who telecoms companies had ignored wasn't a good strategy to make money," the employee said.
In early 2019, for example, Loon announced a deal to licence its networking software to Canadian telecoms company TeleSat's constellation of low-orbit satellites.
Meanwhile, leadership were thinking up other ways to make money.
In 2019, Loon began exploring a strategy to beam internet connectivity to cruise ships, and entered talks with Carnival Cruises for a possible partnership, according to two former employees familiar with the plan. Loon thought its technology could be supplemental to existing satellite coverage, which can often be slow and unreliable for passengers.
This collaboration would have another efficiency: The balloons Loon was sending to Africa would be flying over the Caribbean Sea anyway, so the company could have them deliver internet to the cruise ships during their journey, one former employee said.
Loon also considered beaming internet connectivity to offshore oil rigs, onshore mining sites and other remote work sites, particularly those that involved using internet-connected equipment, according to former employees.
None of these side projects came to fruition, and it was only last summer that Loon even achieved its first commercial deployment, delivering 4G LTE internet coverage to nearly 31,000 square miles of Kenya in partnership with a local telecom company.
CEO Alastair Wingarth told The New York Times last year that the team chose Kenya because it was open to adopting new technologies, yet it took two years for Loon to get the project off the ground. Meanwhile, a contract to bring internet to users in Peru remained stuck in a similar regulatory hell.
"To get these government sign-offs and actually put our balloons on the stratosphere? That was hard," said one former employee. "That was really hard."
Unlocking more airspaces would have also allowed Loon to share balloons between countries. For example, if a balloon in Kenya flew off course, it could be rerouted to Mozambique and used there. Clearing these aerial pathways meant Loon could also deliver its balloons to their destinations efficiently – surfing any particularly good stratospheric winds that appeared en route.
But as time went on, Loon realized how tough it was to land these agreements. The company tried to get clearance to fly balloons over Venezuela, which would have made it easier to travel to other parts of South America, but the country's authoritarian government would not allow Loon to do so, one former employee said.
Some governments were also suspicious of allowing Loon in their airspace. Employees say it was not uncommon for foreign dignitaries to visit Loon's sites and offices to inspect their balloons for surveillance technology.
As Loon wrestled with sticky geopolitics, SoftBank's money was fast drying up, and the company spent most of 2020 trying to attract new investment. Loon had some leads, the most promising of which was a new deal with SoftBank for a second cash injection, according to two former employees familiar with the negotiations.
The amount Loon hoped to get from the new deal could not be learned, but one said they expected it would have had to be at least $100 million to be worth it.
When SoftBank and other leads ran cold, Loon no longer saw a future in which it could finance itself. After a turbulent 2020, the Loon board convened and decided it was time to pull the plug, according to a source familiar with the decision.
The decision to kill Loon was ultimately Astro Teller's to sign off on, they said.Insiders blame timing, politics, and stodgy telecom companies
Few Google projects have attracted as much fawning press coverage and interest over the years like Loon. "It kept Google sexy," as one former employee put it.
What killed Loon? That depends on who you ask. One employee pinned the blame for Loon's fall on underestimating complex geopolitics. Another said it was "stodgy" telecoms companies that worsened the regulatory headaches.
While leadership appeared to downplay what Elon Musk's SpaceX and other rivals were doing in this domain, two of the former employees Insider spoke to thought these companies presented an existential threat in the long term. Loon's decision to play in a nascent space meant it was pouring more into R&D costs but made less attractive to investors.
"The technical feasibility of what SpaceX was doing was tried and tested," said one of those former employee. "We've been launching satellites since the '60s."
Even among Alphabet's gaggle of future-gazing moonshots, which include life sciences division Verily and self-driving company Waymo, Loon was a tough sell to investors.
"Verily and Waymo have the advantage of targeting markets that already exist, so it's easy to quantify," said a former executive, who also asked to remain anonymous, in a conversation with Insider last year. "It's easy to make a case for investment and why it makes sense."
While Loon struggled to find funding, billions were being poured into internet-delivery satellites operating several hundred thousand feet above. In January this year, internet satellite company OneWeb announced it had raised $400 million to continue deploying its network.
$350 million of that was from SoftBank.
Loon's business case has also shrunk over the decade as more of the world has come online, said Ernesto Falcon, a senior legislative council at the Electronic Frontier Foundation who focuses on internet connectivity. He also said he believed Loon had an uphill battle in working with telecomms companies.
"Loon was probably attractive to telcos because it extended their network, but if you were a large player with a larger reach, you didn't need anyone," he said.
Alphabet has become more conscious of its cash-burning subsidiaries over time. In an interview with Fortune last year, shortly after he was appointed CEO of Alphabet, Sundar Pichai hinted that there would be more discipline when it comes to moonshots. A few days later, Google shut down Makani, another bet that was working on power-generating kites.
"I think with the 'Other Bets' we are definitely at a phase where, while we take a long-term view, we also want to marry that with the discipline of making sure they are doing well," said Pichai.
After 10 years in development, Loon apparently failed to meet that bar.
Perhaps, in the end, it was all about timing. Another former employee wondered if Loon was simply too ambitious too soon. "I think that they prematurely sought the path of commercialization, thinking they would get there by winning these deals with other telecoms companies and the satellite industry," they said.
"If they had held off, I think we would still have a wacky internal project at Google X. They simply tried to grow up too fast."
- Starlink's first UK user told Insider what was like to go from zero broadband to zippy internet speeds in rural England
- Google signs multimillion-dollar deal with major Australian publisher as it tries to signal cooperation on news payments
- Gaia was a wildly popular yoga brand. Now it's a publicly traded Netflix rival pushing conspiracy theories while employees fear the CEO is invading their dreams.