It can’t help Osterloh’s nerves that minutes earlier, Pichai was out on the same stage making a grand case for the historical significance of this day. “We’re at a seminal moment in computing,” Pichai told the audience, as he explained how artificial intelligence would create a revolution on the scale of the internet or the smartphone. Google’s efforts centered on Google Assistant, a virtual helper that Pichai had first announced a few months earlier. Assistant promised to create a “personal Google” for everyone on earth that would help them find information, get things done, and live life more efficiently and enjoyably. Pichai made clear that Assistant was a bet-the-company kind of product, and that Google was deeply invested in building the gadgets that would put Assistant in people’s hands. Then he introduced the new guy, Osterloh, who was going to make it happen.
Over the next hour, Osterloh and his new coworkers introduce a half-dozen products, including the Pixel phone, the Home smart speaker, and the Daydream View VR headset. None of them were Osterloh’s idea—the folks in Mountain View had been building hardware long before his arrival. It’s just that most of it wasn’t very good or successful.
Google could no longer afford to make ho-hum gadgets. Alphabet, its parent company, had become the world’s second-largest corporation by building software that worked for everyone, everywhere, delivered through apps and websites. But the nature of computing is changing, and its next phase won’t revolve around app stores and smartphones. It will center instead on artificially intelligent devices that fit seamlessly into their owners’ everyday lives. It will feature voice assistants, simple wearables, smart appliances in homes, and augmented-reality gadgets on your face and in your brain.
In other words, the future involves a whole lot more hardware, and for Google that shift represents an existential threat. Users won’t go to Google.com to search for things; they’ll just ask their Echo because it’s within earshot, and they won’t care what algorithms it uses to answer the question. Or they’ll use Siri, because it’s right there in a button on their iPhone. Google needed to figure out, once and for all, how to compete with the beautiful gadgets made by Amazon, Apple, and everyone else in tech. Especially the ones coming out of Cupertino.
Google does have some huge advantages—its software and AI capabilities are unrivaled. But the company has tried over and over to build hardware the same way it builds software and learned every time that that’s simply not how it works. Its supposedly innovative streaming device, the Nexus Q, flopped dramatically. Its “best in class” Nexus phones were eclipsed by competitors—and even its own hardware partners—within months. And Google Glass, well, you know what happened with Google Glass.
Osterloh wasn’t hired to dream up new products. He was brought in to teach a software company how to endure the long, messy, totally necessary process of building gadgets and to change the company’s culture from the inside. It’s not enough to have great software and the industry’s finest collection of artificial intelligence researchers. To take on Apple, Google had to finally learn how to build good hardware.
The man in charge of Google’s hardware renaissance has always had a weakness for gadgets. Growing up in Los Angeles, Osterloh has fond memories of taking apart the junk computers in his dad’s office and trying, unsuccessfully, to reassemble them into one epic supercomputer. Yet his first love was sports. Tall and athletic from an early age, Osterloh was an all-section volleyball and basketball player, and he enrolled in Stanford not because of its Silicon Valley cred but because it was a great school in California where he could keep playing sports.
In his freshman year, however, he sustained two knee injuries that threatened to end his athletic career. Osterloh hit an emotional bottom. “So much of my identity was in athletics, and I had to totally reinvent,” he says. He started seeking other ways to feel the same highs he did in sports: a team working toward a common goal, the thrill of accomplishment, the joy of the daily grind. He found his way into an engineering program and worked hard to make up for his late start in the major. Something about computers engaged the strategic, problem-solving part of his brain that had once been filled with inbounds plays.
Osterloh is still a sports nut—his Google office is easy to find, it’s the one with the huge poster of Warriors star Stephen Curry on the window—but the tech industry quickly became his home. After graduating in 1994, Osterloh landed a consulting gig, but he didn’t like that all he made was documents and presentations. So he went back to Stanford, this time for business school. After a summer internship at Amazon, he took a job at the venture capital firm Kleiner Perkins Caufield Byers, where he researched possible investments in mobile technology. BlackBerry was starting to generate interest, and Osterloh dove into a case study of it. He set up BlackBerry’s first device, the Inter@ctive Pager, and was amazed by how well the little messaging machine worked. He couldn’t stop thinking about it.
Kleiner had a company called Good Technology in its investment portfolio, and it sent Osterloh to help it figure out a business model. Originally, Good’s plan was to build modules for the Handspring Visor, a modular PDA that many thought would be the next big computing platform. Good’s first device was an MP3 player module called SoundsGood. But the Visor never took off, and the SoundsGood sold terribly. Osterloh presented a new idea: Let’s compete with BlackBerry. He thought Good could develop simple syncing and messaging software, and because BlackBerry by this point had become immensely powerful and valuable, any competitive idea was attractive to investors. Good raised millions.
Good was supposed to be a software company, but it needed a vessel for its code. The leadership team met with BlackBerry, which had recently begun making smartphones. Once BlackBerry execs saw what Good had built, “they hated it, because it was way better than their software,” Osterloh says. “And they realized we were an enemy, not a friend.” Palm and Danger were working on smartphones, as was Nokia, but none could match a BlackBerry. It became clear to Osterloh and Good that the only way to give their software a home was to build devices themselves. They began working on a BlackBerry-like gadget they called the G100.
Osterloh lights up remembering the days he spent building the G100. “It was so fun going through the design process, testing it with users,” he says. Everything about it was new and complicated: getting the keyboard just right, tweaking the trackball until it felt perfect, making sure the battery lasted several days. “It was so hard shipping that product,” Osterloh says. “When we shipped that thing, I was like, ‘This is what I want to do forever.’” Not only had he found his calling, he’d learned that the only way to get the most from your software was to build the hardware to match.
Unfortunately for Osterloh, Good didn’t want to make hardware forever. The G100 shipped in 2002 to rave reviews, but others in the company saw it as a mere reference device, a blueprint of sorts for other companies to follow and tweak. They assumed the phone industry would turn out like PCs: Many companies would produce hardware that all ran the same software. Yet there were no good phones to build for. “We went through this desert of terrible device after terrible device that never ran our stuff properly,” Osterloh says. Good built software for every phone it could find, eventually even working with contract manufacturers like HTC to try to improve the experience, but it never again found something that worked as well as the G100. “Businesses would come to us and say, ‘We love your software, but we hate Treos,’” Osterloh says, referring to the smartphone line from Palm. He never forgot that.
In 2006, Good was bought by Motorola, the onetime feature phone giant whose reign was under siege from smartphone makers. Motorola had no real software expertise and no plan for smartphones, and Good came riding in like a white knight. But the timing couldn’t have been worse. Only days after the acquisition closed, Motorola’s Razr, once an incredible cash cow, stopped selling almost overnight. Apple announced the iPhone not long after. Osterloh knew it was coming: Before the Motorola deal, he and Good had worked with Apple to build Good’s software into the new device. He told his bosses, many of whom dismissed Apple’s touchscreen oddity, that they were laughing in the face of the future.
While he’d been meeting with Apple, Osterloh and Good had also been working to integrate their software with an operating system for smartphones called Android. Now, as a Motorola employee, he saw Android as the company’s one defense against the iPhone. Osterloh became convinced that the only hope for Motorola was to produce a competing smartphone as fast as possible, and that meant using Android. Eventually Motorola came around, largely due to the efforts of new CEO Sanjay Jha, who showed up in 2008 and almost immediately shut down every division but its Android one. Osterloh helped create and ship the Cliq and later the Droid, which was the first great Android phone and the device that saved Motorola.
Not long after, Osterloh left for Skype, where he spent two years as head of product. But his break from the hardware world was brief. Google was in the process of buying Motorola for $12.5 billion and was looking to place new leadership at the company. Dennis Woodside, a longtime Google exec who had been chosen to lead Motorola, and Jonathan Rosenberg, a senior vice president at Google and longtime adviser to the company’s founders, called Osterloh to see if he might want to come back and lead Motorola’s product management team.
Google’s offer seemed like a perfect match, a chance to build hardware within Google, working alongside the now wildly successful Android team. With Google controlling both hardware and software, they could at last take on the iPhone.
Except that’s not how it turned out. Terrified of alienating its other Android partners, like Samsung and LG, Google went to great lengths to keep Motorola at arm’s length. “There was effectively no technical integration,” Osterloh says. “And that wasn’t quite what I expected.” He thought he would bring software and hardware together at last, but instead Motorola was treated as a wholly separate company. “It was tantalizingly close to my dream job,” he says. “But it never quite got there.”
Google’s relationship with hardware has always been awkward. Most of the company’s physical products are born the same way: Someone has a great idea for software, but they can’t find the right gear on which to run it. That person then sets out to build the missing gadget with very little help. Google tends to treat these products as reference devices or sources of inspiration, proving that an idea can work and hoping an ecosystem of hardware makers takes it from there. As a result, Google’s list of orphaned products and abandoned ideas—from the Chromebox to the Nexus Q to the Nexus Player—is enough to fill a Circuit City.
That’s no surprise: Making hardware runs counter to Google’s entire corporate culture. The company shuns process and management, two things a hardware maker can’t do without. In its software development, Google actually encourages and applauds chaos, inviting anyone at the company to just build something and see if it works. (At one point, Google even experimented with a corporate structure involving no managers whatsoever.)
The company’s most successful products are subject to constant refinement. Former CEO Eric Schmidt calls this system “Ship and Iterate,” and in his book How Google Works he makes a consistent case for not even trying to get things right the first time. “Create a product, ship it, see how it does, design and implement improvements, and push it back out,” Schmidt writes. “Ship and iterate. The companies that are the fastest at this process will win.” When Google became Alphabet, all the company’s longer-term projects broke off, to give them breathing room away from Google’s ruthless product scythe. They were all called “moonshots,” as if anything that takes longer than a year might as well be impossible.
Ship and Iterate simply doesn’t work with hardware. A single tweak can cost weeks and millions of dollars. Every small change ripples through the entire supply chain, changing vendor timelines, requiring new tools, and slowing everything down. If one part is late, you’ll miss your ship date, and it’s not like you can move Black Friday. Oh, you want 50 percent more product than you thought? You’ll get it in six months if you’re lucky. There is no bending the hardware world to your whim.
Even when hardware development was going well, company culture didn’t support those teams in getting the software they needed. “We had to go beg and plead to make all the software teams care,” says Rishi Chandra, the Googler charged with building the failed Google TV platform in 2010 and later with developing Google Home. The engineers working on Chrome or Android were used to building products that would touch millions, even billions of people. They’d ask Chandra, how is your thing going to get that many users? And why should we care until it does? The culture is almost the antithesis of Apple’s. There, the software executives always work with specific products in mind; senior VP of software engineering Craig Federighi’s goal is to make the iPhone great, just like everyone else at the company. Google’s priorities are comparatively all over the map, as it tries to support its own products, its partners, and the whole internet-using universe all at the same time.
At Google, the culture revolves around software. That’s what it’s best at and where it makes its billions. With its push for a virtual assistant, that ethos was no different. Except that this time the stakes seemed much higher.
Pichai was certain that this helpful chatterbot would be the way billions of people interacted with Google in the future. Done right, Assistant could be an omnipresent artificial being, able to help with all the tasks and requests people have throughout the day—whether on their phones, a device like Google Home, or the lightbulbs, dishwashers and thermostats that will soon come online. It would connect people with information and services in ways that are more natural, more contextually aware, and more helpful than what is possible with only keyboards and screens. It might even inspire people to use Google more. Oh, and if Google didn’t get it right? Amazon’s Alexa, Apple’s Siri, and Microsoft’s Cortana were ready to swoop in.
In these early days of voice tech, users still struggle to figure out what their assistants can do beyond setting timers and playing music. Yet the only way for Assistant to improve is for Google to convince people that it’s worth interacting with now. Google needs more data to understand more voices, accomplish more tasks, and convince developers to extend its functionality and incorporate Assistant in their own products. A key first step was making sure Assistant was always easily accessible, no matter where you were.
When Google had built hardware in the past, it had done so through partnerships with seasoned manufacturers. But its relationships with its Android partners were souring. Samsung, Android’s most important partner, was developing its own virtual assistant, Bixby, and distancing itself from the Mountain View giant. Google couldn’t even rely on its traditional Nexus program, through which Google would work with a manufacturer like LG or Huawei to build new devices. Those relationships gave Google little control over anything besides aesthetics, and partners would often keep their best tricks for themselves. “Last year, [HTC] helped us build Pixel, and then a few months later, they shipped the U11, and that phone had the best smartphone camera in the industry,” Osterloh says. Virtually every Nexus device was quickly followed by an even better handset from that same partner.
Unfortunately, Google had already divested itself of its own hardware expertise, selling Motorola to Lenovo in 2014 for about $3 billion. Nest, its other giant hardware acquisition, had lost its founder and was embroiled in management and product turmoil. If Google wanted to do hardware, it needed to start over—and this time, to do it all inside Google.
As the company was preparing to bring hardware in-house for real in early 2016, Osterloh was leaving Motorola. He didn’t want to move to China, where Lenovo is headquartered, and he’d landed an offer to become CEO of DocuSign, the electronic document-signing company. He called Jonathan Rosenberg, his longtime adviser and confidant, to thank him for all his help during Osterloh’s time at Google. Rosenberg stopped him in the middle: “You said I was helpful, right?” Yes, absolutely, Osterloh replied. “Well, would you do me a favor and have a conversation with Sundar?” Rosenberg asked. He told Osterloh that Pichai was looking to start a hardware group and wanted some advice. Just advice, nothing else.
Osterloh’s meeting with Pichai quickly turned into a job interview, with lots of questions on both sides. In many hours over a few days, Osterloh realized Pichai was actually, finally, talking about his dream job. He also started to believe that Google was at long last serious about making hardware.
But Osterloh had been burned before. So he tracked down Hiroshi Lockheimer, the head of Google’s Android team, who had worked with him at Good and also happened to be among Osterloh’s closest friends. They spent a whole day together talking about how they might once again be colleagues. Osterloh asked question after question about how hardware and software would integrate, and how building hardware internally could co-exist with the rest of the Android ecosystem. “I didn’t want to join the company if it was going to be like Motorola, where it’s difficult and there’s tension,” Osterloh says. He found the opposite: Google was ready, serious, and prepared to make hardware a priority. So Osterloh called DocuSign and told them that he wouldn’t be taking the CEO job after all. Then he became a Googler again, this time for real.
Immediately upon his arrival, Osterloh set out with Rosenberg to find every hardware project happening at Google, no matter how small. They found more than a dozen projects involving upwards of 1,000 people. Some were working on Nexus phones, others on a new line called Pixel. There were hugely publicized long-term projects like Google Glass and the Project Ara modular smartphone. Some Googlers were building Chromebooks, others were working on a new kind of Wi-Fi router. No centralized structure connected these teams, nor was there an overall plan. Osterloh called it a loose federation, “the European Union of hardware.” And he didn’t mean that in a good way.
Osterloh centralized all that hardware under his leadership, giving 55 percent of those 1,000 employees a new manager. Rather than having an executive in charge of each product, Osterloh chose to implement a “functional” structure, giving his leaders oversight of a larger segment of the Google hardware organization. Ivy Ross, formerly head of Google Glass, was put in charge of all hardware design. Mario Queiroz ran product management. Ana Corrales, a longtime manufacturing exec and Nest’s CFO and COO, was tapped to oversee all things operations and supply chain. The team began to centralize their planning and forecasting, and to streamline their conversations with suppliers. They made five-year plans, which were anathema to Google.
Osterloh’s current and former colleagues describe him as a kind man and a good boss. “The thing I appreciate most with Rick is that he is really about preaching patience,” Chandra says. In conversation he’s voluble and excitable, prone to answering simple questions with a 45-minute response. He is, according to past and present colleagues, perfect for this job: great attention to detail, slow to panic, quick to decisions. Above all, he’s a huge product geek. “He changes phones all the time, and he wants us to change phones all the time,” Corrales says. “I don’t want to change phones all the time!”
Part of the impetus for Osterloh’s new structure was to make sure nobody felt like their job was tied to one product, so they wouldn’t panic if that product were killed. Because Osterloh needed to kill some products.
He went through every hardware initiative at Google, choosing which to continue and which to wind down. None of the decisions was easy, Osterloh says, but two were particularly hard. He’d been around the Ara modular phone project since its beginnings at Motorola and believed fully in its mission: to build a $50 phone with upgradeable parts, which could last longer and be greener than any other device. Yet the device ended up being less modular and more expensive than anyone wanted. “So it was rather like every other phone, except there was an ability to add up to six or so modules to the back,” Osterloh says. He wanted to build one phone, not many, so he shut Ara down.
With Google Glass, too, Osterloh understood the vision but couldn’t figure out how to achieve it quickly. He ticks off the things you’d need to make a great face-worn augmented-reality device that aren’t yet possible: longer-lasting batteries in smaller packages, faster processors that generate less heat, and a populace ready to use such devices. “In the long run, this is going to be a key part of what we do,” he says. “But the timing is a key uncertainty.” In the meantime, Osterloh re-released Glass as an enterprise tool, where it found a surprising niche with factory workers and warehouse employees.
While he was re-writing org charts and culling product lines, Osterloh had also been working with the higher-ups at Google to figure out what, exactly, Google’s hardware strategy should entail. They coined platitudes like “radical helpfulness” and sought ways to communicate humanity and approachability, but mostly they focused on three words, in a very specific order: AI, software, hardware.
He had to embrace the fact that even as Google gets serious about gadgets, the company’s focus is, and always will be, elsewhere. Osterloh is fond of pointing out that Moore’s Law, which famously predicted the rapid rise in computational power, is mostly dead. It’s getting harder and harder to make fundamental leaps in power and capability. Google’s advantage, he says, is in its algorithms and neural networks. Osterloh’s job is to push Google’s AI capabilities more deeply into people’s lives.
For the new hardware team, the task was clear: Find more ways to get Google Assistant in front of people and build a sustainable business around it. Oh, and hurry, because Google is already behind, with Siri and Alexa already entrenched in consumers’ minds. Osterloh poured resources into the Pixel phone, a nascent project between a few Googlers and HTC, in which Google was taking on full responsibility for design and engineering for the first time and HTC was merely the manufacturer. The hope was that with this phone, at last, Google could give its software the physical form it needed. “We have a terrific ecosystem position with Android, but I think no one was really delivering the full Google experience,” Osterloh says.
Designing hardware and software in tandem allows for the detailed decision-making that makes people fall in love with their phones. Seang Chau, an engineering VP on the Google Pixel phone team, gives an example: For scrolling to be smooth and fast requires deep control of such factors as when to turn on the GPU, how to tune the processor, how to manage the power supply, even which cores of the chip run at any given time. “You pick up another phone that hasn’t had all those choices made, all those components chosen,” he says, and you notice the difference. Apple has been touting for years that its products excel because it builds both software and hardware; now Google is following suit.
Osterloh decided to flank the Pixel effort with other devices that were good matches for Assistant. Another team within Google had in the past released two terrific laptops, called Chromebook Pixel, that only saw limited commercial success. Osterloh told the team to go build something even lighter, thinner, and better—and to integrate Assistant. They decided to call it Pixelbook and set off on their way. A different group started working on headphones they called Pixel Buds that would provide access to Assistant without the need for a phone. The Google Home team and the Chromecast crew were also part of the push.
“Eventually, it will be the case that users probably have a constellation of devices to get things done,” Osterloh says. Google’s definitely thinking about tablets, definitely thinking about augmented-reality glasses, definitely thinking about wearables, and many more. But Osterloh speaks of “earning the right” to chase the buyers of those devices, of wanting to prove his team’s viability in existing markets.
With the employees of the hardware division settled into their new roles, Osterloh and his team started working out their production needs. He and Corrales toured manufacturers in Asia, telling them what Google was up to and how they’d be interacting going forward, and they brokered new deals with suppliers. In November 2017, Osterloh oversaw the $1.1 billion acquisition of an HTC division that brought more than 2,000 engineers to Google, many of them the same people who had spent the last decade building Nexus and Pixel devices as outside partners. The deal, Osterloh says, was “very important to help us scale faster. Hiring one at a time takes a long time, and our aspirations are to go faster.” In early 2018, Alphabet brought the entire Nest team under Osterloh’s leadership, giving him control of the company’s smart-home future as well.
There’s plenty of reason to rush. Apple and Samsung continue to push new competitive software onto their hardware, and new classes of devices are getting better all the time. Yet Osterloh notes time and again (and maybe partly as a reminder to himself) that building hardware is a slow process, that that’s a good thing, and that patience is a virtue. This moment is his chance to prove a career-long thesis about putting hardware and software together, and he wants to get it right. “There’s something big at stake here, both for him and the company,” says Ivy Ross, Google’s head of hardware design and one of Osterloh’s key lieutenants. “When you personally have a reason to drive, you’re just that much better.”
It’s now October 4, 2017, a year to the day since Osterloh first showed off the new generation of Google hardware.
The moment is familiar. Osterloh is wearing that same gray Henley and standing in the wings while Pichai explains that artificial intelligence is the future. This time, however, they’re at the SFJazz Center, a larger and more impressive venue. They’ve been rehearsing for weeks, tweaking the words and order of their presentations to better explain what Google is up to.
The biggest difference, Osterloh says, is that this time he knows the story. In 2016, he was trying to retrofit a grand narrative around many disconnected products that turned out to be well liked but only moderately successful. (Amazon’s Echo was still crushing Google Home, and the Pixel didn’t exactly dent the iPhone’s bottom line.) With 18 months behind him, Osterloh now gets to show the world what Google hardware really looks like.
As he takes the stage, clearly more confident than a year ago, Osterloh starts with another overview. He reminds the audience of the 2016 launch and mentions the recent HTC acquisition. “By working more closely together, we’ll be able to better integrate Google hardware and software,” he says. “And our products,” he continues, with a smile expressing something between joy, relief, and aggression, “have built up a lot of momentum going into our second year.” Then he throws to one of Google’s trademark sizzle reel videos, with everyone from YouTubers to Modern Family’s Phil Dunphy loving their Google Homes and Pixels.
Over the next 90 minutes, Osterloh and his leadership team introduce a litany of new products that have Assistant baked in. At every turn, rather than tout spec sheets, Osterloh explains how artificial intelligence can extract remarkable experiences from ordinary hardware. As he introduces the Pixel 2, he mentions that although it only has one camera, its software includes an algorithm trained on faces that can help it turn standard photos into beautiful portraits. Google tweaked the audio processor inside the Pixel Buds to streamline the experience of using Assistant through headphones and to enable real-time translation. The Home Max can adapt its audio output to any space to improve sound quality. A new camera called Clips identifies snapshot-worthy moments and takes photos and video all on its own. Osterloh and his team went through gadget after gadget, showing with each one how Google could make its products smarter than the competition.
The launch goes well, but not perfectly. Some users have issues with the Pixel 2 XL’s OLED screen, which Google selected to show off its cool, contextually aware software. Others find fault with the touch panel on the Home Mini, which was accidentally turning itself on and recording hours of audio. Reviewers praise the idea behind the Pixel Buds, but not every feature.
All these issues make Osterloh angry—“I lose sleep every time customers aren’t happy,” he says—but they seem to energize him as well. He knows how to handle these kinds of challenges: more rigorous process, tighter management. It’s typical hardware stuff, lessons he learned long ago. Unlike the previous year, however, this time Osterloh has a clear path forward. The products under his watch are part of a story that spans the whole company. With its mission ironed out, Google needs to do more than release devices. It needs to learn how to win.