For months and months, as Google’s brain trust searched for the ideal candidate to lead the company’s come-from-behind bid for its biggest new opportunity since founders Larry Page and Sergey Brin began putting ads on their search engine, one name kept coming up: Diane Greene. Little known outside Silicon Valley, Greene is a legend in tech circles and had been friendly with Page and Brin since their grad-school days at Stanford. She is married to Mendel Rosenblum, a well-known computer science professor there. Greene, who is 61, speaks softly and her round face frequently lights up into a warm smile. But behind that mom-next-door demeanour is a sharper edge, which flashes in her eyes when the topic turns to competition. Like the Google guys, Greene brims with ambition, brings an engineer’s mindset and has the entrepreneurial bug. The same year Page and Brin took a leave of absence from Stanford to start Google, Greene, Rosenblum and three others teamed up to start VMware, which revolutionised how companies managed their data remotely using a technique called virtualisation. Greene served as its CEO for ten years, building a business valued at $49 billion at its peak and developing a reputation as a singular talent: A computer science whiz with impeccable executive abilities who understood the business of selling technology to the world’s biggest companies. In a potent validation of her industry status, Page recruited Greene to Google’s board of directors in 2012.
As a board member, Greene began advising Google executives on a missed opportunity: Cloud computing. The idea of renting computing power to businesses had been around for some time, and Google had dabbled in it since 2008, when it first allowed startups to build their apps atop its prodigious network of data centres. But Google, distracted by other pursuits—search, maps, mobile and self-driving cars, to name a few—never got serious about the cloud business.
That didn’t stop cloud computing from developing, nearly a decade later, into a tsunami that is remaking how businesses think about and use technology. First a new crop of startups—Airbnb, Instagram, Pinterest and others—built most of their operations on cloud services supplied by others. More recently, large corporations—the likes of GE, NBC and Shell—have begun moving more of their applications to the cloud.
Now nearly every company is following suit, particularly industrial and logistics giants that see a tremendous opportunity in the so-called Internet of Things. By deploying webs of connected sensors to run their businesses more efficiently, these companies are now depending on IoT applications to manage truck fleets around the globe, monitor soil and ambient conditions in an agribusiness operation, or maintain elevators. For huge amounts of data there’s nothing easier than the cloud. As a whole, the shift to cloud applications represents a generational transition in business computing.
Google, which over two decades has built what is essentially the world’s biggest computer network—a sprawling array of massive, connected data centres spanning the globe and equipped with software to power all sorts of applications—seemed in a perfect position to capitalise on the new era of the cloud. Indeed, many inside and outside the company see cloud computing as Google’s biggest money-making opportunity outside of advertising. Yet, as a juggernaut that built a $75 billion business by providing free web and mobile services to consumers, Google did not have the DNA to create a giant new operation focussed solely on selling technology services to other businesses.
That’s why Greene’s name kept coming up. As Google came to terms with the opportunity, Page asked her to take over the company’s cloud computing efforts. She demurred. Soon enough, other executives and fellow board members followed up with her. It was finally Urs Hölzle, one of Google’s earliest engineers and the man most responsible for building Google’s computing infrastructure, who persuaded Greene to take the job.
There was one wrinkle: Greene was busy with a new, secretive startup called Bebop, which was developing tech to power easy-to-use business-software applications. So in November of last year, Google acquired Bebop for $380 million and named Greene the head of Google Cloud Platform, giving her the reins to build a salesforce and revamp a unit that spent $10 billion on growth in 2015. The appointment thrust Greene into an unusual role: As an Alphabet (Google’s parent) board member, she is, in some sense, Page’s boss. As head of cloud computing, she works for Google CEO Sundar Pichai, who reports to Page.
Greene faces a daunting task. While Google dithered, Jeff Bezos ate the industry up. His improbable bet in 2006 on something called Amazon Web Services, or AWS—a business so removed from the company’s retail core that its announcement left many scratching their heads—has become a runaway success whose growth has surprised even Amazon. With more than a million customers, AWS is on track to generate $10 billion in revenue this year, and its profits ($600 million in the first quarter alone) single-handedly keep Amazon in the black.
What started with a set of three services has since expanded to more than 70 different capabilities, and today entire companies are running on AWS. Its operations, powered by Amazon’s data centres, exert a gravitational pull on the entire internet. On Sunday nights, as Netflix subscribers settle in to watch their favourite shows, roughly 30 percent of all the bandwidth in the US runs through AWS, which powers the company’s videos. Amazon likes to boast that, thanks to AWS, building a tech business is as easy as assembling a set of Legos. “Anyone can now manage a global business driven by technology,” says Amazon CTO Werner Vogels, who recently unveiled a new set of AWS capabilities, including some tailored to IoT applications, at an event in Las Vegas.
For all of Amazon’s dominance, the cloud computing wars are just getting started. As corporate America and multinationals around the world grow more comfortable with the idea of replacing pricey computing hardware that they have to buy and maintain, the market is expected to balloon to anywhere from $370 billion (more than ten times its current size) to potentially as much as $1 trillion.
Regardless of whether the most bullish projections bear out, it is an enormous market. It’s no wonder that it has the biggest superpowers in tech battling one another for customers and mind share. A resurgent Microsoft, driven by CEO Satya Nadella, who once oversaw that company’s cloud computing unit, has bet aggressively on the business and established itself as a solid number two behind Amazon with its Azure cloud service. IBM, which remains ahead of Google in market share, and others are also investing to secure a piece of the pie, so the stage is set for a fierce multi-year battle. Greene understands she’s the underdog, but she’s said over and over that Google has what it takes to succeed: Grit, money and massive technological prowess.
Bezos’s ambition for AWS today is both simple and staggering: To make it bigger than Amazon’s retail business, which pulled in nearly $27 billion in sales in the most recent quarter. Yet no one, not even Bezos, would have thought this was possible when he and other Amazon execs hatched plans for AWS. It was, on many fronts, a clever idea. Amazon had already built a flexible computer infrastructure to store the data and run the applications that powered its fast-growing retail operations. Why not package those capabilities and offer them to others?
The launch of AWS proved to be perfectly timed: It coincided with a new boom in social media and mobile startups. Thrifty entrepreneurs racing to build apps loved the idea of off-loading all the headaches of running servers and data storage systems onto someone else, and AWS immediately became the number one option. It remains so to this day. “They’re not the only game in town, but they’re the automatic choice,” says Jason Seats, a partner at Techstars, a high-profile startup accelerator.
Bezos and crew are now engaged in a balancing act: Appealing to the much larger corporate market without losing credibility with the startup set. That transformation was apparent at the company’s re:Invent conference in Las Vegas. Held in October, the conference attracted some 19,000 techies and developers. Many were there to learn about new AWS tools—like on-demand business intelligence or security—that are essential to the CIOs and CTOs of America’s largest companies.
Among the big reveals was a new IoT platform to track and store the data streaming in by the petabyte, as every toothbrush, trash can, city bus and oil drill becomes a sensor. Amazon’s twist: Its cloud platform keeps a ‘shadow’ version of a device in its system that users can interact with when the device is offline. The list of customers that showed up to hear Vogels and AWS CEO Andy Jassy read like a Who’s Who of corporate America: Accenture, Capital One, Intel and GE. The CTO of John Deere took the stage to demo an app built on the new IoT package that monitors its vehicles in the field. BMW showcased a connected-car app. They joined the ranks of customers like the state of Washington and, in a public relations coup for AWS, the Central Intelligence Agency, which inked a $600 million multi-year contract to move some operations onto the cloud.
Yet as Amazon has cemented its lead in the fast-growing industry, it’s beginning to face the backlash that inevitably follows runaway success. Some companies have become hesitant to use too many AWS products for fear they could end up being locked in to Bezos’s world. Rely on Amazon for too many services, the argument goes, and it becomes prohibitively difficult to wean a company off, even as the bills start to pile up.
Nervousness has increased as Amazon, which lowered prices 51 times in the past decade, has begun to pull back on the cuts. “A CIO starts to take notice when they see the AWS bill become a $50 million line item,” says Sam Ramji, chief executive of the Cloud Foundry Foundation. Jassy says he’s aware of his customers’ concerns: “Lock-in is a question we get a lot, and it’s very understandable. We feel like we need to earn your business every hour, every month, every year.”
Months after its onstage cameo at re:Invent, BMW is back onstage at an industry event in San Francisco to show a sleek new cloud app. Called BMW Connected, the app can tell where a car is going, message a driver’s friends and even spot an open parking space. The conference is Microsoft’s annual developer showcase, Build. And this time it’s not Vogels who is introducing the carmaker but Scott Guthrie, who is the cloud and enterprise boss for all of Microsoft, a position that arguably makes him the most important man after Nadella. “We’ve had many customers switch from Amazon to us,” Guthrie says about the BMW double-dip.
In the cloud race, Microsoft knows that its place is, for now, behind Amazon. But it’s an ambitious number two. Microsoft’s sales pitch to big businesses is straightforward: We have more experience with the demands of large customers. For many, it doesn’t hurt that Microsoft brings to the table 30 years of enterprise sales experience and widely used products like Office 365 that can be packaged with Azure. “We differentiate the most in being enterprise-ready,” Guthrie says, noting Microsoft has more global data regions (32 to Amazon’s 13), better security and more flexibility on where to store data to address global privacy concerns.
Microsoft also beat Amazon to the punch with some IoT features, which Guthrie says helped it win the BMW business. After years of playing catch-up, Azure is hoping to get ahead of its rival with new products that offer its own machine learning and Artificial Intelligence capabilities.
When it comes to hard numbers, Microsoft still has a long way to go. While the company doesn’t break out its cloud revenue from the umbrella ‘Intelligent Cloud’ work group’s $6.1 billion in revenue, analysts peg its market share at around 9 percent, less than a third of Amazon’s. Guthrie says Azure is growing by more than 100 percent annually in users and revenue, adding 120,000 subscriptions each month. “We’re in that phase where the pie is expanding so dramatically that it’s not a zero-sum space,” he says. “Where we deeply engage, we win.” Some surveys seem to validate those claims. A poll of CIOs and CTOs across more than 300 vendors conducted in April by Enterprise Technology Research showed that as a group they adopted Microsoft at a faster rate than Amazon over the previous year.
Azure has had a few recent high-profile wins like NBC Sports, which plans to stream more than 2,000 sporting events from the upcoming Summer Olympics using its cloud. But to close in on Amazon, Microsoft will need its existing customers—roughly 85 percent of the country’s biggest companies—to spend more than they currently do on Azure.
Part of Nadella’s approach to lure them has required nothing short of a U-turn in Microsoft’s culture. The company, known for pushing a Windows-only world on its customers for decades, has become more willing to allow its software to coexist with that of others. If customers want to run some of their business on their own servers, some of it on AWS and some on Azure, that’s fine. “A lot of [Microsoft’s competitors] talk about hybrid, but when you actually look at what they’re talking about, it doesn’t have product truth behind it,” says Azure executive Jason Zander.
Still, plenty remember the bullying Microsoft of yesteryear, and even today’s powerhouses fear a return to those days. Recently, Salesforce seriously considered Azure before selecting AWS as its preferred cloud provider. Amazon and Salesforce both say that AWS won the deal on the strength and maturity of its products. But some believe it was no coincidence that Microsoft lost the deal just as it was beating out Salesforce to acquire LinkedIn in a bombshell $26.2 billion deal. “I wouldn’t say that was a pure technology decision,” Guthrie sniffs.
Greene has a lot of work to do before Google is taken seriously alongside Amazon and Microsoft. Until recently, aside from hosting the popular social platform Snapchat, Google had done little to show it knew how to work with large-scale customers. Greene is working overtime to fix that, reorganising her group to function more as a standalone business within Google and bringing together its technical, product, and sales and marketing teams for the first time. And like Microsoft’s Guthrie, she’s courting customers by offering access to Google’s prodigious technology: Analytics, machine translation, speech recognition, maps and more. “We’ll be doing a lot more so people can train their own data and build their own models using our software,” Greene says.
That pitch has gained some traction among tech companies with heavy data needs such as Spotify, an AWS customer that deployed some capabilities on Google’s platform this year. And Google is now signing up large corporations like Coca-Cola and Disney on the strength of its data chops.
Many of those customers are using Google for data while placing some critical apps with Amazon and Microsoft. “The advantage is flexibility,” says Coca-Cola CTO Alan Boehme, who is spreading his company’s 1,000-plus cloud apps across all three. And while all the cloud companies offer the usual sales incentives to land major accounts—discounts, free trials of new products, engineering help—Google is getting a reputation for buying its way to market share. The company is attempting to lure several large enterprise customers from Amazon with what amounts to a year of free computing, industry sources say. “Google is trying to make up for some gaps by throwing free resources at folks,” says Microsoft’s Guthrie. Greene is unfazed. “All of a sudden, we’re talking to pretty much every top-1,000 company,” she adds.
It’s not difficult to envision a scenario, some industry trackers say, in which all but the more complex cloud tools approach zero in price. While that would seem to favour Google’s strengths and give some indication of its path to success several years from now, Google’s credibility issue remains a serious hindrance. While Google has shown signs of giving Greene total support, it will have to show unusual patience to allow her team to build relationships that even she admits can take years.
More likely, it is Microsoft that will pose the biggest threat to Amazon for cloud leadership. Interest in Azure is at an all-time high, and its leadership team recognises the cloud’s strategic importance.
Microsoft will win out only if Amazon acts less like a Bezos-driven, ruthlessly effective operation. The cracks are there, if you look: AWS’s consistent 20 percent margins are creating a perception that its service is sometimes overpriced. And while plenty of companies across industries already fear the power of Google and Microsoft, Amazon is no different. Its dominance in retail makes AWS unpalatable to some. China’s e-commerce giant, Alibaba, has its own cloud division now; Walmart announced an open-source cloud project called OneOps earlier this year.
But bet against Bezos and his team at your own peril. If Amazon senses a threat from Microsoft in the coming months, expect AWS prices to drop while the quality of its tools continues to go up. “People have underestimated how fast Amazon is still innovating,” says Max Gazor, a venture capitalist. “Are we a hard-charging organisation?” Vogels asks. “Yes.” No one who has faced off against Bezos & Co. in retail doubts it.
Unless Amazon charges down the wrong path, it’s hard to see another company reaching $100 billion in revenue first. But if Google can show real traction and if Microsoft can narrow the gap, increasingly savvy companies will benefit from their fight to keep pace with Amazon. At Microsoft, Guthrie takes another stab at the question of whether Azure will ever take pole position in the cloud. “The real winner from all this competition is hopefully the customers,” he says.