Back in early November, right around the time his name started appearing on the short list of candidates to become Microsoft’s CEO, I had lunch with Satya Nadella in New York. It was eye-opening for a number of reasons, most of them positive for Microsoft. I left convinced that this guy would be a great choice for the job. His comments carry considerably more meaning now that he really is the new CEO.
He seemed genial, relaxed, and likeable. For all the many virtues of his predecessors Bill Gates and Steve Ballmer, both of whom I interviewed many times, those aren’t the first words that spring to mind. Yet like them, he is a strategic thinker and his devotion to Microsoft’s mission appears semi-religious.
Nadella at the time headed Azure, Microsoft’s cloud platform business, which offers software as a service, as well as overseeing Xbox Live and other big cloud services operated by the company. He also ran its enterprise applications business–Windows Server and SQL Server, among others. (Nadella’s job before all that had been running the Bing search service.) His comments didn’t directly address Microsoft’s device businesses or its consumer offerings, and he wouldn’t address the job rumors. But you can hear where he puts the company’s priorities.
In one word, the answer is “cloud.” Microsoft’s traditional software products, which business and individual customers put onto their own machines, confront a growing list of companies offering similar capabilities as a pure online service, including Google, Amazon, Salesforce, Dropbox, Box, and others. But Azure itself is now a major player in that market, even as Microsoft still sells the packaged products. Nadella said that in fact only three companies are now operating cloud businesses “at scale”: “Google, us, and Amazon, in that order.”
Many are starting to wonder if everything will go into the cloud, and a small number of players will thus offer software services to business. Nadella, for all his success with Azure, believes the answer is no.
He explained what he called his “new thesis” about what it will take for any company to be successful with “infrastructure” elements like server operating system software, databases, and virtualization services. In his opinion, success in Microsoft’s on-premises apps that customers install themselves will depend on, and be intertwined with, its success in the cloud. “To be a relevant infrastructure guy in the future you need three things to fall into place: First, a diverse set of SaaS [software-as-a-service] applications. Second, you have to run an actual cloud service, yourself, and then sell it to other people. Now that I see the pain of running my [Azure] organization, I realize you have to run it at scale. It has to be globally available.”
But how much enterprise software will customers still buy? The third part of his thesis explained why he sees a central continuing role for Microsoft’s non-cloud products. “In the post-Snowden world, you need to enable others to build their own cloud and have mobility of applications. That’s both because of the physicality of computing–where the speed of light still matters–and because of geopolitics. My clients are building their own apps, buying my server software, and putting it on their premises. It has to be easy for a Nigerian cloud services provider to offer a service close to its customers.”
“If every sector of business and society will be driven by software–how does that get enabled?” he asked. “By highly-paid computer scientists funded by risk capital in Silicon Valley? Or by lots of engineers who can build it themselves? If it’s the latter, then there will be others who build infrastructure beyond the three or four mega-cloud guys.” Selling to those players will be a key market for Microsoft, he believes.
Nadella surprised me when he explained how Microsoft now develops the traditional enterprise software it sells to customers: “Microsoft has no SQL Server developers. We have only Azure developers. But every 12-18 months we reverse engineer [the code] into a product we can sell.”
How big of a server business will Microsoft have in the future? “There are nine million servers sold annually. Of those, just one million are sourced by the big guys. What we’re trying to predict is: in the future is that all going into the one million category? Or will there be some balance?” Clearly he thinks the answer is yes.
He said the opportunity for Microsoft to grow remains huge. “Three out of four servers today are still Windows OS,” he explained. “Our infrastructure business is over $20 billion, and growing double digits, both top line and bottom line. But worldwide it’s a $2 trillion IT spend. So nobody is a high-share player. In these shifts, all the category definitions get thrown out of the window.”
Nadella said it’s very much in the interest of big enterprise hardware companies like EMC, Cisco, H-P, and Dell that everything not get sucked into the cloud. But he added that some of those companies will face tough choices about whether they should themselves go into the apps business: “Can you run a cloud service on your own? Yes, but it’s $5 billion and there are big barriers to entry to keep up with the big boys.”
The spectre of Google and Amazon loom especially large. But Nadella says working with Microsoft is essential if big traditional hardware companies are to fend off the inroads of the cloud giants: “I tell them ‘I’m your only hope in town to create this infrastructure business.’ ” If Microsoft helps make it possible for many smaller cloud services companies to arise around the world, and for big companies to operate their own clouds, that should continue to generate business for the hardware players, he believes.
His message to colleagues at Microsoft is aggressive. “I say lead don’t follow,” he told me. “And play offense. Don’t even look at Amazon. Let them be successful.”
But Microsoft’s challenges go beyond the technology shifts. One big one is its historically-siloed product-line structure, whose managers often have been more combative than cooperative. Recent restructuring has significantly softened the previously stark lines between businesses, but traditions dies hard. If all the category definitions get thrown out the window, so must the turf battles, for Microsoft to stay successful. But it’s clear this new tech industry leader is thinking hard. He’ll push his company hard, too.
This article was originally published at Forbes.com.