Microsoft's Satya Nadella on Why the AI Push is His Most Important Work of His CEO Tenure

Last week I went to Redmond, Washington for the launch of AI-powered Bing at Microsoft headquarters. This week’s newsletter is a transcript of the conversation I had with CEO Satya Nadella about his AI strategy, nine years as CEO, and the path ahead in a rocky economy.

JON FORTT:

I am here at Microsoft Headquarters in Redmond, Washington. Thank you, Satya Nadella, for having me — CEO of Microsoft. Just got done with a presentation that struck me as very different because of the velocity with which you're getting technology to market here. But before we get into that, I'm seeing these headlines about the toll of the earthquake in Turkey and Syria, the death toll topping 7,000. Microsoft is a global company. What's your outlook on that and the response?

SATYA NADELLA:

Yeah, I mean, first of all, our heart goes out to everyone impacted, and in fact our first priority was to ensure even the safety of our own employees in Turkey and Syria. And we've activated our disaster response so that we can be there for any institution, any organization that we can help.

And we've also activated our philanthropies in giving, so that we can all support the people of Turkey and Syria as they go through this. In the Pacific Northwest, this is something that we think a lot about. And so really our heart, and we'll spare no effort in helping.

JON:

Absolutely. And I do want to turn to your news of the day, which is pretty monumental in technology. Today's announcement, affecting search, affecting browsing. You said it's a new day for search, the race starts today. Google rules search and browsing. Best case, how long is it going to take you to dethrone them?

SATYA:

Well, first of all, it's, to me, these paradigm shifts or platform shifts are a great opportunity for us to innovate. The first thing that is a priority for me is not about dethroning anybody. It's more a priority for us to say, well, how can we rethink what search was meant to be in the first place? In fact, Google's success in the initial days came by reimagining what can be done in search. And I think the AI era that we're entering gets us to think about it. And that's what we’re really motivated by, Jon, and then what you saw today in terms of really building not just a new search experience but thinking of it as what's a copilot for the web look like in this AI era is what's exciting.

JON:

What we saw is a large language model helping to not just compose queries but deliver detailed answers about things like a multi-day travel itinerary. Information about artists in Mexico. Pretty deep. People have been playing with ChatGTP and DALL-E for a while. You're bringing this to market now. What's the economic opportunity?

SATYA:

Well, I mean, we already have an at-scale search business even though our share, as you rightly pointed out, is very, very small. The good news is we start with already a business that is profitable. And here's the interesting thing: the most profitable, large software business is search. So I look at this and say, "Look, I just have to earn one user at a time in incremental GM." I've never, ever felt this liberated in terms of opportunity in the days ahead. So, I'm very excited about innovating, meeting the needs, knowing that the search category is the most profitable and large category. Just on Bing, Google makes more — I mean, just on Windows, Google makes more money on Windows than all of Microsoft. So that alone, should sort of give us the impetus to really go after this.

JON:

I saw a report that sometime ago you weren't satisfied with how quickly AI research at Microsoft was translating into product that people could use. Is that the case? And was there a point where you turned the corner on that and you were reaching the cadence that got you to where you are today?

SATYA:

Yeah, I mean, I grew up in a company where I was taught not to be satisfied by anything that is happening in this place at any time. And so that's just my general persona, I would imagine. I was taught well by a lot of people who came before me on that dimension.

But the reality , innovation is an art. Sometimes you get it right, sometimes you don't. I think we're, at Microsoft what we are good at is to be able to sort of persist. If I think about even with Bing, right? We have learned so much by being in the Bing category that it has helped us. Right? We wouldn't be where we are in Azure if it was not for Bing and the distributed systems infrastructure. We wouldn't even have all of the Azure cognitive services if it was not for what was applied AI.

So I would say, of course I want more pace. And what you're seeing, I guess, this year in particular is what we've been working towards, Jon, for the last, I would say, three years. To build up the partnership with OpenAI was absolutely instrumental. But beneath what OpenAI is putting out as large models, remember, the heavy lifting was done by the Azure team to build the compute infrastructure. Because these workloads are so different than anything that's come before. And so we needed to completely rethink even the data center up the infrastructure that first gave us even a shot to build the models. And now we're translating those models into products.

JON:

Some people are gonna say OpenAI is not the only AI out there working on these things. Microsoft's certainly not the only cloud company. We think about Amazon right down the road. We think about Google. How do we measure whether Microsoft really has a significant first-mover advantage here? Is it on the pace at which you roll out this AI technology to other Microsoft product categories and services compared to the competition? The quality and usage of those?

SATYA:

Yeah, I mean, it's a great question because at some level, I look at, what is it, 48 years after our inception, here we are in 2023. Are we as excited about what comes next as we were, perhaps, back in the day? And I think the only way to keep up with that and being a consequential company as technology platforms shift is to be competitive in each layer. So, for example, we're thrilled about OpenAI's use of the supercomputer. Guess what, Inception, Mustafa Suleyman's company is on Azure. So is Character.ai. So there are others who are going to build on Azure. And we welcome them. So we want to make sure that the Azure infrastructure's open to everybody.

Then we want to make sure that the foundation models of OpenAI are available for everybody. And that's where both OpenAI will have their API's, we will have it through Azure. Then comes our own incorporation in our products. So I kind of look at it and say, "You should measure us and customers will measure us by looking at how competitive are we in each layer of this innovation."

JON:

And so how important is scale and getting to market both quickly and frequently with these updates? Because AI learns from usage. And so the more usage you get, arguably, I think — correct me if I'm wrong here — the faster your platform should be learning and the better it should get compared to the others.

SATYA:

All tech products have network effects. Feedback loops are super important. There are certain peculiarities, I would say, of AI products. For example, the pretraining data is not just about more data. It has to have diversity of data.

So even when we think about scale, you have to think about multiple dimensions, not just scale for scale's sake. But that said, yes, it will be very important. That's why I think it's important not to be in a lab. You have to get these things out safely.

That's why we think about responsible AI and the safety around it, both from the design decisions as well as the technology as first class. You don't think of it later. You think about it right when you're building the model. And then, yes, the feedback loop at every layer of the stack will be important.

JON:

Talk to me about the downsides. And I know you're a technology optimist. It's my job to try to poke some holes —

SATYA:

Absolutely.

JON:

in that. Because even, I mean, the Rust Belt was created through moments and innovations like this. And maybe the AI Rust Belt won't be a place. Maybe it'll be a category of worker. How much resource are you putting into thinking —

SATYA:

That's a great point.

JON:

through those effects and what kinds of mitigations should governments be taking even now to lessen the impact for some —

SATYA:

Yeah, that's a great —

JON:

workers?

SATYA:

Yeah, it's a great, great point. I mean, that's right. I mean, whenever you have any type of, I'll call it, industrial revolution, some people talk about this as perhaps the industrial revolution brought to knowledge work, even.

And so displacement is harsh, hard. And it impacts people's lives in a generation. That's one of the hardest things. And so I think the first thing is we'll have to probably do some of our very best work in skilling, reskilling, and wage support for a variety of jobs. These are certain things which I think we as a society in our form of democracies, will have to figure this out.

But we want to contribute to that. But the other thing that — I'll just give you one anecdote though. Take, say, what's happening with Power platform, product of ours. That's probably the one product of ours which gives me the greatest pleasure. Because when I see —

JON:

Really?

SATYA:

The reason is because whenever I see someone using it, they don't have an IT degree. They are working as a frontline person in healthcare, in retail, in manufacturing, in services. They pick this up, learn how to write an app, automate a workflow.

Now this OpenAI model's built in, which is, you can prompt a workflow. And they suddenly start participating in the digital transformation of their org. Guess what happens? First thing, wage support goes up. Because they're now considered part of I.T.

That is one anecdote. I'm not saying that that's the answer. But we should also not fall for the lump of labor fallacy because there will be new jobs or existing jobs that are now becoming more productive because of AI that will get better wage support. And we should tap into all that while being clear-eyed about what we as a society need to do to ensure that any displacement doesn't cause the hardship like we know from industrial England in the 1700, 1800s.

JON:

Yeah, very often the workers who end up benefiting from the wage support are not the same workers who got displaced in the first place. So I wonder about that. But this announcement is also happening just after Microsoft, and a lot of other companies are reporting layoffs.

We just got layoff news from from Dell, which of course is one of the companies that uses the Windows operating system heavily. How much clarity do you feel you have into the cost structure of Microsoft heading further into 2023? Is that more a reaction to what has already happened? Or is your cost structure now aligned based on what you're pretty sure is going to happen for the rest of the year?

SATYA:

Yeah, I mean, there's, first of all, any of these actions around jobs at Microsoft is painful for the people who have been impacted. And we're trying to do the best we can to ensure that they find their next play. And we also thank them for all the work that they have contributed to Microsoft.

So the one point I'll make there is, if you look at least the displacement that's happening at Microsoft as well as the rest of the tech industry, this is something that we saw even multiple years ago where the number of tech jobs outside of what is considered the tech industry are higher than what is in the tech industry.

It does require relocation, maybe even adjustment in salaries. Because tech industry salaries are different than, let's say, in the auto industry or in the energy sector. But that said, there is going to be labor market movement here which, in the long-run by the way, is going to be beneficial for the tech industry and everybody else.

But to your point about our own cost structure, we did have a challenge with our cost structure because the growth did moderate. I mean, that said, just to put it in perspective, this is one of the things that I think especially our investors should think about. Is our last quarter, our Commercial Cloud constant currency growth was 29%. That's, what, ten times U.S. GDP?

You know, yeah, it is not 40%. But it's ten times U.S. GDP. It's the largest business at Microsoft. That's the other unique thing about Microsoft is our largest business is a fast-growing business. And so we want to make sure we invest in it while being disciplined. And where we overshot, or quite frankly, it's not even about overshooting on demand. Sometimes you have to take bets in tech which don't work. And if they don't work, you’ve got to just fast-fail on them. And that's something that we have to learn and get better at.

JON:

Speaking of fast fail, and maybe this isn't the best transition. The Activision Blizzard acquisition looks to be running into some real headwinds in the U.K. How certain are you now compared to where you were that that's going to go through? And how important is that to Microsoft?

SATYA:

I saw that Bobby was on your network this morning. And I thought he did a good job of explaining why he and I both are enthused about this combination because I think it'll only bring more competitiveness to the gaming industry.

And look, I think I look at this and say, "At the end of the day, the regulators around the world have to make the choices." And I would only submit to them that if they really seriously think about competition, they have to sort of really reflect on, is this going to be helpful to bring more competition? Right?

Think about this: There are people who make more money in gaming who don't even build games today. Like, maybe we should look at that. But even if you look at the console market, we should probably look at Microsoft's share of the console market in Japan. As perhaps a question that somebody should ask and say, "Oh wow, I wonder why that is that small."

And maybe they should actually start competing more. So I hope that the regulators take an approach that is going to truly be beneficial to gamers, it's going to be beneficial to all publishers and make all of gaming more competitive.

JON:

Finally, I believe it was nine years ago this week that you were named CEO of Microsoft. Where does this AI announcement and, I take it, the series of announcements we should expect from you from here, where does that fit in, in terms of strategic importance of all the things that have happened in the previous nine years?

SATYA:

It's a great question, Jon. I mean, one of the things I think a lot about is you can only be relevant in technology if you are good enough to see the waves of change and then to reorient your technology and innovation agenda and the business model agenda.

So I would say we have gone through some very harsh ones. The last one we went through was obviously the mobile and cloud. We caught one, we missed one. But the cloud transition was very hard. Like, most of what I — and I look around, even my senior leadership team, we're all the veterans of that transition.

In fact, even Charlie Bell who came from Amazon is also part of our team. And he was there when AWS was getting created. So I feel it's kind of like that. It's 2007, 2008 cloud. But the one good thing here is the business model side of this, because it builds on the cloud. So one of the foundational things about AI, who knows how it'll all reshape all software categories. This is one of those things, time will tell. But I'm much more optimistic in terms of both our capability to lead from day one versus having to do that catch-up which we've done too in some cases.

JON:

If you're going back 15 years, then it's the biggest thing to happen since you've been CEO.

SATYA:

Oh for sure. I always look at what elite devs are buzzing about. It's sort of the best early indicator. I've not seen something like this-- since, I would say, 2007, 2008 when the cloud was just first coming out.

JON:

All right, well, looking forward to seeing how this plays out and to talking to you more about it before too long. Satya Nadella, thanks for joining us on CNBC.

SATYA:

Thank you so much.

Forethought CEO Deon Nicholas on the Potential of AI in Industries: CNBC Working Lunch

Alteryx CEI Mark Anderson on Earnings, AI, and Equipping Workers with Low-Code Data Tools

Freshworks CEO Girish Mathrubootham on Earnings and the Velocity of the AI Wave

Aisera CEO Muddu Sudhakar on Building Enterprise AI Applications and the Significance of Microsoft’s Announcement