Amazon.com Inc. executives, from Chief Executive Jeff Bezos on down, are famous for refusing to comment about competitors. No such reticence, though, from Andy Jassy, CEO of the retail giant’s cloud unit Amazon Web Services Inc.

Earlier this month, for instance, Jassy (pictured) tweeted in an ongoing spat with Oracle Corp. Chairman and Chief Technology Officer Larry Ellison, “In latest episode of ‘uh huh, keep talkin’ Larry,’ Amazon’s Consumer business turned off its Oracle data warehouse Nov 1 and moved to Redshift…,” one of Amazon’s own cloud databases.

It’s a good bet Jassy will take a few shots at Oracle and perhaps others at his keynote Wednesday morning, but he’ll spend most of his time as he customarily does at the signature event of the conference: debuting a mind-numbing array of new services intended to keep that competition at bay.

In this third of an exclusive three-part pre-re:Invent interview, Jassy talked about how he views the competition as well as the issue of potentially competing with customers and how he’s thinking about another route to growth: acquisitions. Like part 1 and part 2 of the exclusive interview that ran Monday and Tuesday, this portion is edited for clarity.

Also, check out re:Invent coverage all this week by SiliconANGLE in its special report on the cloud, including insights from its market research sister company Wikibon. Not least, SiliconANGLE’s livestreaming studio theCUBE will be interviewing dozens of AWS and other cloud leaders through Thursday from the show floor.

Q: How do you view the segmentation of the cloud market? Some such as Microsoft and Google include their Office 365 and G Suite revenues, which are SaaS, not infrastructure like AWS’, in overall cloud revenues.

A: They’re not apples to apples. This is where I think people get it wrong when they look at the relative growth of these companies. First of all, if you look at somebody like Microsoft’s cloud revenue, they have their comparable type of platform to AWS, but also Office 365. And we pay Microsoft a lot of money licensing Windows and SQL Server. They have that revenue, too, that’s in there.

We’re the only ones that really release our cloud numbers in a clean way. If you look at what the analysts triangulate at in terms of revenue for some of the other providers, let’s take what most people consider the second-place player in the cloud computing space. If you look at what they say about their cloud computing platform, I think the last results were something like growing 75 percent year over year.

People say, “Wow, 75 percent. That’s a higher number than 46 percent.” But the percent is really only a reflection of what the base is. When you have a business like AWS that is several times larger than the next five providers combined, 46 percent on that base is a much larger growth curve. If you look at the third quarter, 75 percent equates to about a billion dollars of year-over-year growth. Our 46 percent year-over-year growth on our much larger base equates to over $2.1 billion of growth. So we’re actually growing in terms of revenue more than anybody else, too.

Rivals galore

Q: How do you view the overall competitive landscape, which includes big companies from IBM to Microsoft to Google to Oracle?

A: For a value proposition that’s as strong as the cloud for customers, you’re going to always have lots of big-pocketed technology companies going after it. But at the end of the day what matters most is what customers want and whether you’re satisfying customers and the customer experience is right.

There’s so much noise and so much FUD and so many wild claims by lots of folks. And the thing that people sometimes get wrong is that it’s so inexpensive and easy to try in the cloud that it doesn’t matter what people randomly claim, people are going to try for themselves. And people get to the root of what’s real much easier in the cloud than they ever could in packaged software.

Q: You’ve made no secret of AWS’ rivalry with Oracle, and you recently tweeted about Amazon’s consumer business mostly moving off Oracle’s database to AWS databases. Do you regret that tweet?

A: I definitely don’t regret tweeting that. [Laughs]

When you have a company like Oracle that so often claims things that are untrue, at some point somebody has to call out the truth. All I was trying to do in that tweet was react to a number of statements that were completely made up and completely wrong, and just set the truth straight.

Q: Was the migration off Oracle hard or easy, and how do you talk to your prospective customers about that kind of migration?

A: We as a company have spent a fair bit of time over the last couple of years moving all of our database workloads away from Oracle to Aurora and to DynamoDB. This is pretty consistent with other customers who are making that significant shift. They’ll tell you that they’re saving a lot of money, they have a lot more visibility, they have a lot more control over their performance and their workloads, and they feel like they are being treated with a lot more respect. And they have more portability should they decide that they don’t want to run on top of something like Aurora or they want to move back to MySQL or Postgres or whatever they want.

Q: What do you hear from those Oracle customers when they move to Amazon?

A: It’s like they’ve been granted a new lease on life. I think that the cost savings, because it’s a tenth of the price; the operational performance; the ease with which they can actually operate; the way that they’re treated; not being constantly audited and fined or extorted for more business if they find some kind of violation in the audit. The ease with which that database works with the rest of the AWS platform; not having to fear that at any moment those database providers are going to raise their prices, or tell them that they can’t use their licenses on whatever infrastructure platform they want to use.

Q: Thomas Kurian just left Oracle in the rift related to the company. Now he’s heading up Google and Diane Greene left. What do you think about that?

A: I don’t know much about the decisions that were made. We don’t spend a lot of time thinking about who is leading the different cloud providers, so I don’t know very much about it.

Q: Would AWS allow other clouds to run its databases? Would you let other clouds run your databases?

A: If we were going to build software that could be run anywhere, and we were going to deliver in that type of package, sure. To date that’s not how we built our services. We build services which are a combination of software, hardware and data center services, and I think to change that approach and deliver just software on-premises is a really different delivery model, which I wouldn’t rule out, but just isn’t our focus today.

Q: How important is any one customer at AWS? We’ve heard from Google that they’re putting huge eggs in a few baskets to win enterprise business. It seems to be improving their position in TensorFlow and AI, for instance. Microsoft seems to be doing the same thing. How do you view that competitive posture?

A: If you look at the numbers and in practice what’s happened, we have more than twice as many machine learning reference cases and customers in business than any of the other providers. We have a really significant machine learning business and [customers said] more so than you’ll find elsewhere. So people are using us across the board for infrastructure, for machine learning, for the edge, really in all those areas.

Every customer for us is important. I think that you see some of these other providers who are trying to buy business, to have some marquee references just because they feel like they’re behind. But we don’t think about our business that way. We think that every single customer we have is important, and I think our customers will tell you that we put them at the center of what we’re doing. And we will treat customers who are small customers or startups every bit as importantly as we do big enterprises and public sector customers. You can’t build a business that is as big as what we’ve built, and where we’re trying to get to, and only treat a few customers as important.

Competing with customers?

Q: How are you handling the dilemma of a platform provider that wants to recruit new software companies that build on Amazon, but those companies might fear Amazon eventually will compete with them?

A: It’s one of those topics that sounds really interesting and really juicy, but in practice just hasn’t really been a significant issue when you look at the detail. When we launched AWS, we launched EC2 with one instance size and one instance family and one data center. Linux only, no load balancing, no persistent storage, no monitoring, no [auto scaling]. It was [literally] feature-poor. And some of it, we naively thought, “Maybe we won’t have to build it; others will build it.” And then what we found over time was that our customers said, “No, no, no; those are things that have to be part of the platform.”

I remember when we launched monitoring, which is CloudWatch, people said, “Oh my god, that’s the end of New Relic.” And of course it wasn’t the end of New Relic. New Relic has done really well, even though we have a good business in CloudWatch. And that has been the case consistently.

I can’t find a situation where a company was building a solution in an area that was more than just a shim, a loose shim on top of our platform, where when we launched something that was tangentially in the same space they lost all our customers or they didn’t compete. If you have a capability that has a good set of features and you’ve done a good job for customers and you keep iterating on top of the platform, you’re going to have big business forever.

Q: The one quote I heard was, “I hope Amazon doesn’t turn into the old Microsoft.” Because Microsoft used to have that kind of strategy.

A: Think about monitoring. Think about big data. Think about key value store or document store databases. Think about logging. These areas are so large that there’s not going to be one single winner. There’s going to be lots of successful players. Maybe we’ll have an offering because customers want us to and we’ll have some success, but it absolutely will not preclude others from being very successful as well.

More acquisitions

Q: Would AWS consider large acquisition in the SaaS area or anywhere else in the stack? Maybe Salesforce; they’ve got a big tower in San Francisco. Or some of the other big application providers. It seems farfetched, until you think about Amazon buying Whole Foods.

A: There is nothing off the table for us. We think about these things, big and small, all the time. We’d actually done quite a few acquisitions in AWS over time — Elemental of course was an acquisition. One of our most successful acquisitions was Annapurna, which was the chip design company in Israel which … has totally changed the network performance and capabilities of our compute and storage offerings. It’s been wildly successful for us. I expect that we’ll continue to do acquisitions.

Q: Mostly tuck-ins, off the radar?

A: It depends. We acquire a bunch of companies. We probably do less of the really expensive ones than others for two reasons. One is that we tend to be — right, wrong or indifferent — probably a little bit more rigorous analytically about what we think the payback is of that acquisition.

The second thing is that we have confidence that if the valuations are just really heated, and a lot of the technology acquisitions are very heated, we can put together a team to build capabilities that give our customers what we want. And because we’ve built so much, we have confidence building, and we tend to hire builders and be inventors.

I expect we’ll acquire more. It’s possible that we’ll buy something significant over time, but it would have to be the right offering that we thought made sense to the business and where we saw a reasonable payback and where the price we felt was reasonable.

Photo: Robert Hof/SiliconANGLE