Adam Selipsky, Amazon Web Services Inc.’s new chief executive, doesn’t want to show all his cards just yet. But it’s clear, as he looks ahead in anticipation of the No. 1 cloud computing provider’s 10th annual re:Invent conference starting today, that he aims to set AWS on a new course. 

On the surface, it doesn’t look like he needs to. When Selipsky took the helm of AWS in the spring of 2021, he inherited a business that was run by Andy Jassy, since elevated to head all of Amazon.com, that was running on all cylinders. Revenue from the cloud computing unit of Amazon will top $60 billion this year, and SiliconANGLE sister market research firm Wikibon estimates revenue will grow 36% from last year — remarkably, for its size, an acceleration over recent quarters.

But Selipsky faces unprecedented challenges. He must navigate AWS into the next decade with well-funded competitors such as Microsoft Corp. and Google LLC nipping at its heels in an attempt to get a larger share of the lucrative cloud computing market. Microsoft, for example, has developed its cloud business by building software as a service application offerings that run on top of its Azure infrastructure. Google similarly leverages its Workspace suite of productivity apps to compete in cloud computing.

Thus far, AWS has not copied its competitors by offering the same kind of SaaS products for business users. Instead, it has primarily relied on delivering a wide array of infrastructure products and services — more than 200 of them — from myriad methods of compute and storage to data analysis, artificial intelligence, fraud detection, enterprise search and much more. Developers then use the services as building blocks to run sophisticated and scalable applications that more directly compete with traditional SaaS applications.

However, that focus has left organizations that have limited in-house cloud expertise and whose information technology services are still locked inside their own data centers — the vast majority of $4 trillion in IT spending — to cobble together AWS services, Lego-style, as they creep into the cloud.

Now, that focus on “foundational services,” as AWS calls the infrastructure services it’s still building at a breakneck pace, is about to expand. “More and more, customers are asking us to provide them with higher-level abstractions on top of AWS services,” Selipsky told SiliconANGLE in a sit-down interview at AWS headquarters in Seattle, where he provided exclusive details about what he will talk about in his Tuesday keynote. 

One iconic example of this “horizontal” software push that Selipsky waxes eloquent about is Amazon Connect, a complete customer contact center in the cloud. Using AWS’ own artificial intelligence and other services, Amazon Connect — a business that Selipsky said is “on fire” — has helped the likes of Barclays plc reach more people at lower cost than traditional software solutions.

“We’re looking at a whole bunch of those horizontal use cases,” Selipsky said. “In the coming years, I think you’ll see us continuing to look at horizontal use cases such as the call center. We’ll undoubtedly try and help by releasing higher-level capabilities that both have brand-new functionality as well as bundling up, if you will, some of our existing capabilities.”

At the same time, Selipsky says, AWS needs to continue to expand vertically as well, by providing more complete solutions for specific industries such as health care and manufacturing. To do so, the company is bringing its services to the edge of the network, including traditional data centers, the factory floor, and even the field, and establishing an operating model that integrates AWS more deeply into businesses in virtually every industry.

AWS is doing this by creating purpose-built services that can be deployed horizontally across myriad types of organizations in support of their digital transformations. AWS’ insights into vertical industry dynamics – gleaned through daily interactions with millions of customers globally — lead to unique perspectives into how to tune the software and AI to solve hard industry-specific problems that are traditionally expensive and require long lead times to deploy. AWS for health, manufacturing and media are some examples of AWS’ industry-specific solutions.   

Strategic moves have already been made to advance these trajectories. Several years ago, in a move that supports both horizontal and vertical expansions, the company embarked on an ambitious initiative to design its own silicon chips, reducing reliance on traditional microprocessor suppliers and charting a course to control its own destiny in performance and cost. In addition, befitting Selipsky’s recent history as CEO of the business intelligence firm Tableau Software, acquired by Salesforce.com Inc. two years ago, AWS is digging deeper into providing enterprise customers more ways to provide artificial intelligence and data analytics services usable by mere mortals, not just data scientists.

It’s a heady set of challenges for Selipsky, despite his previous long history at AWS. Venture capitalist and early Amazon investor Matt McIlwain has been close to the core team at AWS, including Jassy and Selipsky, since the early days of the business and sees Selipsky as the right leader to steer the AWS rocket ship for the next 15 years. And he views Selipsky as an appropriate executive to take AWS through the coming transition.

“Areas of change for AWS and the industry will include an increasing focus on delivering whole solutions to all types of customers, building many more first-party applications and partnering better with third-party software and solutions providers, and providing a greater amount of industry and functional specialization as platforms on AWS,” he said. “I am confident Adam will continue to focus AWS on being genuinely customer-obsessed.”

A new cloud architecture takes shape: “superclouds”

Although there continues to be a dramatic shift to the cloud by all types of businesses, nonprofits, startups, education institutions and governments, the reality is that the majority of IT remains on-premises today. IT budgets still haven’t fully transitioned to cloud yet as analysts estimate that only about 15% of IT spending has moved to the cloud. Massive growth remains to be had by all current providers. According to Selipsky, this shift to the cloud is accelerating, but it’s still early days. 

“We used to say that in the fullness of time, we believe that almost all workloads will move to the cloud,” he said. “Today, we see overall adoption continuing to speed up due primarily to the overwhelming value of the cloud’s many benefits [including agility, lower cost, scalability, elasticity and security] and in part to environmental forces, such as the pandemic, which has served as an accelerator in the move to the cloud by several years.”

Along with increased adoption of AWS, Selipsky is seeing increased sophistication from both longtime and newer AWS customers. According to Selipsky, “the world has changed, and customers and developers are expecting more from AWS.” AWS is developing a “new playbook” to raise the bar for its customers while changing the game on the competition.

Some investors, independent software vendors and entrepreneurs are predicting that we’ll see an acceleration of technology and business model innovation over the next 15 years, vs. the previous span. The rise of cloud computing has ushered in an era of technology innovation that is only accelerating. They see a shifting of focus from cloud apps to cloud platforms such as Snowflake Inc., Databricks Inc. and others that are now full-blown platforms and built on AWS, not just point SaaS applications.    

Today, any company can have a fully built-out platform in the cloud. Venture capitalist Jerry Chen of Greylock Partners calls this phenomenon “castles in the cloud,” but they might also be considered “superclouds” since they’re built entirely on cloud infrastructure.  

“Castles in the cloud is a large opportunity with 31 markets, such as machine learning and AI, and 173 further submarkets such as fraud detection and facial recognition,” Chen said. “No longer is there one giant market like ‘databases.’ There’s a database market and 30 or 40 submarkets that serve the needs of developers for a specific type of database that is specialized for the ‘castle’ they are building. Essentially, it has opened up a lot of white space for anyone to innovate within broad or niche areas.”

The success of Snowflake, MongoDB Inc., Databricks, HashiCorp and others illustrate how startups can grow rapidly by leveraging the capabilities of this new platform capability. “AWS is helping to drive this broad market opportunity,” Chen said. “The more that AWS grows and succeeds, the more opportunity it creates for the tech industry overall, and others are able to capitalize on it.”

Cloud: the next 15 years

Fifteen years into the cloud, AWS has a rich set of tools, lots of data and increasing plans to use artificial intelligence and machine learning to build solutions. Selipsky hinted that this is one way he intends to add incremental value for AWS customers. Specifically, AWS is looking to abstract its underlying foundational products, services and application programming interfaces to build solutions that use Amazon’s deep machine learning capabilities, so the company can tap new markets with purpose-built offerings that span industries horizontally and vertically. 

Asked if this might be a new technical model for AWS, Selipsky was unwavering in his answer that this is not new. He explained that AWS is moving fast to enable companies to build modern applications based on built-for-the-cloud services and are increasing their cloud investments by adding more scaffolding, infrastructure and AI capability — and AWS is clearly looking to escalate its commitment to help them with new purpose-built solutions.  

“I think you’ll see machine learning and AI baked into probably every other service that we have for customers,” said Selipsky. “If you look today, you can already see the early, but powerful, signs of that. Like in Amazon Connect, our call center solution, we’ve incorporated a primitive to do text extraction, a primitive to draw inferences from conversations that customer service reps are having. We expose these as high-level capabilities, and our customers don’t have to be technical geniuses to interact with them.”

So if you take Selipsky’s example and apply it across all of AWS services, then everything can be a platform, not a point solution or service. This is a new and potentially powerful kind of AWS flywheel, both for AWS’ customers and channel partners who can choose to add value on these platforms. Done at large scale, AWS could be unstoppable, because for AWS’ customers and partners, it’s a speed advantage in deploying new large-scale applications.  

“As cloud gets easier to adopt and lower-level services are replaced by smarter, faster and easier to use higher-level services, like we saw during the pandemic, we expect that this new capability makes adopting the AWS technology, infrastructure and purpose-built software easier for customers to seamlessly span applications across the cloud and the edge,” said Selipsky.  

New growth markets and the pandemic

The cloud is expanding globally. At the same time, AWS capabilities are being embedded everywhere — in 5G, in factories, on premises and in other remote and edge locations. Experts are finding that the rapid growth of cloud over the last 15 years, combined with the surge of machine learning and AI, is waking up slower-moving markets, such as public sector, health care, energy and other industry verticals that were originally slow to adopt cloud.

Because the pandemic is driving a new cloud shift toward purpose-built software, AWS is expanding the release of industry-specific capabilities. “It’s purpose-built, not custom,” said Selipsky. “We don’t want the cloud to be custom. That would destroy all the favorable economics and even impact things like operational excellence and probably security at the end of the day.”

Not surprisingly, companies are also looking for more remote and edge use cases. The pandemic magnified the idea that the cloud must move to where the people and applications are – at the edge. More and more use cases are emerging such as cloud in cars, cloud in factories, cloud in farm equipment and so on. AWS wants its cloud to be everywhere. “We are going to work aggressively on ‘internet of things’ solutions, on ML solutions and these horizontal use cases and applications as well as bundle it together in ways that are attractive to solving customer problems,” said Selipsky.

From on-premises to cloud

Selipsky is very careful to draw a distinction between what AWS refers to as “the old guard” and how Amazon thinks about so-called hybrid cloud.

“If you’re talking about a classical old-school, walled-off data center that all of our old-guard competitors have known and loved for so long, that still happens,” he says. “And that is truly not part of the cloud, of any cloud. It’s not part of any edge. And there’s still a vast sea of workloads that operate in that fashion. And of course, you see more and more companies moving away from that.”

But he acknowledged that the size of legacy infrastructure means it’s not necessarily easy to move. “That will take years for some of those workloads to actually become part of the cloud in any fashion, despite the velocity with which it’s happening,” he says. “[But] we continue to believe that in the fullness of time, almost all workloads will be in the cloud and not siloed off behind four walls.”

Data value drives new value

Over time, most believe that legacy systems, even those that plan to stay on-premises for some time, will need to address their siloed data. Some are even saying that without data sitting at the center of every company’s innovation strategy, there’s no path to viability in the future. According to Selipsky, the role of data has evolved from being an asset stored in some data warehouse or locked down for security reasons. Now, data is being intelligently leveraged to bring more value and competitive advantage to the organization.

“Data goes on a journey,” he said. “It’s not a snapshot. Data comes in from somewhere, and it might be from an industrial sensor, or it might be from web logs. It might be metadata on photos that are stored in the cloud.” But it has to land somewhere, he added, and that’s why AWS has an increasing array of databases. 

“[Data] will often land in a data lake and might go from there to a database,” he said. “Or, it might come into a database and then go to the data lake, and now you’ve got all sorts of capabilities to query, analyze and extract insights from that data.”

Selipsky believes that data is a powerful ingredient for competitive advantage. Data has to be freely accessible with software, and AWS is looking to be the preeminent cloud provider to power it as the data volume continues to grow beyond predictions. The data deluge is forcing businesses to transform how they effectively store it and create dynamic policies to govern the data to extract the value of insights, visualizations and predictions that impact business.  

“I think it gets back to the organization, not the developer, the organization setting up a governance structure that makes sense for it,” he said. “And some organizations will actually choose to be more restrictive about who gets access to what data, and other organizations are going to be incredibly permissive. And they’re going to say, let’s expose our data lake to everybody because I can’t predict what some associate product manager will go and dream up, and start cross-correlating different things out of my data lake. And then she’s come up with some amazing insight.”

Selipsky agrees with the premise that developers want to experience what I call “data as code” ethos similar to infrastructure as code — having machine learning and AI programmable and baked in from the beginning.   

“I think it’s happening now,” he said. “You will have developers who will be doing self-service analytics, who will be querying data lakes and databases, who will be doing so with drag and drop interfaces, who will be, in fact, applying machine learning without being machine learning experts, and more and more you’ll see us democratizing machine learning that becomes an analytics tool not for the professional practitioner.” 

AWS’ secret weapon: custom silicon

Amazon Elastic Compute Cloud or EC2, AWS’ second service after Amazon Simple Storage Service, or S3, provided in 2006, provides secure, resizable compute capacity in the cloud. It’s widely believed to be AWS’ largest revenue service offering, and it most likely comprises the largest portion of its profits, which last quarter accounted for all of Amazon’s overall operating profit, masking a loss in international retail. AWS offers more compute options than any other provider and increasingly is innovating new compute options on its own chips. 

In fact, industry watchers believe that one of the more prominent examples of AWS’ innovation agenda is the way it approaches custom silicon development to support various forms of compute.

AWS created the modern cloud business by giving easy access to virtual machines that could be deployed and spun down on demand. Early on, AWS tapped the Xen Project, an open-source “bare metal” hypervisor that enabled the concurrent sharing of hardware resources. At the time, AWS used x86 chips primarily from Intel for its Amazon EC2 offering. 

It soon realized that the overheads of using Xen were a problem and began working with Advanced Micro Devices Inc. to create alternative compute infrastructure. But it still wasn’t enough to satisfy the AWS engineers. So, in 2014, AWS began a partnership with Annapurna, a then little-known Israeli-based company specializing in the custom development of Arm-based chips. In 2015, AWS acquired Annapurna for a reported $350 million to $370 million. 

It could turn out to be one of the most productive acquisitions in the history of enterprise technology. The Annapurna acquisition spawned a new era of innovation for AWS — across both hardware and software at the virtualization layer — culminating in the form of the AWS Nitro System, which was introduced at re:Invent in 2017.

The Nitro System is AWS’ underlying platform enabling the company to deploy a range of compute and hypervisor offerings, where virtualized systems run more securely and at performance levels that are substantially comparable to running on bare metal. 

In the following year, AWS announced Graviton, its own custom chip based on the Arm architecture. Subsequently, AWS has introduced a series of custom Arm-based chips, including Graviton2, a custom machine learning chip called Trainium, and Inferentia for deep learning workloads. This is a technical advantage so far unmatched by AWS’ competitors. 

Although AWS stresses the importance of its relationship with x86 and GPU chipmakers Intel Corp., Advanced Micro devices Inc. and Nvidia Corp., increasingly, customers are adopting EC2 instances based on AWS’ in-house silicon. SiliconANGLE reported earlier this year that in 2020, nearly half of new EC2 instances were based on Graviton2, AWS’ custom chip. That percentage could increase in the coming years. 

“We continue to see rapid adoption from our customer base,” Selipsky said. “We put it underneath other AWS services as well, so you get better capabilities, whether it be in Amazon RDS or AWS Lambda or other AWS services.”

The CEO said AWS will continue to iterate rapidly on general-purpose chips but also on more specialized chips. That being said, it’s clear that AWS sees the silicon advantages of faster performance at lower costs, while also being able to bring new, specialized chips to market faster. If AWS is left uncontested in this area, then ISVs will flock to AWS cloud over the competition. It’s one of the company’s secret weapons and its software developers, and it ultimately wins the ecosystem.

All this adds up to the reality that by developing its own custom, Arm-based silicon, AWS is on an innovation curve that is probably one of the fastest in the industry. It’s no coincidence that Tesla and Apple have similarly chosen to develop custom silicon based on the Arm architecture. These firms design their own chips and outsource manufacturing to fabricators such as Taiwan Semiconductor Manufacturing Corp. and Samsung Electronics Co. Ltd. Microsoft and Alibaba have also followed AWS’ moves in this direction, choosing to design their own Arm-based silicon.

But in silicon, experience matters, so the sooner you get up the learning curve, the more benefit accrues. And Amazon was first in the cloud business, well ahead of its competitors. Six to seven years later, AWS’ main hyperscaler competitors are following suit, though Google has also had its own tensor processing unit chips since 2016.  

This in-house chip design and manufacturing enables a much faster time-to-market compared with traditional general-purpose chip cycles — 18 months versus four years — and a performance improvement curve that is at least double or triple that of traditional Moore’s Law, in which performance doubles every 24 months. Given these new cycles, Graviton3 and other Arm-based silicon can’t be far away. 

In addition, with this early start, the AWS Partner ecosystem and customers are already building applications to take advantage of AWS’ higher-performance and lower-cost processing, making their offerings more attractive and thereby conferring advantage to their customers running on AWS. 

“We have enough scale that it makes economic sense for our customers, and for us, to have purpose-built silicon for different use cases,” Selipsky says. “So even if a use case is only relevant to a very small percentage of our customer base, that’s still going to mean many thousands of customers.”

When it comes to serving software developers, it’s been proven in every inflection point where new software models emerge that the platform that attracts the best developers, architects and software developers is the one with the fastest and lowest-cost processing system of the software and data from the application developer. 

If you’ve seen this movie before, you know that it doesn’t end well for most. AWS’ differentiated custom silicon such as Graviton gives software developers, including AWS partners and customers, the ability to build faster applications that take advantage of the higher performance and lower cost processing. 

To top that, AWS is now giving cloud buyers purpose-built AWS solutions if they prefer that, and their developers can continue to build unique cloud applications from foundational AWS services — or they can do both.  

The next-gen cloud has arrived, and there’s a new sheriff in town. 

His name is Adam Selipsky.

Look for the full interview with Selipsky in four installments starting Monday in SiliconANGLE.

Photo: Tableau/kareni/Pixabay