For the first decade of Amazon Web Services — call it AWS 1.0 — Amazon.com Inc. devoted much of its time to proving the cloud could work. The companies that fully embraced AWS as a platform tended to be startups in technology and media, which had the business models, processing requirements and skills that aligned perfectly with AWS’s platform. But for the majority of enterprises, AWS was the Costco of computing, where you went for bulk storage or bulk compute for application development and testing, or dev/test.

Wikibon attended AWS re:Invent 2016 in Las Vegas last week in some force. Our take? It’s AWS 2.0. AWS proved the cloud can work for your enterprise: Your workloads. Your big data. Your complex business problems. We encountered a thriving ecosystem that, while still relatively adolescent in its structure, is poised to make AWS an increasingly attractive option for even the most rigid, demanding shops. AWS 2.0 will continue to expand as an industry force because:

• AWS is big and fast. Bigger is slower, right? Not for AWS. Execs told us the company adds the equivalent of a Fortune 500 data center’s capacity every day. Every. Day. How? By vertically integrating. AWS is designing itself from customer experience back to hardware. It’s a risky undertaking, but it makes technical sense. They’re operating at unprecedented scale, and as James Hamilton, an AWS distinguished engineer, told us, “Any software that has to operate at our scale, some part of it is going to move into hardware.”

• AWS is going after big business problems. Bulk storage or compute? That’s passe. AWS 2.0 is going after big enterprises (and their legacy workloads), big data (especially artificial intelligence), and big, hairy problems (such as IoT). AWS discussed more than 20 new services or functions to their already pretty broad service portfolio, many in decidedly non-Costco-like domains such as workload migration and AI. The AI story is especially compelling. Amazon, AWS’s parent, is an unqualified leader in the commercialization of advanced AI-related technology. Packaging consumer services like Alexa for the enterprise, such as the new AWS “Lex” service, is a phenomenal win for a lot of businesses.

• The AWS ecosystem is full of potential. AWS is moving fast, but their ecosystem is stepping up, too. Start-ups like ATADATA were demonstrating software tools that could be used to mostly, and quickly, move shops from on-premises to AWS in Linux-to-Linux environments. This is very hard stuff, but AWS is providing a target that gives a lot of smart people a much better palette for innovation than the usual product-based stacks of hardware and system software.

Plenty of Challenges Remain

Future Wikibon research will explore the limits to AWS’s growth in some detail. Here, we’ll note a few of the challenges that not surprisingly didn’t make it onto to the agenda at AWS re:Invent this year, but will powerfully shape enterprise IT decisions going forward:

• Developers want more than virtual machines. While AWS is touting its support for modern software development tools such as containers, the cloud industry is starting to bifurcate specifically around the needs of developers. Microsoft Corp., in particular, is positioning Azure as the developer-friendly platform, presenting its significant portfolio of tools and apps to developers seeking to build the next generation of people-friendly applications. Oracle hopes to take it one step further, turning its suite of enterprise “clouds” into services. These are big companies with big money and big brains. If they, too, can speed up, they are not going to roll over for AWS.

• The system of record still matters. If there is a firebreak constraining the AWS conflagration, it’s systems of record and the databases they run on. Our clients — and many re:Invent attendees — still tell us that they’re skittish, at best, about putting personal and private information and financial systems in the cloud. Is this for technical reasons? Not really. Rather, it’s because brands blow up or execs go to jail if they have big problems with these classes of systems. Social innovation often is more important than technical innovation, and AWS has a bit of work to do on that front.

• Sometimes it’s better to extend the cloud than move the data. Until recently, Wikibon has been the lone voice in the woods about IoT edge computing, but we’re right. Our research shows that, given current circumstances, as much as 95 percent of IoT data will stay at the edge. AWS’s response? Very clever bulk data movers, such as its Snowball storage appliance, that let you physically ship huge amounts of data back to the cloud at an effective speed about equal to fast Internet connections. Snowball has been enhanced with enough compute to run a new service, Amazon Greengrass, which embeds AWS’s Lambda event processing service. It’s an important start, but the IoT edge is going to be a determining factor in setting the industry’s direction and AWS has more to do.

AWS Still Has To Learn: Leadership Is Hard

One final thought:

The show floor of this year’s re:Invent was dominated by a central AWS pavilion (above) that was unlike any I’ve seen at a trade show before. It was a two-story building, built inside the Sands Convention Center, that featured seating all around the periphery of the second level. People could be seen peering down upon the show floor, intently watching the proceeding of the customers and partners below. At times, it cast a benevolent presence: the new captains of the tech industry, steering us through the cloud transformation. At other times, the AWS folk looked like guards, ironically dressed in orange, constantly monitoring the yard.

That last metaphor may sound harsh, so let me explain. AWS’s quest for scale and speed is delivering significant customer benefits, especially in the form of customer cost savings and simplified delivery of IT infrastructure capabilities. But the quest for scale sometimes has the effect of limiting options. As digital becomes a more significant feature of customer experience, some partners and customers that we spoke with at re:Invent were starting to worry about AWS subsuming their technology segment or value proposition. Real or imagined, it’s a concern that AWS will have to address.

For example, AWS collects huge volumes of operational data. It uses that data to make impressively good business, design and service development decisions, among others. How will it share that data with partners and customers? Will they get operational data that can help with decisions about loads, service consumption and other mainly operational issues, like they do today? Or in time will they get deeper insights that could lead to superior business opportunities?

At AWS re:Invent 2016, AWS picked up the mantle of industry leadership. As IBM Corp., Microsoft and others have learned, leadership in the tech world doesn’t come without cost. Followers are demanding, especially regarding something as important as their digital capabilities. What will AWS’s leadership challenge be? Not all partners and customers can move with the same alacrity as AWS. As the company’s community gets more diverse, AWS’s approaches to partnership will have to become more nuanced.

We bet that will be an important element of AWS re:Invent 2017.

Wikibon analyst Stu Miniman, SiliconANGLE Media Co-Chief Executive John Furrier, and Jeff Frick, general manager of theCUBE, provided their take on AWS on the last day of re:Invent:

(* Disclosure: TheCUBE was a paid media partner for the show. Neither Amazon nor other sponsors has editorial influence on content on theCUBE or SiliconANGLE.) 

Photo by Robert Hof