If you could spend $1 to make $10, you’d do it, right? By the same token, if you could spend $9 billion to make $90 billion, you’d do that, too, right? With hundreds of billions at stake in cloud computing, and winner-takes-most economics in play, Amazon Web Services, Microsoft, and Google spent $35 billion last year—and roughly $10 billion in the first quarter of 2018—to earn the right to take home multiples of that later. “The cloud isn’t cheap,” rightly reasons Geekwire’s Tom Krazit.
Failure in cloud, however, is much more expensive. Given the rise of serverless computing, it’s getting ever more costly, with some signs that AWS could turn its early lead into long-term dominance.
What all that cloud infrastructure money buys
As cloud pundit Bernard Golden highlights, the copious quantities of cash that AWS, Microsoft, and Google are plowing into datacenters reflects the magnitude of the payoff for winning: “This investment pattern reflects the realities of the cloud computing market: It’s a platform-based industry with enormous network effects that requires sufficient capacity to support exploding, spiky demand—much of which can emanate from geographies that require local infrastructure. That’s a recipe for needing to plow huge amounts of money into the business.”
We’re not talking about a few billion dollars up for grabs. We’re talking about hundreds of billions in play, as the future of enterprise computing gets defined today.
For those that remember how the search wars were won (and lost), capital expense (capex) was a big component. Google not only had the best algorithms, it also invested the most to ensure it had adequate capacity, as Expedia engineering executive Subbu Allamaraju stresses. Going cheap on capex turns out to be really expensive.
AWS is poised to dominate serverless computing
One of the more potent avenues for reaping a return on these investments is serverless, precisely because it locks enterprises more deeply into their respective clouds. As evidenced from the huge capex investments, “serverless” is a misnomer: Scads of servers sit behind services like AWS Lambda. Such serverless functions invite enterprises to write business logic atop the server infrastructure provided by the big three cloud vendors.
Getting serverless right, from the vendor perspective, involves billions in both hardware and software innovation. AWS, Microsoft, and Google are spending those billions; everyone else is toast.
But even among these big-spending cloud behemoths, AWS stands out. With its two-year head start and its relentless focus on helping enterprises to adopt this new computing paradigm early and often, AWS has the broadest, deepest array of serverless functions to choose from, all working together to improve enterprise productivity—and dependence on AWS.
It’s not nefarious; AWS’s approach is breeding willing dependence on Amazon to take the brunt of infrastructure innovation so that enterprises can focus on business logic innovation.
So far, it’s working, and it’s keeping all competitors at bay. The Cloud Native Computing Foundation’s survey data suggests that AWS Lambda owns 70 percent of the burgeoning serverless market, prompting industry observer Simon Wardley to note: “Let me translate that for you. Amazon is currently positioned to own 70 percent of the future of all software.”
If that causes you to do a double take, Wardley makes it clearer: “As Amazon’s serverless ecosystem grows, the more metadata it can mine, the faster its rates of innovation, customer focus, and efficiency. Once it gets to around 2 percent of the market then it’s game over for all those not playing at scale.” You can quibble with that 2 percent threshold but not with the idea of a serverless virtuous cycle of data, scale efficiencies, and customer lockin is harder to challenge.
On that last point, AWS doesn’t think of it that way, of course, but AWS Lambda is becoming more like an operating system than anything else. As AWS’s Jeff Barr describes, “People started to think of [AWS Lambda] as this nervous system they could connect up to the incoming flow of data into S3, to message queues, and to notifications that are wired into different parts of the AWS infrastructure.”
Happy as developers were to assemble their infrastructure from AWS’s broad menu of services, serverless computing makes it even easier by hiding the complexity. As companies embrace this convenience boost, they give up a little latitude to move to rival clouds or back to their own datacenter.