Amazon’s cloud business just recorded its weakest growth to date | Hacker News Reader

Amazon’s cloud business just recorded its weakest growth to date

cnbc.com

91 pointsbelter38 comments

>Amazon said Thursday that revenue growth in its cloud-computing unit slowed in the third quarter to 27.5%.

27.5%. It is lower that their previous 33% over the past few years, but at the current size of AWS growing 27.5% is still ridiculously good. To put this in perspective, if AWS continues to grow at 33% in 2022 and 2023. Then the whole 2023 33% growth alone, would equal to the size of the entire AWS in 2018. It is not the first time Amazon said they are limited by how fast they are building out Datacenter and getting hardware resources ready.

It will be interesting to see further details given out in AWS re:Invent 2022. Especially on Graviton roadmap.

It's also worth noting that a year of growth of "just" 27.5% after growing 33% the prior year, the 27.5% represents more growth-in-dollars this year than the 33% did last year.

With a growing expense. It’s not so clear cut math.

Weakest growth. Crazy. Apparently growth needs to growth. Every year. More. Very sustainable, I'm sure.

Accelerated acceleration or your company is a failure. I wish my company had the problem of only 24% growth.

"Keeping up with the Bezoses" is eating the world...

*listed company

Exactly this. It’s not enough to take over the world. Growth isn’t enough. The rate of growth must grow as well.

I was happy to see another buying opportunity, but I definitely rolled my eyes at Wall Street’s overreaction.

How long until the growth of the growth is expected to grow quarter to quarter?

Well that’s a derivative joke

"In the fall of 1972 President Nixon announced that the rate of increase of inflation was decreasing. This was the first time a sitting president used the third derivative to advance his case for reelection." - Hugo Rossi

https://www.maa.org/press/periodicals/convergence/quotations...

It's gotta grow at a rate of e^x so that the rate of growth is also the rate of the rate of growth and also is the rate of the rate of the rate of the rate of the rate...

The trillion-dollar question is whether the slower growth rate is just a temporary blip, or the beginning of a paradigm shift.

I don't think anyone knows for certain, but anecdotally, I have found there's (1) more skepticism about the cost savings promised by cloud providers and (2) a growing realization that the financial incentives of cloud providers are not aligned with those of customers.

In the early days of cloud providers, their services were both easy-to-use and cost-effective for many growing companies. Now, cloud providers offer what I'd describe as a "giant hairball" of specialized services, built on top of other specialized services, built on top of yet other specialized services... such that no single person can comprehend all of them, let alone manage them cost-effectively. Moreover, as all these intertwined services become deeply entrenched in IT infrastructure, migrating away from them becomes incredibly disruptive. The more a customer is sucked and trapped inside those giant hairballs, the better for the cloud providers. To paraphrase the Dark Knight, cloud providers may have been the heroes at first, but they have lived long enough to become the villains.

I think AWS itself has actually eased up on the "cloud saves costs" story, shifting more to "be more agile in how you spend" (see for example [1]). Arguably one of their biggest advantages over regular hosting is that deploying resources is as far removed as possible from any mention of payments, thus skipping the procurement processes that usually occur in companies.

That way around internal red tape keeps attracting large companies, and the free credits keep attracting startups. I don't think this dip is anything more than a readjustment to a post-pandemic world, where more happens offline and thus slightly less online.

1: https://docs.aws.amazon.com/whitepapers/latest/aws-overview/...

I'm hearing you say "a large catalog of services can be hard to navigate" (true), "dependencies can be hard to manage" (true), and "dependencies can be hard to migrate away from" (true). What I'm missing is how these considerations make a platform a villain.

Their incentives are not aligned with those of customers. For example, anecdotally I've heard that many cloud services are configured with (a) out-of-the-box defaults that make it deceptively easy to get bigger invoices than anyone ever expects, and (b) byzantine billing that makes it deceptively hard to figure out how to make those invoices smaller.

> or the beginning of a paradigm shift

Could just be reaching market saturation.

I agree with this. When organizations move their stuff into the cloud, they justify it as driving down complexity and reducing cost -- we don't need to rent space for a data center, hire people to build and maintain that stuff, and so on. But the costing models seem to assume that other burdens like billing and cost management will be handled by the current in-house team.

Except the billing for cloud services is byzantine and entirely unpredictable. So the org hires new specialists to manage this, and then new IT folks to setup and run some kind of CMP, and then they end up in a confused cost spiral that nobody can really explain but it certainly isn't saving anything.

The developers meanwhile really just appreciate all the prebuilt lego bits and relatively simple scaling which they could get in theory and on-prem, avoid the cost spiral, and get way better visibility into what costs what, where, and why.

The cost models get even more wrong when the org tries to go multicloud because its like nobody who builds those models understands egress costs.

> Except the billing for cloud services is byzantine and entirely unpredictable.

I think AWS is the biggest offender in this area, by not enabling per resource usage by default.

Aren't all of these indictments of CTOs/COOs, rather than cloud providers?

Yes. But I think in a sense, they're also somewhat handcuffed by the industry. Talent wants to work "in the cloud", and if you don't offer that you restrict your talent pool significantly. A big part of creating a desirable workplace is creating one in which people want to be hired into and in tech that really just means "the latest tech" regardless of it being a good idea or not.

I think it's also part of the reason we're doing so little with so much in terms of efficient software. "The latest tech" also just means huge abstraction layers which eat up tons of compute but offer only easier development at the cost of much more expensive O&M. It's honestly part of the drive to the cloud as well as what we used to do on a few boxes with software written in C or a mainframe is now a scalable cloud architecture running mostly Python or some such.

Every time I look at cloud offerings I can’t help but remember, “there’s money in confusion”.

Right there at the top

> Amazon said Thursday that revenue growth in its cloud-computing unit slowed in the third quarter to 27.5%, missing analysts’ estimates. It’s the slowest expansion since at least 2014, the year Amazon started reporting on the group’s finances.

Really these kind of 'weak revenue based on _predictions_' articles are getting ing ridiculous. That's still a crazy high growth. AWS has better cloud competition than ever before, cloud computing is nearing fuller market penetration, and 27.5% growth is "weak". It's only framed as "weak growth" if you compare it to unsustainably crazy fantastic growth the last decade. At a certain point everyone that's likely to move to the cloud will be in the cloud.

Yup, at biggest, this slowing of the rate of growth could be some kind of market saturation being approached, but it will take many quarters to establish such a trend.

More likely just a completely ordinary variation, with some headline writer desperately trying to hit his click numbers...

My take-away from "missed analysts predictions at 27.5%" is that analysts need to make predictions more firmly based in reality, rather than AWS needing to grow more.

Better title: AWS records more than 25% growth for third quarter in a row

That title is better for understanding, but is not better for CNBC.

Both titles are incomplete. It's useful to know that AWS is still growing whole integer percentages and that those percentages seem weaker than normal.

Titles about second and higher derivatives rarely feel complete or accurate.

Massive jerk in Amazon's cloud business as revenue growth rapidly decelerates! Is this the revenue jounce competitors have been waiting for? (/s)

Ready for sensational headlines about the concerning reduction of the deceleration of AWS' decrease in YoY growth. Will they ever be able to turn this ship around?

That doesn't come close to telling the whole story either though. Namely two things:

1. All of Amazon's profit last quarter, and then some, came from AWS. Declining revenue growth in that segment could have an outsized effect on earnings/losses.

2. Amazon's PE ratio is still an eye watering 90+ much of which is predicated on continued (perhaps even accelerating) growth. Unless they pull a rabbit out of their hat with new sources of income or cost improvements, the market cap of the company is going to continue to depend a decent amount on growth within AWS or that PE is going to look more and more detached from reality.

AWS still has a great unrealised revenue backlog which are not on the statements for the current year/quarter (from long-term commitments).

$104.3 billion as of September 30, 2022. [0]

[0] https://s2.q4cdn.com/299287126/files/doc_financials/2022/q3/...

From the perspective of a small business person complexity is a big deal. Amazon has amazing services with more than a hundred cloud products for tuning every aspect of a web site. But it is possible to get something going on Linode or Digital Ocean without all the fancy unnecessary options and a base price that is half to a third the cheapest AWS service level available. There is a big and long tail that goes small and AWS is increasingly too high altitude for small scale applications.

I'd argue that small scale apps are pretty great on AWS, because of the "always free" tier of AWS:

- 1 million Lambda invokes, 3.2 million seconds of compute time

- 25 GB of DynamoDB storage, 25 provisioned read and 25 provisioned write units (about 200 million reqs/month)

- 1 TB of CloudFront bandwidth

You can run a good sized service completely free on AWS forever.

The sweet spot for Linode and Digital Ocean pricing is when you want to run a monolithic app hosted in a VM, but for small scale apps they may not even need that yet as they haven't reached the scale where they can fully utilize a VM, nor have they reached the level of traffic where the per request cost of a serverless architecture starts to feel expensive

If you can fit your app into a serverless architecture and take advantage of the always free tier of AWS then that's a super low cost (even free) option.

What surprises me is that Amazon's market cap is back to a pre-covid level.

The market things the whole Covid thing after all did not boost the value of Amazon?

And neither does all the newly printed money? During Covid, the monetary base doubled:

https://www.federalreserve.gov/monetarypolicy/bst_recenttren...

Anyone else moving away from the cloud? From AWS to the utility room server :)

In my personal projects? Absolutely. It's fun and satisfying to build something on a server sitting in the corner of my room, something I can actually look at and kick.

For professional projects? Only if given a _very_ compelling reason.

Reply on news.ycombinator.com