Cloud Profits May Be Slowing at Microsoft and Amazon (yahoo.com) 69
"Once-booming demand for cloud-computing services is slowing..." reports Bloomberg. "When Microsoft and Amazon report results next week, analysts are anticipating the slowest revenue growth for their cloud-computing businesses since the firms started breaking out performance last decade."
For years, demand for cloud-computing services has steadily driven growth at both Microsoft and Amazon... Microsoft's Intelligent Cloud unit, which is home to its Azure cloud-services business, accounted for 38% of its revenue and 39% of operating income in 2022. Amazon Web Services was the fastest-growing of the Seattle-based company's major businesses last year and generated $22.8 billion in operating income. The rest of Amazon's businesses combined posted a $10.6 billion operating loss.
For both companies, cracks are starting to appear. In the first three months of 2023, growth for Microsoft's Azure unit and Amazon Web Services is expected to fall to 31% and 14%, respectively, excluding currency fluctuations, according to the average of analyst estimates compiled by Bloomberg. A year ago, Azure sales expanded 49% and Amazon Web Services 37%.
In a shareholder letter released last week, Amazon said AWS "faces short-term head winds" related to the economic backdrop that will "soften" the growth rate. This echoed what it said in its most recent results. Microsoft also warned of a slowdown in cloud software sales last quarter. Wall Street has been getting more cautious. UBS lowered growth estimates for Azure last week, warning "customer efforts to optimize/trim their cloud spend will be deeper and last longer than most think...." Jefferies [financial services company] sees slowing cloud demand as "a key concern" for Amazon. Analyst Brent Thill said that because AWS generates so much of Amazon's operating income, "a stabilization in cloud is crucial for shares to outperform."
For Alec Young, chief investment strategist at MAPsignals, Microsoft and Amazon remain attractive despite the slowdown, which he expects to be a temporary pause before growth re-accelerates. "There's still a lot of runway ahead for cloud computing, so I don't think investors should obsess too much over the level of growth over a couple quarters," he said.
For both companies, cracks are starting to appear. In the first three months of 2023, growth for Microsoft's Azure unit and Amazon Web Services is expected to fall to 31% and 14%, respectively, excluding currency fluctuations, according to the average of analyst estimates compiled by Bloomberg. A year ago, Azure sales expanded 49% and Amazon Web Services 37%.
In a shareholder letter released last week, Amazon said AWS "faces short-term head winds" related to the economic backdrop that will "soften" the growth rate. This echoed what it said in its most recent results. Microsoft also warned of a slowdown in cloud software sales last quarter. Wall Street has been getting more cautious. UBS lowered growth estimates for Azure last week, warning "customer efforts to optimize/trim their cloud spend will be deeper and last longer than most think...." Jefferies [financial services company] sees slowing cloud demand as "a key concern" for Amazon. Analyst Brent Thill said that because AWS generates so much of Amazon's operating income, "a stabilization in cloud is crucial for shares to outperform."
For Alec Young, chief investment strategist at MAPsignals, Microsoft and Amazon remain attractive despite the slowdown, which he expects to be a temporary pause before growth re-accelerates. "There's still a lot of runway ahead for cloud computing, so I don't think investors should obsess too much over the level of growth over a couple quarters," he said.
Is E-Bay still there? (Score:3)
If so, a quick search for "World's Smallest Violin" would be useful here.
I don't get this "slowing growth" issue (Score:5, Insightful)
Although I have a good grasp of economics, stock market etc, I cannot understand how growth, just slower growth is considered such a disaster for a company. Everybody and their mother have moved to the cloud, whether it was a good idea or not (spoiler: it often was not), so the market is a bit saturated compared to years ago, not as many businesses left to move to the cloud. And existing businesses don't have crazy growth rates to be increasing their cloud footprint at 30% yearly rates either. Not to mention we are talking about Amazon who owns the biggest market share for this huge, although probably a bit saturated business, so a 14% growth is still a huge number.
I mean, the population is finite, human activity is finite, no business can keep a tremendous growth rate forever, does that mean that every business is doomed?
Re:I don't get this "slowing growth" issue (Score:4, Informative)
I mean, the population is finite, human activity is finite, no business can keep a tremendous growth rate forever, does that mean that every business is doomed?
According to current ecconomic doctrine, yes. Yes, it is that abysmally stupid.
Re: (Score:3)
Growth now is going to be new innovation. Different market focus. New customers that previously did not seem value. In other words the more expensive customers.
Re: (Score:1)
Re:I don't get this "slowing growth" issue (Score:5, Informative)
It's not a disaster for the company, it's a disaster fro the stock price that was based on the higher growth lasting just short of forever.
Re: (Score:3)
Re: (Score:1)
The problem is too much money companies generate are from Stock prices. Not by actual profit, at least directly. A lot of people buy stock in fast growing industries, so slower growth makes it less appealing stock options.
Basic accounting has Assets= Liability+ Share Holder Equity. While Share Holder Equity is really the same as a loan, which would be a liability, it isn't considered the same.
Re:I don't get this "slowing growth" issue (Score:5, Informative)
Its that the companies value won't continue to grow at the same rate so all the people who assumed it would will remove their cash and the stock price will fall. As a shareholder you expect a certain percentage rate of return.
Stock prices are not based on the actual value of the company but what people think it will be worth. CEOs get paid in stocks so once again terrible for them.
Its all speculation and contributes nothing to the productive economy.
Re: (Score:3, Interesting)
Although I have a good grasp of economics, stock market etc, I cannot understand how growth, just slower growth is considered such a disaster for a company.
It seems you do not have as strong a grasp of stock marklets as you think.
There are many ways to estimate the value of a company. But, academically, the most accepted one is the "Discounted Cash Flows" Method.
This method calculates (using various techniques that depend on the taste of the person doing the calculation and the nature of the company) and a set of assumptions, all the money (earnings) that the company will produce during a reasonable "forecast time horizon" (typically 10 years) + another number
Re: (Score:2)
I don't have anything big to disagree with your points, but to expand upon it, there are some useful ways to reason about DCFs in order to understand the implications of this news.
Firstly, you might consider the rule of 72, which gives an approximate doubling time for a given growth rate. This gives a sense of how different growth rates are non-linear, a 1% growth rate will take 72 years to double while a 2% growth rate will take 36 years to double.
GDP growth in the modern era is often 1-3% a year and has
Re: (Score:1)
Re: (Score:2)
Re: (Score:2)
I mean, the population is finite, human activity is finite, no business can keep a tremendous growth rate forever, does that mean that every business is doomed?
On a long enough time scale, we're all dead...
Re: I don't get this "slowing growth" issue (Score:2)
Because in capitalism, the suits n ties are always looking for the bucks they think they should be making and not at the bucks they are in fact making.
So a saturation means to them they just have not shoved it in enough faces - and that they quickly need to find new faces to get on board.
Never auch a thing as it should be enough with these people.
Have you done the math? (Score:5, Informative)
"cloud services" are expensive. Like we compared buying our own server over a5 years period. Microsoft cloud was the most expensive then owning your own hardware, and hiring out maintenance, then a "local"cloud provider who also did maintenance .
That worst part is. Cloud services are fees and locally owed hardware is assets that can be deducted off of taxes a little every year.
Re: Have you done the math? (Score:2, Insightful)
Well yes, that should be obvious. Even if you prepay AWS hardware 3 years in advance, you still pay as much as buying the hardware outright.
The main advantage of these solutions is that you can have the hardware pretty much instantly.
Re: (Score:1)
The problem is it's not just the hardware, a server by itself isn't of much use - you typically need networking, connectivity to the outside world, security (both physical and digital), staff to physical install and maintain these things and so on. Once you bundle in all those costs you can't remotely do it cheaper than the cloud - this becomes even more true if you need to distribute across regions for redundancy or security, you just can't come close to achieving that yourself for a similar cost, in fact,
Re:Have you done the math? (Score:4, Insightful)
But the bosses were sold on the idea that it is cheaper. Just like the Idea of having homemade software, is more expensive than some huge Enterprise solution. The companies that sell these products push how so much liability is taken off your shoulders. However having a crack team of lawyers to make sure any contract are so one sided, that any issues will fall on their customer.
For an org of substance, having your own data center, with a dev team will be cheaper, and have better running products. But bosses are convinced that their staff is actually a bunch of idiots.
Re: Have you done the math? (Score:2)
Okay we are replacing a windows 2012 server running ms sql 2012
With a new ms server running windows something and ms sql 2012 compatibility mode
Because our erp software hasn't upgraded the backend support for something else. The front end has shifted entirely to a web based front end and duplicates the old applications UI only slower cause it's local web based.
However support for the old system ended a couple of years ago and we have been limping along.
Your idea sounds great. It won't happen until compati
Re: (Score:3)
That worst part is. Cloud services are fees and locally owed hardware is assets that can be deducted off of taxes a little every year.
I am not an accountant, but I don't think this is really accurate for the tax implications. When you buy hardware that is a capital expense and the business can write it off via depreciation over time, however cloud computing is an operating expense which you can write off immediately. At least that is my layman's understanding.
Cloud sucks (Score:5, Interesting)
As someone that went through an "upgrade" where we moved from an on premise ERP system to a cloud system I can confidently say that cloud sucks. Why?
1) The system we replaced allowed us to fully customize the application in any way we wanted using a delivered scripting language. The new cloud system? No customizations allowed. You are limited to configuration changes that may or may not suit your needs.
2) In the old system the vendor would deliver fixes and feature updates that we could apply - or not apply - when it suited us. The new cloud system? The fixes are applied to our database instance on a schedule that the vendor sets and cannot be changed without an act of God. By the way, every other Tenant that shares our instance also receives the same fixes on the same schedule.
3) Reduced functionality. Apart from the inability to customize the cloud application the on premise application that it replaces had broader and richer functionality. Eventually the cloud version might catch up but right now it's behind.
4) Pricing. We went from an on premise system that had a straightforward maintenance contract (18% of the software cost) to a cloud system that seems like it costs less up front but actually costs more in the long run. They build in complex rules around how many API calls you have and how much network traffic you have. You have to structure your interfaces very carefully in order to avoid tricky scenarios that will have you paying through the nose.
5) Issue resolution. With the old system we had experts in house that could troubleshoot issues and often fix them without any need to contact the vendor. At the very least, we could put in our own temporary fix while we waited on a permanent one from the vendor. With the cloud system we are at the mercy of the vendor. The code behind the scenes is a black box that we can neither view nor modify. Fixes also take longer because the vendor has to ensure that it won't break any of the tenant instances that share our instance.
Some companies are starting to figure this out and are going back to on premise systems, where you control the data, you control when patches are applied, you control when the system goes down for backup etc.
Yes there are benefits, such as being able to scale up storage and memory quickly. But in our case those benefits are negated by the reduced functionality, cost and issue resolution policy.
Re:Cloud sucks (Score:4, Insightful)
Well, yes, some people are finally figuring out that if you are large enough to run your own or you have somebody local doing it for you, the cloud is basically a rip-off. Of course, Microsoft does not want you to own your own infrastructure, because that brings them less money. Hence I expect the next move from remote computing to local computing may be a bit more difficult. Still, this idiotic hype-cycle has now been running for more than half a century and there is not reason to expect it going away.
Re: (Score:2)
Well no kidding. If you are operating at any real scale it's almost always going to be significantly cheaper to roll your own "private cloud" than to pay a 3rd party to run your applications on their hardware. Public cloud is awesome if you're a small startup that doesn't want to invest in a bunch of infrastructure, lack the staff to manage your own infrastructure, or if you simply want to "burst" into pub cloud when you have temporary capacity needs. There's lots of the first two types and a vanishingly
Re: (Score:2)
Indeed. It is also interesting that European regulators for financial institutions require a cloud-replacement strategy or re-insourcing strategy for all critical IT infrastructure. This is really no accident. On the plus-side, running a small "private cloud" for yourself is as easy as it was never before and if you are smart and did not use vendor-spefific features, you can deploy the same infrastructure on your own hardware with reasonable effort.
Of course, MS does not want that. So "cloud AD" is not the
Re: (Score:3)
You have to be very careful. SaaS solutions can you suck for the exact reasons you stated.
Going either IaaS or PaaS are totally different.
Re: (Score:1)
Re: (Score:2)
Some companies are starting to figure this out and are going back to on premise systems, where you control the data, you control when patches are applied, you control when the system goes down for backup etc.
As an ERP technical consultant, I completely agree with everything you said. Except this bit. What on-prem ERP systems are available today new that offer the functionality of these older systems that were replaced? Of the big systems that I am aware of (and I will happily acknowledge my ignorance here and look forward to learning something new), namely the Microsoft Dynamics (formally Axapta) range; the Oracle systems and the SAP systems, none of these offer on-prem any more. Everything is cloud.
Sure, Micro
Re: (Score:3)
PeopleSoft is the on premise system I was referring to. Oracle is pushing hard to replace it with Oracle Cloud. Apparently the salespeople won't even show it to a prospective customer unless they ask for it. Some customers want it but Oracle makes more money from Cloud products so they push that. When Oracle did a hostile takeover of PeopleSoft they inherited the codebase and at first threatened to get rid of it until customers complained enough to convince them to keep it. To their credit they have added o
Re: (Score:2)
Epicor's Kinetic ERP still has on-prem as an option. They do heavily push towards their cloud offering (which is the same code base), but they have found too many of their customers resist going to the cloud for them to try to make it mandatory.
Re: (Score:2)
"Functionality in many cases is actually better in the cloud versions rather than the on prem ones as the on prem systems are now frequently lagging." - Not for Oracle Financials Cloud. They limit you with a fixed number of something called Descriptive Flex fields for creating your Chart of Accounts. They don't provide enough and there is no way to expand it. We have run into issues with reports. If the output is above a certain size it simply times out and won't run. We literally have to break up the repo
Re: (Score:2)
On premise? Why not in a data center (with colocation of course)?
Re: (Score:3)
None of these sound like cloud problems, they sound like the company picked the wrong system.
Re: (Score:2)
The clouds are not recession-proof (Score:4, Informative)
Hyperscalers are very flexible but they come at a high cost. Even if you leverage all the possible discounts, negotiated deals, savings plans, etc, it's still probably going to be more expensive than on-prem AND there will be a high risk of overspending.
On-premise is not as flexible as the clouds but it has some advantages: it's sunk cost and you know exactly how much the amortization and operation is going to be each year, and there's no risk of developers going crazy and terraforming a humongous architecture that will cost you a fortune.
My prediction is, if recession persists, many businesses will:
1. Downscale on cloud
2. Downscale on cloud and add something on-prem, thus being hybrid cloud
3. In the most extreme cases, where cloud cost control has not been achieved, move completely to on-prem
Cloud is a tool, not an evolution, like offshoring (Score:5, Insightful)
Every company in my area was tripping over themselves to migrate off on-prem to cloud...and my company spent a FUCKTON of money to get many of our servers off the cloud...and our bills increased...the servers are PAINFULLY slow. OK, we chased the fad. The old leader of our migrate-to-cloud initiative resigned and will sell her nonsense to someone else and declare her tenure a victory. Now, several hundred million dollars later, we're moving mission-critical servers off cloud to save money. No one is happy about the migration. The customers notice how much slower everything is. The cost is way higher. The bills keep going up every year. TMK, nothing got easier or better.
I'm old enough to see the parallels between offshore outsourcing and this. 20 years ago, articles on slashdot were running op-eds from "tech leaders" claiming you're an idiot if you're not having ALL of your technical projects done from India or another developing nation in Asia. Now, the industry has largely moved away from offshore outsourcing, replacing with something smarter...global offices...actually hiring talented people throughout the world, including India.
In both cases, there was a long-established paradigm. A new disruption came about...and instead of viewing the new option as an option, it was proclaimed to be the next evolution and you're a fool for not jumping in. In both cases, both paradigms were too good to be true. So...talented software engineers want to work for you for 1/5 of the salary they could get if they moved to your region?...also while companies around the globe are starving for tech talent and happy to pay extra to get some talent?...yeah, it never made sense in theory...and most who went down that road experienced what I could have told them...yeah, you can offshore your entire department...but you won't save money!
Today, I see the same with cloud hosting. So...Amazon and Microsoft are SOOOOO efficient they can run all this hardware for cheaper than you ever could...while these high-profit enterprises are keeping their unprofitable divisions afloat? AWS and Azure are so automated you can cut your IT dept in half? Sorry...none of that makes sense. Is their hosting model more efficient than your IT team's...probably....do they pay so much less for hardware, bandwidth, services, and staff that they can do everything you do for cheaper and make the legendary profits they report?...no...that just doesn't make sense. The math doesn't support that level of wizardry.
Without a doubt, cloud hosting is ideal for many companies. For others, it's just a more expensive way to do the same thing you did before. You still need a dedicated staff to manage the cloud. Your bills go up with time. The whole elastic thing never really materializes...which in fairness is more the customer's fault than the hosting provider's...but server deployments tend to be pretty static.
Tech leaders need to evaluate cloud hosting options rationally. Sometimes they lower costs and increase availability. Sometimes they decrease them. It's non a one-size-fits-all solution. I think many companies have took the plunge and realize not all of the hype materialized. For many, the cost savings never materialized. I am sure we'll see articles very soon from tech leaders patting themselves on the back for moving things back to on-prem to cut wasteful cloud spending.
buggy garbage (Score:2)
Microsoft's cloud is buggy garbage, it gets a little better over time but always buggy. Good luck breaking through the wall of consultants that do support, you'll need rock-solid steps to reproduce and a large number of customers affected. Have an intermittent issue, even with evidence? They wait until a day it works and tell you its fixed and close the case.
No wonder people walk away. I'll take stability over new features..
Re: (Score:2)
You know whats wild? I've been asking ChatGPT questions about Azure and 9/10 its answers have been vastly superior to the nonsense i get out of Azure support. Of course 1/10 its absolute hallucinated gibberish, but we kinda already know thats a potential outcome when we roll the dice. And its STILL vastly superior to the human alternative.
Kinda feeling thats the way a lot of things are gonna go. Which is a bit of a worry in my field actually. But I'm resigned to it, its not like as a coder half the work I'v
Re: (Score:2)
I'll give you an example of a Microsoft support case.
User: I tried to download a message trace and the server returned a 503 error. Here is the error message that includes all of the details, server name, timestamp, operation, etc.
Microsoft: Okay, please install Fiddler, man-in-the-middle all of your HTTPS connections, and grab a Fiddler trace showing your problem. Then, do Steps Recorder and grab all the screen shots of the steps you are taking to produce the 503.
User: Can you just look at the server provi
Well shit (Score:1)
proifit is not slowing (Score:2)
We tried cloud and it was dumb (Score:2)
I work for a small company. 10 employees total. Before I came in a few years ago, the company was paying $2,400/month to some hosting provider in Arizona for 3 small VMs with 10Mbps connection. I thought it was stupid and told them that anything would be better. Even Microsoft Azure. So we moved. And it was better and cheaper for the primary VMs. They do get you on storage costs. We were averaging about $1200/month with Azure. We stuck with them for a few years. And then I discovered OVH. I went
Re: (Score:1)
Re: (Score:2)
How do you get 3 128GB servers for $300 total per month?
Re: (Score:2)
According to the OVH pricing calculator, you don't. I can't get anything under $140/server/month.
Re: (Score:2)
Have fun with OVH support when something goes wrong ;-)
On the whole, I like OVH, and they are almost impossibly cheap - but man, their support is terrible.
Cost? (Score:2)
I notice for many high compute configurations we are paying within three or even two months for a VM what it would cost to build/purchase a machine. It's almost like returning to the older days of physically hosting machines in data centers like hurricane electric may be the answer.
Cloud is expensive (Score:1)
If Amazon (Score:2)
Cloud going up in smoke? (Score:2)
Title says it all.
Wow. Just as if cloud hosting cost a LOT. (Score:2)
LIke a LOT more than on premises hosting.
Capex vs Opex arguments only appeal to fools, MBAs and other types who have no idea how a company actually runs.
Business Opportunity (Score:1)
And, my friends, that means there will be oodles of money to be made in digital detransformation (i.e. migrating from cloud to onprem).
Cha-ching, baby.