These numbers are very misleading considering they wrap up things like Office into “cloud revenue.” When talking about compute cloud they are still way behind Amazon although you don’t get that from the misleading headline.
The article discusses "cloud revenue" of $26.7b but the more relevant Azure revenue is less than a third of that.
"Microsoft hasn’t disclosed how much revenue Azure brings in, but the Evercore analysts estimated that it generated more $7.74 billion in the 2018 fiscal year."
On the flip side, I believe Amazon counts their Office and Drop/Box competitors under AWS as well so it may not be as apples to oranges as first perceived.
I suspect there are a lot fewer customers for Amazon WorkDocs / WorkMail / WorkSpaces than for Microsoft's equivalents, though. I've certainly never heard of anyone using them.
In that scenario, the Microsoft numbers are even more mysterious as OneDrive for Business just another non-billable component of SharePoint that's consuming that storage, which AFAIK isn't Azure storage, but in dedicated Office 365 nodes. The Amazon Drive service AFAIK is just a frontend to S3.
The interesting thing about O365 is that it combines a very valuable licensing cash stream with a variety of services.
The story that Microsoft SEs used to sort of tell a few years back was that Microsoft needed so many spinning disks to support Exchange workload IOPS that they had ridiculous amounts of "free" storage available for low-IO workloads like SharePoint/OneDrive. No idea if that was true years ago or is true today.
Office is one of Microsoft's most lucrative lines, whilst I doubt that most AWS users pay much attention to WorkMail/WorkDocs. Many big companies play games with reporting numbers, and this is just a rather transparent example. Unfortunately, the fact that it got the Forbes story rewards that this kind of trick.
I think GP comment actually meant that AWS has their own products that compete with similar products like Drop/Box and Office and they include revenue from those products as part of their AWS revenue.
Yes, that's what I meant. Haven't figured out a clever way to say Dropbox + Box, Drop/Box seems OK. I'm also led to believe they compete with email as well.
This is nearly a "no true Scotsman" definition of cloud revenue, where if Amazon doesn't have a product in a market segment it shouldn't count. Google and Microsoft both have offerings in the cloud SaaS office suite market segment, and it's a real liability for Amazon that they do not.
Not really. When people buy Office 365, they're primarily purchasing a the MS Office product, and the cloud just happens to be the mechanism of delivery. That doesn't really make this cloud revenue.
Imagine if MS was forced to "unbundle" Office 365. Consumers could purchase a license to run MS Office on any platform. On of which might be hosted Azure. What percent of the Office 365 revenue would accrue to the products division compared to the cloud division. Sincere guess: 90%+.
Another hypothetical scenario. Imagine if Apple offered a deal on iPhones. Get an iPhone for free as long as you sign up for a grossly overpriced $1000 iCloud contract. All of a sudden Apple would have hundreds of billion in "cloud revenue". But no reasonable person would all of a sudden say that they're now the biggest cloud provider in the world.
I disagree, totally. Most companies who buy Office 365 want Active Directory just as much as the office products. Azure AD is an official Azure product; it's got Azure right in the name and lives on the Azure console. AD, hosted email inboxes, OneDrive storage; that's what companies are increasingly looking to O365 for. The Office suite is just a checkbox next to those things.
You can't punish Microsoft's categorization of these things just because Amazon doesn't have directly competing products. Well, actually they do; they have Workdocs and hosted AD. But no one uses them. Both of these are official AWS products; are you going to suggest we should start sharding out Amazon's product suite by revenue per product just to get a perfectly fair comparison of revenue?
G-Suite is categorized as a Google Cloud product. I guess we should just completely cut G-Suite out of GCloud's revenue total? Even though it does directory-like functionality and competes directly with Office, Outlook, etc?
Where do you draw the line? Ultimately, you draw the line at mechanism of delivery and payment, which is what Cloud has always been about. Office 365 and AzureAD are pay-per-seat, and thus by extension correlates with pay-as-you-go, just like any other cloud product.
>Not really. When people buy Office 365, they're primarily purchasing a the MS Office product, and the cloud just happens to be the mechanism of delivery.
And when people buy Amazon RDS, they're primarily purchasing a database product, and the cloud just happens to be the mechanism of delivery.
Imagine if Amazon was forced to "unbundle" Amazon RDS. Consumers could purchase a license to run the database on any platform. On of which might be hosted AWS. What percent of the Amazon RDS revenue would accrue to the products division compared to the cloud division...
--
Yes, this is the literal definition of a "no true scotsman" fallacy.
Amazon RDS is literally just a hosted offering of an off-the-shelf database server (MySQL, PostgreSQL, SQL Server). There isn't really anything there to "unbundle".
Amazon Aurora does appear to be a somewhat unique offering, but it appears to be tightly coupled to their infrastructure; I doubt that it could be unbundled without major rearchitecture.
I don't think this is that fallacy. It's also fallacious to dismiss something as wrong if an argument in its favour uses a fallacy.
If something is described as 'cloud' but uses a definition most people don't recognise or describe as 'cloud', that's at least worth noting. It's arguably prescriptive linguistics.
>Imagine if MS was forced to "unbundle" Office 365. Consumers could purchase a license to run MS Office on any platform. On of which might be hosted Azure.
You can literally buy office as a standalone product. That's the ONLY way you could buy it a decade ago. So the current situation is exactly what you're describing.
As a counterpoint: I purchased office 365 purely for Onedrive. I was totally happy with my ancient office products (2013?) before, but I wanted to get some cloud storage with good synchronization and sharing tools as well as a good web interface. The Onedrive/Office365 offering provided that.
However it might be true that I'm the exception on this, and that most of the Office365 licenses are sold to corporations that are more interested in the basic office tools.
Office 365 also has the cloud sync and collaboration features that are a pre-req these days for a cloud service. Whether you use them or not is another story, but they didn't just change how they price it.
I'm not really sure how my copy of Excel and Word running on my work laptop (and subscribed to monthly) should qualify as "cloud revenue".. We do host our email on their cloud servers, but according to their own sales team, that is worth $5/month per subscriber for email only. Yet we pay $20/month per user for the full office suite installed on their systems. (and more for people with other tools like Microsoft Visio and Project) Its not a "no true scotsman" its that many of their products have no cloud component at all.
Microsoft gloms together the pay-per-use/subscription part that totally doesn't need "the cloud" (Office apps) along with the other bits that do (email, Skype/Teams, OneDrive, etc.). It makes sense that they do from a financial reporting perspective, because they sell them and you buy them as one thing.
How exactly should they split out the email or OneDrive component from Word or Excel of your $10-20/month? I don't think there's really a sensible way to do it without overcomplicating things.
The copy of Word and Excel on your laptop have cloud integration. You can do collaborative document editing a la Google Docs through the desktop copies or the web app versions.
Amazon does have a cloud SaaS office suite, actually -- Amazon WorkDocs / WorkMail / WorkSpaces. I've hardly ever heard of businesses using them, though; Office 365 and Google Apps have the market pretty well locked down.
I feel like there is a fundamental distinction to be made between infrastructure services (like AWS S3, EC2, RDB, etc) and end-user services (like Office 365). Combining the two doesn't make a lot of sense; procurement for the two typically happens separately, and businesses frequently mix-and-match vendors for infrastructure and end-user services.
I was an engineer at Amazon for a couple years (left about a year ago) and I've never heard of anybody using these services. It's possible that some of the services that we used internally were simply frontends for those AWS services. Having an internal frontend for an AWS service was common.
Yes, WorkDocs is used extensively for document sharing. Chime is the go-to chat/meeting/VC app. People use WorkSpaces when they need a Windows or Linux test machine.
I would disagree, it's not a "no true Scotsman" argument.
Microsoft is actively competing against AWS IaaS and PaaS services, and leveraging embedded product features and license terms to force the issue with enterprise customers. Microsoft's spin is that those embedded relationships will compel customers to just join Azure. (Example: Windows 10 and Azure AD dramatically simplify PC management. How receptive is the market to that proposition? (I don't know, btw))
If they aren't getting heavy traction with their millions of existing enterprise customers adopting cloud services, they are doing something wrong that Amazon is doing right.
Additionally, the fact the Microsoft can sell Office and Exchange is a known, irrelevant fact -- they've been extracting a toll from every business since like 1995. What isn't clear is if they have the ability to compete toe to toe with AWS (the market leader in IaaS/PaaS).
I think there is another perspective here. Consumer vs Developers. When we speak of the ongoing cloud vendor wars between AWS, Google, Azure and the likes of Digital Ocean, we mean cloud platforms for others to build and host things on, not consumer apps delivered through the cloud. If that was the case, we would have to count something like Dropbox or Google Photos as well, right?
To me cloud provider in this context means a vendor that provides cloud platform for others to develop on top of. RDS counts, Office doesn't.
That's how I look at it and I'm sure that's how MS looks at it. Certainly they wouldn't be blinded by Office revenue and think Azure is close to beating AWS. Office competes with Google Work Suite which isn't even part of GCP (even though it's linked).
> To me cloud provider in this context means a vendor that provides cloud platform for others to develop on top of. RDS counts, Office doesn't.
I completely agree with you on this one.
However all the cloud providers include their SaaS subscription services revenue in cloud revenue. So when comparing market, revenue and stocks, that's what we should compare them on. But here we are in this thread looking at their "cloud revenue" to figure out which is a bigger "cloud vendor". That's like comparing Apples and Oranges.
When someone is directly comparing Amazon and Microsoft "clouds", I absolutely assumed they meant Azure vs. AWS. To the exact generic definition I shouldn't have done that, but when it's specifically up against Amazon, well, things can be true but also misleading.
Its not so much "no true Scotsman" as, "what is a cloud and what is useful to track?".
For comparing infrastructure offerings, it would be really misleading to include things like Amazon Music or MS Office (both of which are cloud offerings).
But they have transitioned office to a Saas model to do this. I used to buy office every 5 years, now I pay per month. It is also so much more than just the office apps, there is Onedrive and Sharepoint and email too. I would say that Microsoft have done a pretty good job of making it's products cloud based.
Does Google include G Suite in their cloud umbrella? Microsoft does a really good job of using products like O365 or other IT kinda services as a lead-in to Azure infrastructure. Which is ironic because O365 kinda sucks.
They do. On O365 I’ve said the same thing for a while, but the front end rewrites that are rolling out are really quite good. (Admin is still a tire fire)
In the world of large enterprise which is just now waking up and jumping all-in with "digital transformation" projects and "cloud strategies" Microsoft appears to be winning the battle. I work for a company that services large enterprises and anecdotally I'm seeing the majority of them choosing Azure over AWS. I believe this is in large part because Microsoft knows how to sell to enterprises, they have great tie-ins with existing on-prem core infrastructure like Active Directory which is a big deal to large companies. When large enterprises start jumping on board they will move the needle pretty significantly as they ramp-up. I anticipate Azure closing the gap on AWS pretty significantly over the next 2 years.
The numbers are NOT misleading precisely because they compare the cloud revenue. Google includes GSuite revenue in their cloud revenue. So does Amazon which includes Work Docs.
Using their "cloud" revenue to figure out which is a bigger "cloud vendor" is a fallacy.
But if we're talking about cloud computing and microsoft offers these these services in the cloud, they should not be penalized because their customers once started on the desktop. They're demonstrating that they are making the transition.
Yeah cloud and ML numbers are becoming a running joke in financial circles, at this point I'm pretty sure Oracle counts JRE downloads as cloud revenue somehow.
I wish these large corporations didn't demand so much creativity from their accountants to bluster an image.
Not to mention that Microsoft VARs have been doing things like offering discounts on Enterprise Agreement licensing that correspond 1:1 against pre-committed spend on Azure.
Ben Thompson of Stratechery precisely wrote about this last week:
>Given this week’s theme, though, I wanted to focus on AWS: revenue was up 46%, and while that may be lower than Azure in percentage terms, it is almost certainly higher in absolute terms. AWS had $6.7 billion in revenue last quarter, while Microsoft’s entire Server Products and Cloud Services — the majority of which remains on-premises — was $5.7 billion. Microsoft of course has other cloud revenue, including Office 365 and Dynamics 365; those are SaaS products though and generally don’t compete with AWS.
wonder how much of those were attributed to Serverless revenues....probably negligible.
2018 is almost over and I still don't know if I should go fully serverless, more importantly, which cloud vendor has fixed the "cold startup time" for serverless functions. Google seems quite promising but still no recent study has been done on this. Would somebody like to see a test on this done? I've been wanting a straight up answer for a long time but hard to find a consensus, seems to be all over the place.
It will depend a lot on "sporadic". The dominant factor is where your workload sits in the warmth hierarchy. RAM and disk space are limited, so less-frequent workloads will eventually be evicted from worker nodes.
Very anectodal, but some things are better on Azure, because they were late to the party, for example azure blob storage.
It costs the same as s3, but it supports journaling, so you can do journal archives.
You can run your main thing on aws and keep backups on azure.
If you wanted backups with journaling on AWS, you'd have to use EFS at minimum, which is orders of magnitude more expensive.
Disclaimer: also very anecdotal: However, most things are way worse than either GCP or AWS.
Azure is just now rolling out availability zones. There are strange constraints everywhere, like you can only use premium storage accounts with specific VM types. You cannot use the same premium storage account for blobs. In general, not all configurations of skus for storage accounts, vms and disks are supported.
You need to match availability set skus between scalesets and load balancers.
If you connect an internal 'standard' load balancer to an instance, then it cannot reach the internet anymore. You can only attach one internal and one external load balancer to a scale set. You can't use the internal load balancer to loopback (same as internal AWS NLBs, but AWS has other options).
And so on. It all seems poorly integrated. Budget 5x as much time for a similar deployment on Azure as compared to AWS or GCP. Add more buffer if you need flexible automation, as you'll bump into those edge cases often as people change parameters.
They have a ridiculously huge amount of features compared to the other two major cloud providers, but integration between them is iffy. It is improving.
AKS seems pretty good, although it's still new and people are reporting issues. But it is promising, you can actually tweak quite a few things, and I have high hopes for the compatibility story (even more so than GKE). AWS's EKS is a marginal implementation at best.
On Azure AD, the tables are turned: it's AWS that provides an inconsistent story. There are many options, none are as well polished compared to Azure.
They did learn from some AWS mistakes. But technically speaking they have a bunch of polishing to do.
EKS is, without a doubt in my mind, among the worst products Amazon has ever released. You pay $0.20/hour for the "benefit" of a managed high-availability control plane, no matter the size of the cluster, then are forced into a seemingly incomprehensible mess of CloudFormation templates to spawn a set of VMs which you now have to manage yourself.
It makes no god damn sense. They had every opportunity to look at how Azure or Google were doing it and just straight-up copy it. For Christ's sake, they were over a year late to the game.
I remember during the original announcement, maybe its documented somewhere, they announced that Fargate was coming to EKS "eventually". They don't even have managed worker nodes for normal VMs, if their intention is to bring Fargate as well we literally won't see it for another two years. It took them over six months to get from announcement to GA.
The only good thing I have to say about it is that it's pretty reliable. Which isn't saying a lot because it doesn't do much, but they're happy to charge you $150/mo for what it does.
Did I mention it takes over 20 minutes to create a new cluster? Well, actually, they "call" it a cluster, but its really just the control plane. The actual cluster can take up to 30 minutes for everything to actually connect and be usable. My company was interested in doing a "pristine" staging environment for every/some PR, throwing it away afterward. Interesting idea. Not worth it on EKS because their performance is so god awful.
EKS is literally the reason why we abandoned AWS in favor of Google Cloud. I'm not suggesting that the rest of AWS isn't overall pretty great, but when nearly all of your workload is defined as "90% inside kubernetes, a few cloud-managed queues, and blob storage" Amazon isn't doing themselves any favors by ignoring the most interesting infrastructure technology to be released in the past five years. Oh, and our bill dropped by ~50% because preemptible instances on GCloud are actually easy to use, instead of the sheer insanity of trying to integrate spot fleets with an eksctl cluster.
And some things are very much worse. I feel like if you stick to MS technology feature you are probably going to be ok but my anecdotal evidence around other features built around open source tech has been not good.
For example AKS (azure kubernetes service) has been problematic for us. Similarly AWS RDS is much better than Azure PostgreSQL offering.
We've had nodes die (more than what we've experienced with simple vms). We've had cluster go into failed state multiple times and require full rebuild. We've had nodes simply stop responding but aks claims they are fine. It's been relatively unstable and we've had to work with support multiple in the past year because of issues. It has improved but it was pretty awful when they announced GA. We were surprised when they did. We had been on AKS for a bit and it was no where close to ready for GA.
Azure Blob Storage also supports immutable blobs (they can't be overwritten or deleted once written to), which is interesting for various compliance and auditing scenarios:
Interesting. S3 does have object versioning, which I believe once enabled, causes the object versions to preserve always be preserved. I think the AWS equivalent though is using AWS Glacier and creating a vault lock.
These numbers are so misleading. I understand why they are the numbers published, but AWS vs Azure last time I heard real numbers was approximately a 3:1 market share ratio.
Including Office 365, Dynamics, etc doesn't make sense to me. We don't include Prime in Amazon Cloud, AFAIK.
I'm also an MS employee. Are you sure no Office service revenue is included? I assumed from the size of the gap they must be piling it in. If not, that leaves an even larger unaccounted-for gap between Azure and the claimed 24 billion.
Under C+E, the last time they tipped the hand on cloud revenue it was ~18 billion with about a third from Azure AFAIK. To me, that is woefully misleading to compare us on 1:1 terms with AWS.
AWS includes their PaaS AWS service offerings in their revenue, no (Cognito, dev tools, Workspaces, etc)? Why wouldn't Microsoft include O365 and Dynamics then?
Raw compute and storage isn't where the money is at long term, it's the lock in value add services (especially with Kubernetes and Hashicorp making multicloud and portability first class).
Sidenote: I have seen A TON of open roles looking for Azure and GCP skills over AWS lately.
Open roles for Azure & GCP might remain open for longer because of lower supply and poorer matching, not because of higher positions being searched for. The AWS talent market probably clears more easily given that it has existed longer.
I don't disagree. I'd also argue it's because certain retail clients won't use AWS for obvious reasons like not funding your competitor (based on conversations I've had with retail clients).
AWS has WorkMail and WorkDocs, which is in the same space as Office 365. Amazon Connect sounds similar to Microsoft Dynamics, though I'm not really familiar with either.
Amazon isn't currently a major player in that arena, but they are still competing under the Amazon banner.
They're completely separate products. Connect is call center facilitation software that can utilize lambda (and correspondingly a suite of other services) when routing and leverage AWS resources for metric collection and logging. It integrates with both your typical desk phone or a computer via WebRTC.
Microsoft Dynamics is a CRM product. Completely different spheres.
How does it work with the Enterprise Agreements these days? A few years ago it was common to hear that customers got O365/Azure included in their EA (likely free), but weren't using it. Don't know how much truth there was to it, since I don't have first hand experience. Are EAs with bundled cloud still a thing? (Were they ever?)
If you are a core cal customer there is a bridge cal to link your subscriptions to the EA. They do not want you to carry office as perpetual software and will encourage you to negotiate away your ability to terminate O365 and get your licenses back, if you had that right.
The push now that I hear about is subscription windows, pstn dialing/O365 E5 and azure commit.
Assuming Amazon continues on the same 25% YoY trend, it would be 15B in 2018. Microsoft could beat that number by reaching a 400% growth rate. That would be pretty amazing. But it's not the case.
This is comparing totally different numbers than what is being compared in the OP.
I'm not exactly sure what Gartner includes in its "IaaS" revenue figures, nor am I sure what Forbes is including in its "cloud" revenue figures, but even Amazon's official press release [1] (which is what the OP is comparing) gives completely different numbers for 2017 AWS revenues than Gartner.
Even if you exclude Office 365, etc there your statement that AWS has 80-90% penetration is absolutely false. AWS has about ~45% of the market, Azure, GCS, IBM and Rackspace round out the top 5. Amazon & Google also both include their Office Suites in their Cloud numbers...so, it's not "a joke." Everyone does it. No one denies that for compute AWS is #1, right now.
You are blatantly wrong. Man, I've been in this industry for a long time, and I met literally thousands of cloud computing customers, from startups to large corporations.
The current state of affairs is that AWS still dominates. Your claim of 45% shows that you are simply repeating what some myopic analyst has written. Talk to industry experts, and you'll see why I'm right.
Your anecdotes mean nothing, there are plenty of people gathering data and market share numbers out there. Also, you're on HackerNews and you honestly think you're the only person on here that has dealt with a lot of cloud customers and vendors? C'mon.
> The current state of affairs is that AWS still dominates.
Nobody is saying otherwise.
> Your claim of 45% shows that you are simply repeating what some myopic analyst has written. Talk to industry experts, and you'll see why I'm right.
This is anecdotal. Provide some verifiable source.
Does AWS not report their WorkDocs (which nobody uses, and hence less revenue) and similar revenue in their AWS revenue? Does GCP not include G-Suite revenue in their Cloud revenue? I don't see what 'game' is Microsoft playing? Are they saying they are the #1 cloud provider because of their revenue? No.
This report is not a dick measuring contest to find the best and the biggest cloud provider. This is for Microsoft to showcase to investors why they should invest in Microsoft.
It fluctuates a lot, the last number I saw from Gartner pegged them around 45%, but it could be less or more at this point. The whole space is growing.
I work for a company that sells software which is often deployed in the cloud. I'd say the ratio is about 30 to 1, with AWS far ahead. I'm always surprised when somebody starts talking about Azure - they usually work for a company where some executive was talked into Azure by Microsoft.
There are arguments on both sides about "Should Office count as Cloud revenue?". We have the same decision regarding GSuite. If you take the "Cloud is anything a business depends on" argument, then it obviously should count (but then all sorts of things like SaaS would count).
I like to think of this from an investor's standpoint: what do they really want to know? In my estimation, they want to know about potential hyper-growth market opportunities for the company.
For Microsoft, more than anyone else, Office isn't a high growth opportunity. There is some amount of revenue uplift from customers "upgrading" from Office on prem to Office 365 (like Adobe when they forced Creative Suite to be subscription) but I doubt there are going to be lots of net new customers just because Office is on a subscription model. By contrast, Cloud infrastructure is a large opportunity for Microsoft.
For Google (and to some degree AWS), "productivity suites" is actually a reasonable growth market, particularly as market share often comes from Office. However, the raw productivity market is still "only" $XXB (it's predominantly $X/seat, and there are only so many employees in the world) while people talking about "cloud" usually are talking about the opportunity being "all money spent on IT".
tl;dr: I watch the YoY "Azure growth was XX%" numbers and ignore the headline revenue.
To be fair, the VP of Strategic Communications at SAP in 2011, and Chief Communications Officer at Oracle from 2012 to 2016, surely has no dog in this hunt.
I came to the comments fully expecting first comment to show false advertising.
This is what I come to expect from Microsoft, I just don't trust anything coming from their PR department. This can't be good for Microsoft.
In seriousness, isn't this where Microsoft should have a competitive advantage? If Excel integrates with Azure easily, doesn't it incentive Azure adoption? I guess I'm picturing a bank analyst running a derivatives model that in the background, spins up 10 Azure server-less functions in parallel that hit a database hosted on Azure? The banker is now happy Excel "works quickly" but it really is just connects a front-end people are comfortable with alongside cloud resources?
