Rideshare giant moves 200 Macs out of the cloud, saves $2.4 million

Singaporean super-app company Grab has dumped 200 cloudy Mac Minis and replaced them with physical machines, a move it expects will save $2.4 million over three years. Grab is Southeast Asia’s leading rideshare and food delivery outfit and therefore needs to build apps for iOS to connect with customers. In a Thursday post, the …

[Tim99]

Multi-user?

MacOS is a (heavily?) modified UNIX based system. I run my iMac with several different accounts, does anyone know if you can use multi GUI users on the same remote machine? Like I do with VNC on a Raspberry Pi5 "server".

[that one in the corner]

Re: Multi-user?

> Like I do with VNC on a Raspberry Pi5 "server".

Just for curiosity's sake, any particular reason for the quotes around "server"?

[Tim99]

Re: Multi-user?

Because: sometimes it is, sometimes it isn't; and when it is, it may be several distinct types (web, application, and database), and I have several other Pi5s which also alternate roles. Possibly, also, the servers I deployed when I was working cost somewhere between $thousands and $hundreds-of-thousands? I forget that I am old, and that the Pi5 is roughly 1,000th the price and 10,000 times more powerful than the DEC VAX that I wrote stuff for when it was a general SQL, file, and application server back in the 80s - Somehow it doesn't seem right...

[Kurgan]

Re: Multi-user?

Because the Rpi is such a poor excuse for a computer, stability-wise, that calling it a server is like calling a Reliant Robin a car.

[that one in the corner]

Re: Multi-user?

What on Earth are you * doing* to your R'Pi?

(Just checked, my 24/7 R'Pi has uptime of 287 days - which is about right for the most recent power cut).

[Anonymous Coward]

Re: Multi-user?

How do you patch it?

[Dwarf]

Re: Multi-user?

@Kurgan

Because the Rpi is such a poor excuse for a computer, stability-wise, that calling it a server is like calling a Reliant Robin a car.

You seem to be on a different planet to the rest of us. A Rpi(*) is a very capable computer and stable, as long as you build and manage it correctly. If you find it not to be, then you are doing something wrong.

* Of course "Rpi" is a collection of different systems, not just one model, so if you were referring to a first generation Pi A, I might agree with you, but the recent stuff (read last couple of years) is miles ahead.

Similarly, the definition of "computer" varies widely. If you want to compare it to a current generation super computer, then it will fail significantly, so probably worth clarifying what you think the problem is and some people here might be able to help you get over those problems.

The same logic about "car" can be applied to a Robin Reliant and to a Bugatti, or even a Tesla if you squint your eyes enough. So, if you want a super reliable computer, you might have to spend more than £50 on it.

[VicMortimer]

Re: Multi-user?

Yes, you absolutely can.

[GoneFission]

Re: Multi-user?

Since 10.7 actually:

https://9to5mac.com/2011/02/27/10-7-lion-allows-multi-user-remote-computing/

[Henry Wertz 1]

Re: Multi-user?

I' run my iMac with several different accounts, does anyone know if you can use multi GUI users on the same remote machine?'

Yes, but their absurd licensing makes it so EACH user must pay for 24 consecutive hours at a time. (You may still see some clioud provuders offering macOS instances of much older macOiS instances, this is partly because these terms to put the screws on virtualized users kicked in aboout 5 years ago, probably in resoponse to this very type of use where peoole may have only had to pay for a few minutes of use at a time otherwise.)

[gnasher729]

Re: Multi-user?

“ I' run my iMac with several different accounts, does anyone know if you can use multi GUI users on the same remote machine?'”

You can have multiple GUIs running, but only one GUI is being displayed. (Only actually tried two, using one while the other burnt a DVD, and it worked just fine).

[talk_is_cheap]

Re: Multi-user?

It is best described as UNIX like rather than UNIX based as its original base was FreeBSD. The development cycle of FreeBSD means that by 1994 it was free of any UNIX code from the original AT&T code base.

[rcxb]

Re: Multi-user?

UNIX like rather than UNIX based

Yeah... nobody cares. Sorry.

its original base was FreeBSD

Apple's XNU kernel is based on Mach.

[Stephen Hurd]

Re: Multi-user?

It's actually certified real UNIX, no qualifiers.

https://www.opengroup.org/openbrand/register/

[Nate Amsden]

Yet another case..

of a company filled with cloud folks who don't care about costs... until they are forced to care... just need to keep on repeating this again and again across more organizations.

[doublelayer]

Re: Yet another case..

Maybe being a little less general, though, unless you want to be wrong sometime which doesn't help. Cloud can be more expensive a lot of the time, but if you want a way to make it a lot more comparatively expensive than it otherwise would be, try to run Macs in it. Macs clearly weren't designed for this because you can't run Mac OS in a VM on commodity hardware, or custom hardware, and they don't have DC-scale hardware you can buy from Apple. Combine that with the requirement to rent them by the day, which the article and Grab's statement suggest is Apple's doing although I thought that was AWS's choice, and you've got a recipe for something that is even less connected to hardware prices than normal cloud hosting would be.

If you use this example to predict what other price comparisons would look like, you're going to be making overestimates which won't convince many people who check your calculations. Comparing prices requires creating detailed plans of what infrastructure you're going to use and what you need to set it up. People with an agenda, usually a sales commission, who decide that it can all be done from the cloud, it's definitely less expensive, and let's not bother with those boring plans, are often wrong and always annoying. Unfortunately, the people who have an agenda and say that running in the cloud is always less functional and more expensive, and you don't need to consider those boring plans to know that, are equally so.

[Nate Amsden]

Re: Yet another case..

I am unsure what you are trying to say here.

I happily admit there are some great use cases for public cloud, they involve extreme elasticity with low amount of time spent in the high mode. For example you need 2 thousand CPU cores for not more than ~60 hours per month. Once a month 60 hours, or say average 2 hours per day over a given month. I haven't run the exact math but hopefully you get the idea, buying/running infrastructure 24x7 for something that only runs for a short time doesn't make a lot of sense (again haven't run the math so maybe it still does).

Another use case I have seen here on el reg(can't find the article) is similar just larger scale, one off super scale tests, the article I'm thinking about was a HPC test of some kind where they spun up a few thousand systems to run a benchmark to show scalability of some software stack. Maybe it cost them $250k or something to run the test, a lot cheaper than leasing/buying a bunch of hardware to do the same thing. Assuming of course you don't run such things often(maybe once every few years or something).

Of course the number of these workloads are few and far between.

Pretty much anything that has some kind of steady state workload will cost a lot more to run in a public cloud IaaS. Even those with "bursty" workloads you should do the work to determine exactly how "bursty" your stuff is, in the last 25 years of my career I have not worked with a company where their "bursts" would come even remotely close to needing public cloud(talking on the scale of say ~100 CPU cores bursting to several thousand). Also keep in mind it is a BAD IDEA in almost all situations to burst from a data center, to a another data center (including public cloud). Latency between the two locations will often seriously hurt performance (you could get around this to some degree if you were to locate your primary facility very close to your cloud provider). Too often, people think "oh my traffic surges 500% so I want to burst to cloud", not taking into account how much resources that 500% surge actually requires. In all cases in MY experience that surge is nothing to get excited about capacity wise. If you are talking a 100,000% surge then yeah..but your bigger concern is probably application scalability at that point anyway.

I came to this new realization recently though, that many organizations are filled with "cloud people" (I am starting to use the term "Cult of the Cloud") who have bought the marketing BS hook line and sinker over the past decade, and as mentioned in previous comment don't really care about the costs, until they are forced to. Look at examples like SAP (https://www.theregister.com/2025/09/04/sap_sovereign_cloud/), Geico (https://www.thestack.technology/warren-buffetts-geico-repatriates-work-from-the-cloud-continues-ambitious-infrastructure-overhaul/), 37 signals (https://www.theregister.com/2023/09/18/37_signals_cloud_repatriation_savings/), DropBox (https://www.trgdatacenters.com/resource/dropbox-left-the-cloud-in-2015-and-never-went-back/), and my personal example in the small business space moving my previous employer out of AWS in 2012 with 7 month ROI saving $10M+ over the following decade and provided BETTER availability than Amazon did provide in us-east-1 region. Ask yourself why it took Geico many years to realize public cloud was so expensive, same for SAP and others .. there is only one answer.

I can't count how many conversations/chats/threads I have had over the past 15 years talking with people in the "Cult of the Cloud", I swear it's like asking a "MAGA" person how they think Trump is doing as president. The recent AWS outage that number has spiked quite a bit as I try to clarify things for people on LinkedIn. I was in one thread with a person who is deep in the cult. At one point she said she had "1 million endpoints, how can I deal with that many without auto scaling?". My answer was simple: without knowing more I can't really say for sure, but I can say I worked for an advertising company where we served several billion requests a day(very light weight mind you) without a lot of infrastructure(I think it was about 30-40 servers in 3-4 data centers running active-active). She was convinced you can't build a CDN without leveraging public cloud, and I provided evidence that most of the CDNs do not use public cloud you just have to do a WHOIS on their CDN endpoint see who owns the IP(completely unrelated thread a few days later a Cloudflare employee specifically stated to me that they do not leverage public cloud for anything, I sort of assumed they may "burst" into public cloud for their workers but he said that is not the case). She said public cloud really pays off when you are at high scale, and said I don't work on high scale stuff so I wouldn't know. I happily admit I don't want to work for a big company. I gave her examples of Geico and SAP spending hundreds of millions of dollars on cloud and they are moving out. She stopped replying at that point(was about 7-8 days ago). You can present these people with overwhelming evidence and for whatever reason in many cases they just can't believe it, it really is like a cult.

If you(or anyone) is happy with your public cloud provider in what they provide to you at whatever price you pay for it, that's great, keep using them. Just don't pretend you are saving any money (I'll admit again there are use cases where you would save, but they are few and far between). There are certainly technologies out there that I am willing to pay a premium for myself.

[jvf]

Re: Yet another case..

"cult of the cloud"-Awesome! a perfect description that I will use from time to time. Thank you for that.

[Nate Amsden]

Re: Yet another case..

sure thing, ironically enough yesterday I purchased the domains cultofthecloud.com and cultofthe.cloud with the idea that I want to write some of these thoughts out and put it on that site(I host my own sites on my own server in a co-location facility) in a more organized fashion rather than try to come up with the stuff on the fly every time I go to write a comment. Not sure when I will get around to actually doing the rest of it, but maybe soon...

[doublelayer]

Re: Yet another case..

I am saying that generic "cloud is always more expensive" arguments are wrong enough of the time that they aren't useful to make, and I don't just refer to the extreme spikiness exceptions you've mentioned. There are various situations where renting machines can be cheaper, and you have to evaluate those costs manually*.

I was also saying that Apple machines make for a very unusual price comparison. I've been saying that for some time. When AWS launched their Macs as a service, I commented that thirty days of rental equals the cost to buy the Mac concerned. That's nothing like normal cloud, where thirty days rental of a server is not enough to buy anything as powerful.

* One example which also often comes down in favor of cloud is a place with very low requirements. There's a base cost to have even one server in a colo. I've done that calculation several times, and until you reach a certain amount of hardware, you can have more capacity in the cloud, including redundancy, than running the bare minimum on your own hardware costs. Depending on the environment, that can sometimes be handled by running the one or few servers in the office of the organization involved, but places that need a few small VMs are also commonly the places that don't have a suitable facility or, in some cases, any office at all. The ones in the middle are where you often need to calculate for much longer to estimate what the equipment and facilities would cost and what the rental in the cloud would cost to compare.

[Nate Amsden]

Re: Yet another case..

I can certainly agree there are likely other use cases out there where public cloud makes sense, in my 25 year career those are very few and far between, and many use cases I see of companies moving in our out of cloud line up with what I have seen.

The one example of very low requirements is a decent one I suppose, there is a whole "VPS" hosting market out there catering to those kinds of folks. Myself I used to pay for web hosting space on small ISPs back in the 90s. Just pulling a number out of my ass here of course, but if your IaaS spend is more than say $5-10k/mo then you should look closer at how things are being used.

It's jaw dropping astonishing how many organizations spend $100k/mo+ without blinking on public cloud stuff, this is going back to my first public cloud hosting 15 years ago where the company's bills were upwards of $500k/mo for a SMALL startup. Completely clueless doesn't even begin to explain things. The costs were in part driven by very high turnover, no documentation, we literally had 100 maybe 200 cloud servers that nobody even knew how to access(I worked on a project to try to identify unused things and shut them down to cut costs there), and nobody knew if they were being used by anything, so everyone was scared to turn stuff off.

The company I helped move out of cloud in 2012 my manager told me years later he had to put his job on the line in order to get the board of the company to approve the plan. We were spending ~$90k/mo in AWS for a few months. He hired me explicitly to move us out(he hired me at the previous company that was spending $500k/mo at times). So he knew what I could do. The board was skeptical, fortunately we had a very supportive CTO and my boss told him he guarantees this will work or he'll quit(or something). They agreed and it was a massive success, even on the first day. The performance improvements on day 1 were dramatic(honestly more than I expected), I still have the email the CTO sent to everyone on that day. Over the following decade as management changed I went through several more rounds of "I want to use public cloud" and then I prove to them what the costs are and they are just shell shocked. So. many. times. Even at my current org, I just went through that again in the past few weeks(technically I just gave the data to a manager who went and got the pricing which was 4-8x more depending on the cloud provider, I try to not waste my time with direct involvement).

Years later, our new CIO at that company left for another company and as some layoffs mounted others joined him. They were in a data center running old stuff, I don't know much details other than they decided to go to Google Cloud. Costs were huge, and the funniest part was they spent 4 years trying to move stuff and they never got everything moved, so they were paying both bills at the same time. Leadership refused to commit to "reserved instance" pricing to cut costs out of paranoia(???), simultaneously refused to reverse course on the cloud migration - eventually they ran low on cash and started laying people off themselves eventually being acquired and I think most of the staff were axed?

Another similar example early on, take a look at the image showing this company's budget/spending in this news article from 11 years ago https://www.geekwire.com/2014/moz-posts-2013-5-7m-loss/

You can ignore the text in the article(or don't..) because it does not mention their move out of AWS. I am friends with a guy who knew the CEO there at the time and told me the CEO was so incredibly angry after realizing how much $$ they were wasting. Beyond moving out I have no idea what happened in the years that followed.

But yeah, if you are spending $5k or so a month or less perhaps my posts aren't as relevant to you(?)

But seeing so many clueless idiots doing this stuff over the past 15 years sorry has just driven me crazy. Because even though I host my own stuff and have for a long time, the marketing crap that people eat up still comes back to me and drives needless stress (at times). So I feel compelled to try to fight back/inform where I can, even if it's a futile effort, at least I tried.

[Anonymous Coward]

Re: Yet another case..

"cloud folks who don't care about costs" ... This is not my experience of Cloud; it's the shining example of what happens when Bean Counters get there first.

I think what we're seeing is a consequence of this and recognition that Cloud is not a universal panacea and 'Cloud First' strategies are fundamentally flawed.

We can expect to see lots more of this sort of thing and a mass exodus to physical data centres and "sheds".

[cdegroot]

Hosting on macOS is one of the sillier things i read today.

[frankyunderwood123]

this comment is one of the most misinformed and pointless comments i’ve read today.

Go and read the daily mail or something equally fitting for your level of thought.

[Anonymous Coward]

But is it though? If they are saving 2.4 million a year from roughly 200 Mac "servers", then I'd argue it was never a great platform to host the application in the first place.

A smart option would have been to develop a cross platform app that didn't rely on Apple, lol.

[Anonymous Coward]

According to the article, they're STILL running Mac servers, just their own in their own co-lo... and are saving a packet. So the issue isn't the "Mac Servers" themselves it seems. Your argument is flawed, AC1.

[MonocleRB]

...unless you have no other option

If you read the article, you'll see where they lay out their use case: building their iOS app. You just can't build iOS apps on Linux or Windows, Xcode is the only way to do it and Xcode is only on macOS. That's a major reason why cloud providers offer Macs at all.

And before someone suggests that they run Hackintosh on off-the-shelf PC hardware: consider that the time spent getting such a solution to work (especially reliably over the long term) would completely wipe out any savings from cheaper hardware.

[Dan 55]

Re: ...unless you have no other option

But do they really need 200 Mac servers to build their apps? If they're running CB/CI they'd still need a handful at most.

So what are they doing with the 195 remaining Macs? It seems to me they've built a cross for their own back somehow.

[Richard 12]

Re: ...unless you have no other option

Yes, that figure jumped at me too.

Clang is very fast, building very large and complex macOS applications usually takes under ten minutes.

iOS apps are very rarely large, build times over a minute are rare. Swift also builds really quickly.

Notarization takes a long time, but you don't do that locally for iOS, it's done by Apple after uploading.

Unit test suites might tie up the build machines for longer, but I'm still wondering why they're apparently doing 200 concurrent builds.

[steviesteveo]

Re: ...unless you have no other option

That's the perks of being a core business function. I wonder if it's linked to their capacity on android - a $23bn company really could get by on a handful of CI/CD machines but it's not worth the brain cells of minimising linux vms specifically for building the thing that makes all the money. Then, if the Android team have all this (elastic) capacity, the iOS team are going to want at least parity

There's developer value in the luxury of never having to wait for a build machine and 200 machines must realistically get you there

[Alan Birtles]

Re: ...unless you have no other option

Yep, aws mac pricing is crazy, we calculated a month of usage would be the same cost as buying the hardware, with the minimum 24 hour pricing you can't even use it for a temporary burst of higher capacity which is where cloud usage would make sense

[talk_is_cheap]

Re: ...unless you have no other option

It may just mean that they do integration testing well. I've seen large systems complete all their tests within a few hours and only once a day for the whole development team based on just the merged branches, while other systems have run the whole CI suite for each developer every night for each branch they are working on. Guess which approach costs a lot more in CPU cycles and development time, but far less overall for the business long term.

[katrinab]

Re: ...unless you have no other option

And MacOS26 is likely going to be the last version of MacOS that will run on a Hackintosh. Unless someone can figure out how to hackintosh a Snapdragon, which I think is unlikely.

[gnasher729]

Re: ...unless you have no other option

Now ask about the last version of macOS running _legally_ on a hackintosh. Running on VMd that run on a Mac is legally fine.

[katrinab]

Re: ...unless you have no other option

I'm not talking about whether it is legal or not. There is not going to be an Intel build of MacOS 27. The Arm build is going to expect an Apple Neural Engine and other Apple specific stuff. Maybe you could emulate those in software, but it is not going to work very well.

[Nate Amsden]

Re: ...unless you have no other option

I realize of course Apple probably doesn't care, but I do believe they have a large data center footprint for some of their things. I also believe they do spend a lot on public cloud providers (their mindset is "we have so much money we don't care[1]", and that's fine...).

But my point is they probably have a good opportunity to make a data center with their hardware(probably change the form factor so it is more rack friendly/etc like they had with their rackmount servers years ago) for the sole purpose of this ability to remotely develop apps for their platforms(and host it themselves, selling access to the stuff to others). It may even benefit their own employees (for all I know they probably already do this to some degree for their employees). With their vertical integration they can probably do a ton of stuff from a (Apple) hardware perspective that other cloud players can't/won't do.

[1] I think this is their mindset due to something I read years ago where the claim was Apple would say something like this to people who interview there "I don't care if you can save us money, I care if you can make us more" (something like that), and obviously they have tons of cash.

[Pascal Monett]

"it was a no-brainer"

Hardware costs next to nothing these days. What is a no-brainer is buying a properly configured desktop case. Hell, even RAM doesn't cost all that much these days.

Who are these morons who think that The Cloud^TM is going to save them money ?

Don't they read El Reg ?

Ah. Probably not. You have to be smart to do that.

[Henry Wertz 1]

Re: "it was a no-brainer"

'What is a no-brainer is buying a properly configured desktop case. Hell, even RAM doesn't cost all that much these days. l

a) These are Macs. RAM is quite costly. Nonexpandable so you can't just buy afternarket RAM.

b) Have you oriced RAM recently? Spot pice has hit about as high as it was back in 2010. Pricing is downright dystopian.

(This doesn't negate your point though.)

[Bodhix_]

Re: "it was a no-brainer"

The cost of self hosting is not only in the hardware, but in requiring additional technical staff to configure and maintain the infrastructure with inferior uptime numbers than a hyperscaler. Not having to pay for engineers to rack and stack hardware, swap out failing components, manage on prem network infrastructure, etc more than makes up for the cost difference in almost all cases except for Mac. On top of that, if your business does not revolve around managing infrastructure, hiring infrastructure staff is like having a staff plumber instead of hiring one when the toilets back up. Besides, the problem with mac on cloud is not a technical one, it's a political one since the reason you have to rent macs by the day instead of the second is because apple says so. It's is an evil company that should be shuttered.

[Bebu sa Ware]

I had to laugh

200 mac minis, 42RU.

Sometimes the direct brute force approach has its own peculiar elegance. Certainly doesn't want for simplicity.

These are by definition development machines, not production machines so a bit more leeway is permissible.

I guess for redundancy they could collocate an identical arrangement in, say, a Sydney datacentre.

I would have thought a cross build environment would be possible - at least under Darwin but only ever compiled Unix text/terminal code on a Mac so I am only guessing.

[Henry Wertz 1]

Re: I had to laugh

'I would have thought a cross build environment would be possible - at least under Darwin but only ever compiled Unix text/terminal code on a Mac so I am only guessing.'

Yoiu'd think so wouldn't you? Franly, I would guess distcc or something could do exactly this, but, ither than OCLP (Opencore Legacy Patcher) and associtated projects, I just haven't seen macos users pushing the envelope, i think the kind who would justr use gentoo or arch etc. instead and simply steer clear of Macs. I would love to develop without a macos vm (I'm not about to buy the physical hardware) but found it to not even be something peiple pursue. I mean visual studio code has a thing for it but it literally just connects to xcode over some remote procedure call or socket or something, not move any work off the Mac.

[that one in the corner]

Re: I had to laugh

According to TFA, the Macs *are* running a cross-build environment, that is the reason they have to *be* Macs: they are used to build programs for iOS.

[that one in the corner]

Re: I had to laugh

> These are by definition development machines, not production machines

WTF is a "production machine"?

If you meant "server hardware" then say so.

'Cos all the time I spent writing applications, the "dev machines" were the large ones* on our desks, with extra RAM and drive space, whilst everyone who was productively running our apps in their production environments was doing do on laptops with far less grunt but a lot more portability.

* So we could do things like run VMs to mimic the smaller target machines *and* simultaneously a realistic model of all the networked resources they would be using the app to interact with.

[telveer]

The true savings may be somewhere in between, or not at all. Running their own infrastructure requires data center space, network, storage, backups, etc ... and people to manage all of this.

[Throatwarbler Mangrove]

You've got to love the armchair CFOs who assume that the company is too stupid to compute the total costs of this change. I would assume, by contrast, that the company has extensively crunched the numbers on the migration, especially before spending hundreds of thousands of dollars on hardware.

[doublelayer]

I totally expect they've done the calculations and are correct, especially with Macs. However, remember that these are exactly the same finance people who are getting blamed repeatedly throughout this comment section for failing to calculate the costs of the last version. Sometimes, financial decisions are not made with the complete sets of calculations you expect which often leads to having to do similar calculations later on to clean up the mess caused the first time.

In my experience, this especially applies to cloud versus non-cloud, where a lot of the decision gets made by IT and engineering and finance doesn't directly contribute to it at the start. It doesn't matter which decision they make; I've seen bad choices in both directions. Usually, finance does come in later and ask or demand for prices to drop, but they don't know or care what the choice was or try to figure out which approach is better. Maybe if finance had more tech knowledge and the other teams had more attention to financial success, but in my experience, they don't, there's often a lot of compartmentalization between the teams, and everyone wants the easiest solution for whatever they're doing, which leads to little or not-so-little problems.

[Joe Gurman]

And people….

Study after study has shown that human (admin) costs for Windows are way higher per seat than for macOS seats. Whether that’s true for systems used as servers, I don’t know, but based on my own experiences ~ 7 years ago and earlier, it at least held true fir our mix of server and desktop deployments. The Linux per-system cost for admin appeared to be similar to the Mac cost for servers, but my experience was limited, so it’ll have to remain “appeared” until I read a credible study.

[Richard 12]

Re: And people….

The marginal support costs of Windows and Linux servers are both almost zero, as everything is automated.

Mac servers don't exist at all. Everything is a laptop/desktop, so the marginal support cost of a Mac "server" is roughly the same as the marginal cost of a Mac desktop. Possibly higher, as remote desktop is sloooow and there's a lot of things that are impossible to do on the command line.

Even things like installing certificates requires a GUI confirmation at first use.

Study after study...

That's the exact opposite of our experience. Care to provide some links?

[Tim99]

Re: And people….

The JNUC presentation by IBM CIO Fletcher Previn? Multiple links including: easytechsolver.com.

I had a close contact who was a very senior tech person at IBM who confirmed that the majority of people he worked with used Macs and Linux until a couple of years ago (when he retired) - He used mostly used macOS. The Windows users were generally those who supported customers' Windows based stuff...

[Richard 12]

Re: And people….

That site is worse than Retch!

I couldn't find any studies in there, just lots of assertions, with no evidence.

I admit I didn't read to the end though, I bounced off the massive amount of abusive advertising.

[Tim99]

Re: And people….

Ignoring the rest, it is a reasonable précis - A YouTube video of his IBM presentation is here.

He is now SVP & CIO at Cisco (like IBM, not necessarily near the top of my favourite companies) YouTube update 2023.

[Graham 32]

$4,000

$2.4M / 3 years / 200 Macs = $4,000 per year per Mac. That's the saving. Original cost must have been higher. Wow.

[Androgynous Cupboard]

Going the other way

I look forward to the legion of linked-in thought leaders penning earnest article on how to manage a “ground migration” to save money and reduce latency; can’t think what else to call exiting the cloud…

[gnasher729]

Real prices

The cheapest Mac Mini (quad core, gigabit Ethernet) is £599. One level below the M3 Ultra is a Mac mini with M4 Max, 8 cores, 10 Gbit Ethernet, for £2100. 200 of them are £120,000 or £420,000, and of course you spend money on racks etc. Anyway, a lot less than 2.4 million over three years. And I don’t know if they had 200 cores or 200 CPUs. 200 cores would be just 50 M4s or 25 M4 Max.

[Taliesinawen]

VM minis run in the past ..

> Singaporean super-app company Grab has dumped 200 cloudy Mac Minis and replaced them with physical machines, a move it expects will save $2.4 million over three years.

And save you from the key and mouse responses happening two seconds in the past /s