back to article IT crisis looming: 'What if AWS goes pop, runs out of cash?'

The big public infrastructure-as-a-service (IaaS) players may be on the brink of a crisis as cataclysmic as the 2008 banking crash, as they slash prices while spending billions on building out and staffing their operations. Or so claimed Steve Brazier, CEO at channel beanie Canalys, who warned resellers steering their …

  1. Destroy All Monsters Silver badge

    I like this thinking!

    "The big public infrastructure-as-a-service (IaaS) players may be on the brink of a crisis as cataclysmic as the 2008 banking crash"

    Yes, but the whole economy is also on the brink of a crisis far more cataclysmic than the 2008 banking crash and the 2001 crash. Where is your credit now?

    Let's invest in a couple of on-premise servers.

    How Financial Bubbles Fester And Burst—Even As The Fed Says Not To Worry

    In today’s post Wolf Richter offers some solid insights on the dynamics of financial bubbles which merit further comment. The starting point is to recognize that once they gain a head of steam, financial bubbles tend to envelope virtually every nook and cranny of the economy, creating terrible distortions and destructive excesses as they rumble forward. In this instance, Wolf Richter explains how Silicon Valley has once again (like 1999-2000) been transformed into a rollicking capital “burn rate” machine that has spawned a whole economy based on striving for bigger losses, not better profits.

    This latter development—- currently exemplified by 44 VC start-up companies in the IPO pipeline with a valuation of more than $1 billion each, despite no earnings and scarce revenues—-is indicative of late stage bubble dynamics. Say January 2000!

    And also:

    "He said nobody predicted the financial meltdown in the banking sector"

    He says bullshit

  2. Anonymous Coward


    ...random analyst, plucks random data out and makes 1+1 = -50.

    Many companies run many projects at a loss as they build up, heck even most little shops run at a loss in the first year or two. This is not news or even newsworthy.

    They are expanding at a massive rate and building out at huge cost. As the market slows down and the "me too's" drop off, the big boys will pick up a lot of the slack, and slow down the expansion as the market matures. That's when they will start to make money.

    1. YetAnotherPasswordToRemeber

      Re: Sigh....

      I think that was the point of the article but, as I read it, there's only going to be Microsoft and Google being there in the end because they're the only ones in the market with pockets deep enough, and while AWS is big it's losing $2B per year, which is a pretty unsustainable business model.

      1. Tom 13

        Re: I think that was the point of the article

        No, that's the starting point of the article and a bit obvious. The point of the article is that because of the hype surrounding the cloud, and because the cloud has so undercut the pricing structure of traditional deployments, a large percentage of businesses have moved to the cloud. If/When these cloud services other than Google and MS fail, those businesses will take a huge hit and some of them could be headed under. If enough of them are headed under, he equates it with the banking problem. In other words, the contraction process will happen faster than the market can react which causes a catastrophic collapse.

        Not sure I concur with the level of alarm he has, but it certainly is a risk that should be evaluated.

        1. Peter2 Silver badge

          Re: I think that was the point of the article

          I don't use the cloud for anything (well, Azure Multifactor but that's not exactly mission critical) and personally i'd find it hillarious if the cloud rained and took out much of our competitors IT infrastructure with it.

          Sadly, it'll never happen. The other cloud suppliers would buy and operate Amazon before they would see it die, because if a major cloud supplier just evaporates then quite honestly then I think that's going to be curtains for virtually the entire cloud computing sector.

      2. Anonymous Coward
        Anonymous Coward

        Re: Sigh....

        "there's only going to be Microsoft and Google being there in the end "

        Microsoft are already making a lot of money in cloud with Azure - and are about to overtake Amazon AWS in revenue. Google aren't even close to being in the same league as Azure and AWS. Rackspace probably have a larger cloud business than Google does.

        So if Amazon fold - its likely going to be Microsoft Azure as the one big remaining player - and lots of much smaller bit players.

        1. Smitty Werbenjaegermanjensen

          Re: Sigh....

          My concern is less that AWS might fold and take systems offline. I doubt it will go pop, rather it would deflate more slowly and give people an off-ramp.

          I'm more concerned that, as the level of competition drops away, both pricing and lock-in techniques will increase. Fewer players means less choice and I'd not be surprised to see a big duopoly happening, especially when the financial barriers to entry become significant.

  3. Rich 2 Silver badge

    Yes they did

    Off topic but....

    "Nobody predicted the financial meldown..."

    Actually quite a few people did but nobody listened.

    1. Anonymous Coward
      Anonymous Coward

      Re: Yes they did

      I was listening to a Radio 4 program while waiting in a traffic jam in the spring before the meltdown happened, and the general theme was that the financial derivative market was too complicated and convoluted to be sustainable. The guests specifically mentioned the examples of Freddie Mac and Fannie Mae sub-prime mortgages being packaged into derivative products such that their risk and real (lack of) value was obscured from the investors.

      So when it all kicked off, I was not surprised. I was surprised by the then UK Government's reaction, which was too slow to start, and then became too severe, and all the time talking about Prudence (whoever she was) while draining the last out of the already depleted coffers.

      Should become a case study in how you should not spend more money that you have in a boom period, but keep some reserve back in case something happens. Would be worth studying both by governments and large companies, all of whom believed that the good times could not end!

      1. Roland6 Silver badge

        Re: Yes they did

        >I was listening to a Radio 4 program while waiting in a traffic jam in the spring before the meltdown happened

        Yes 2007 was an odd year, with many people with a London financial services focus regarding the sub-primes being a wholly US problem and nothing to worry about, and were shocked when things started to crash around them in London ...

  4. John Tserkezis

    They're not losing anything. At least not what the doomsdayers claim.

    If AWS really did lose $2bn, they would have folded by now.

    On paper, Amazon's profits as a whole make the local street lemonade stand look like a cut-throat money making venture. Whether or not you like it, the books are cooked to put any and all profit back into the business so it looks like they're making nothing. If this doesn't give you a nice number that you can ooh and ahh about, then tough, deal with it.

    Face it, they've been running 20 years, and people are still throwing money at them - and getting their returns too.

    Smells a bit like the competition wants to make them look "flakey" so they can drum up a bit of work for themselves.... Funny thing is, a lot of the cloud competition offer services that don't directly compete with AWS, so they're not even treading on each other's toes anyway. Someone's really insecure here, and it ain't AWS.

    1. Charlie Clark Silver badge

      Re: They're not losing anything. At least not what the doomsdayers claim.

      I don't understand either the analogy or the conclusion.

      Amazon's profits are wafer-thin because it's in low-margin, high-volume, high-capex business. Any ROI for investors in stock-price only and that because of the expectations of future business because it certainly isn't justified by current profits. I personally like the way Bezos tries new stuff, though I do wish he'd get out of the business of owning warehouses and actually shipping physical products.

      As someone else pointed out the other day: in hosting you don't pay for what you use but for what you provision and this is the achilles heel of the whole model. At some point vendors will have to drive up their yield in order to make money or go out of business. That will make the whole system much less flexible unless derivatives are introduced allowing resource trade between users.

    2. Tom 13

      Re: they've been running 20 years,

      So were Freddie and Fannie before the collapse.

      If you're cooking the books, the market will eventually catch up to you. The hurt that hits you when it does I don't give a rat's arse about. The catch is, with something this big there's a lot of collateral damage. That I do give a rat's arse about.

    3. Getriebe

      Re: They're not losing anything. At least not what the doomsdayers claim.

      I'm with this analysis. I am amazed he can pull the 2billion figure out and say its is correct. I don't doubt they are running at an operating loss, but how much, what proportion of corporate turnover, what's the cash and so on are all ethereal in large companies.

      Bezos also runs the company, afaik, so as not to state big profits

      AWS might go tits, but this analysis has a low percentage chance of being correct.

  5. Captain Server Pants

    Something tells me the bankruptcy Brazier really cares about...

    Is his own.

  6. ElNumbre

    Race to the Bottom

    I suspect in the race to the bottom, only the services which are supported by businesses with other interests will survive, particularly at the massive compute scale. AWS has a bookstore to help them out, and Microsoft have O365 and its continual licencing to support their Azure compute cloud. The smaller players need to find a niche and play to that strength.

    The other thing to bear in mind is that cloud is not necessarily cheap; in our business the break-even point between doing it on cloud and buying tin averages around 21 months. But you do move your spend from CAPEX to OPEX and you struggle to pull the eject cord once you're in flight, especially if you run a complex environment in there.

  7. Anonymous Coward
    Anonymous Coward

    Do the math

    It probably takes an upstart cut-rate VPS provider 6-12 months to pay off its hardware, maybe 2-3 years to turn a profit. I'm running bloated friggin LAMP+Java stacks on it (most of the world runs at least 5-10 years behind the leading edge) and I can't complain about price or performance or uptime... It does one thing well... It's the best ever. AWS costs more but offers more. So I'm not worried about its profitability.

  8. Quentin North

    Wouldn't it be funny if...

    Amazon was to go bust and Apple bought the whole lot.

    1. Anonymous Coward

      Re: Wouldn't it be funny if...


  9. Anonymous Coward
    Anonymous Coward

    Amazon need to piss away billions across their whole organisation...

    That way they can avoid showing any profit so they can continue to avoid paying tax.

  10. James 100


    The "banking crisis" was fear of a collapse in clearing bank services: that our salary payments would fail to transfer, we'd be unable to pay our bills and mortgages or buy food. That really would have been a crisis: headlines of "Nobody paid, direct debits all bounce, starvation and looting in the streets".

    An Amazon outage? We've had those before: my company had a pair of EC2 instances in Dublin knocked offline back when they had the mysterious cascade power failure. Yes, it meant some downtime; if it had lasted more than a few hours, I could have picked up a pair of VMs from some other hosting company and restored backups to it - meaning a day's downtime for DNS caching to update. No missed mortgage payments or starving kids: "Big service provider closes suddenly, everyone moves to rivals"?

    It's very hard to imagine Amazon suddenly pulling the plug, too: more likely, either they'd hike prices to sustainable levels (if current levels really aren't), or announce a planned shutdown, giving us time to move. Yes, maybe the alternatives would cost more ... where's the scary headline there? "Hosting providers hike prices, web hosting to cost more next year"? Sorry, not scared.

    Apart from anything else, in the estimates of AWS revenue, are they really factoring in Amazon's own heavy use of the infrastructure for the corporate web presence? (That use probably won't show up in "revenue" since it's internal.)

  11. Anonymous Coward

    What, me worry?

    Amazon takes their client's OpEx and uses as CapEx and OpEx for their datacenters so that they have the infrastructure for their their season, the holidays. It's a real nice gig and if you bother to pay attention, this guy ain't paying any attention to the bookkeeping involved here. I suppose one could mothball it for ten months out of the year but that doesn't make any sense either. Lighting off is when most electronics gets downright cranky.

    Doesn't understand the (warehouse store with deliveries!) business model. Seems Google does, getting people to do the same for their (search) business model. Some analyst! And I have a feeling that the one's needing a bailout aren't Amazon nor Google. Yes, the government. "May we hear from the esteemed Senators from Google and Archer-Daniels-Midland?" That's the logical next step towards person-hood.

    1. SiliconSlick

      Re: What, me worry?

      re: holidays

      Exactly... their losses for AWS are far outweighed by their ability to have capacity during the Christmas season which I'm sure more than makes up for it. Lost sales are much worse than losing a bit of money to maintain capacity.

  12. Ken Moorhouse Silver badge

    Cloud Ownership

    I think one of the issues with the cloud is ownership of the bit that you want to rent. Businesses will want a lot of reassurances that their data is safe with the provider that they have rented through, even as much as to say that they have the ability to go to the premises holding their data and be able to touch the machines that hold their data. In some cases this is to put their minds at rest regarding Data Protection legislation. These assurances may be a "false sense of security" if you look deeply enough, and that is a reason not to put too much faith in the cloud, full stop. (Everything depends for security on logging in to your data: if your connection to your server is compromised, so is your data).

    But ignoring that for this scenario, businesses may well go through thorough "vendor evaluation" and plump for their chosen cloud provider. Now if that provider sells up, or changes their "feed", there is the worry as to who they will pass the data to. It could be anywhere in the world, to an organisation they have no control over. There has to be a clause in these types of contracts where cloud users have the option when a cloud moves from one owner to another to opt out and move their data to another provider of their choice without penalty, regardless of how long the contract has locked you in for.

  13. IanW

    Author fell into the usual trap

    Better to see how Amazon are really doing by looking at the volume of their cash flow and its trends. That demonstrates that Bezos has more than enough coverage to sink costs into AWS for years to come. As such, the authors headline is an attention grab, but I'd fully expect Amazon AWS (and Digital Ocean for that matter) to be there as the market consolidates.

  14. Sirius Lee

    Have you taken a look at Amazon's 10Ks and 10Qs?

    These are the documents all public companies are required to file with the SEC quarterly (Q) and annually (K). They paint a picture of a company investing for future. Sure, the company made a loss of over $100 million. But Amazon's stated intent is to not make a profit. On their revenues the loss is within a very small margin of error and covered many time over by their cash reserves. In other words, this is a company managing its finances very precisely and erring on making a small loss [relatively speaking] rather than pay any corporation tax. And this despite making huge capital investments. You can see that Amazon lowered it's cash holdings to 'only' $5bn by an amount almost as much as their investment in kit. Of course this was not the stated reason as a note to the accounts points out that sellers are getting paid earlier.

    May be it's because we live in a time when companies no longer make major capital investments for future because it's no longer acceptable to Wall Street who want the money NOW! No software company needs much capital (for hardware) and even those that 'produce' hardware such as Apple offload the expense and risk to Chinese companies like Foxconn.

    So it's not normal, and therefore note worthy to an analyst, when a company does something different. However, it's not unusual for new businesses sectors to require unusual investments upfront. Trains, telephones and broadcast TV are examples of sectors that were being funded in far greater amounts than their incomes justified the respective early days. With hindsight these were sensible investments though it probably didn't seem that way at the time.

    As for the CIA, my understanding is that they are using AWS kit and software by the container load but in-house. It seems low risk to me. If AWS were to collapse, the CIA already have the hardware and software and would then have ready access to people who were looking for a job. The kit being used is commodity, right? That mean's it's available everywhere, right? So presumably even a government could procure some.

  15. The Godfather

    Things pan out..

    Headline grabber even if it does touch upon the potential frailty of any such outsourced service. This is largely about investment and it's nothing like a share price based on hot air IPO's.

    AWS remains pretty sound although they do face quite a battle with MS and Google.

    Has anyone considered the 'soundness' of 85% of channel partners?

  16. The Empress

    Who knew?

    Who knew IBM would actually get something right by forcing the cloud sector to at least make their budget?

POST COMMENT House rules

Not a member of The Register? Create a new account here.

  • Enter your comment

  • Add an icon

Anonymous cowards cannot choose their icon

Other stories you might like