AWS Goes AWOL: Are We Too Dependent on the Cloud?

In this podcast, Motley Fool analysts David Meier, Tom King, and Tim Beyers:

• Discuss the failures that led to the AWS outage this week and which companies and services were impacted.
• Debate whether companies have become too dependent on AWS and its peers, especially when virtually all the in-demand AI services we're banking on are hosted in these clouds.
• Play another game of Faker or Breaker with three companies impacted by the AWS outage.

Don't wait! Be sure to get to your local bookstore and pick up a copy of David's Gardner's new book -- Rule Breaker Investing: How to Pick the Best Stocks of the Future and Build Lasting Wealth. It's on shelves now; get it before it's gone!

Tim Beyers: WTH at AWS. We need to talk about computing in the Cloud fools. You're listening to Motley Fool Money. Welcome Fools. I'm your host, Tim Beyers. With me are longtime fools, David Meier, Tom King. It's great to have you both here. So, guys, just a few hours ago, recording on Monday morning, Amazon Web Services suffered a pretty catastrophic outage in one of the Eastern US regions. And businesses that offer essential digital services have taken a hit here. This isn't the first time we've seen AWS go down, and it's not going to be the last, especially with the scale of the AI buildout of which AWS plays a pretty big part. So quick reactions here first. Tom, I'll start with you. Are you at all surprised by this news? Why or why not?

Tom King: I'm not. This is a complicated system, and in a complicated system, little things can compound. So it's almost inevitable that something like this would happen. It's very similar to what happened with CrowdStrike last year. Remember when CrowdStrike caused that big outage. It recovered from that. It got on. It is the cost of how things are these days.

Tim Beyers: Yeah. I mean, that's a fair point. I mean, these things are made up of, we've seen these commercials before, Dave, where you have the dominoes all lined up. You hit one of those dominoes and they all start falling. What's your reaction to this?

David Meier: To be frank? I'm surprised it's not happening more. Really, really think about this. Something bad happened there was an outage for a few hours. It's slowly everything's getting back online. Everything's back up and running. This is a massive system.

Tim Beyers: Yeah.

David Meier: The fact that it has the reliability that it does is an engineering marvel. So yeah, like, I get it. It's disappointing. Thank goodness it happened at 3:00 A.M. Eastern Time. I mean, I don't know exactly how much traffic is flowing through there. I remember in my backyard in Northern Virginia, 35% of the world's Internet traffic used to flow through my backyard at peak times. So I'm not surprised, but quite frankly, I'm impressed how quickly they recovered it. I'm surprised it doesn't happen more. Maybe it does and we just don't feel it.

Tim Beyers: I don't know if there is a Jeff Bezos shakes head approvingly meme, but I think you'd be getting that right now, Dave. I mean, it is interesting. So let's talk about this. AWS going dark, again, this has happened many times. Here's what we know. AWS, like you said, Dave, went offline around 3:00 A.M. Eastern Time for what appears to have been a couple of hours. The rolling effects could continue, and we are recording this on a platform called Riverside. As soon as I logged in, we got a nice little notice from Riverside saying, Hey, AWS went down and services may be affected. I'm paraphrasing there. So lots of companies have been affected here. At issue were some errors in the company's DynamoDB database. So let me just briefly explain what this is. Amazon's primary database is DynamoDB. It's a transactional database, and it's most famous for being the homegrown Amazon database that they use really to process transactions. So this is meant to be a highly resilient, highly replicated database throughout the Amazon and the Amazon Web Services ecosystem. This is what they've built their business on. What it was subject to, we don't know if it was an attack or just an error. It was subject to an outage in the DNS system, the domain name server system. When that happens, when the DNS goes out, and you probably have heard of this, like a distributed denial of service attack, that's meant to attack the DNS. If you do that, if you take down the DNS, it's like you having your phone and suddenly all your contacts, like, you can't press to call anybody because your contacts are frozen. The phone book in your phone just no longer works. You can't call anybody. When the DNS is down, that's what happens. The DNS determines when you say google.com, there's an underlying IP address that's a series of numbers, and it translates google.com to those numbers. When that's down, you can't do that, which means the Internet just stops functioning. So, this appears to be what happened here. I'm curious because the affected companies include Coinbase, Robinhood, Roblox, punch of others, two questions from me. I'll start with you this time, Dave. You said, effectively, you're surprised this isn't happening more. But I'd like to gauge your concerns. Are you at all concerned with how much influence AWS has over modern compute? And are we creating any kind of single point of failure here?

David Meier: So the answer is yes, there is concern.

Tim Beyers: Okay.

David Meier: I don't know how much I should be concerned, though, because this is the direction that everything is going. Many years ago, I was talking with some folks, and I'm like, When the Internet came up, we are always going to be going more digital, not more analog. AWS Azure Google Computing Platform, GCP, they all sprung up, and they all provide these types of services. So I don't think we're going to a single point of failure. There's not going to be one company that always does everything for everyone. But we don't have nearly as many companies doing this as when the Internet was first being created. There aren't as many IP companies things like that. So I don't know. You know, it's just like I said, it's kind of one of those things where it's sort of the cost of doing business. If you're not going to own your own hardware and manage your own hardware, you're essentially paying for the risk of someone else having an issue and your business being affected.

Tim Beyers: Yeah. I mean, Tom, I'm curious how you think about this because this is this is an issue. Digital businesses need to have that digital infrastructure to do business, and that infrastructure is now largely outsourced. So how do you think about this?

Tom King: It is one of the downsides that we must live with. We gained many advantages from the Cloud providers, and this is one of the downsides that if one of them breaks, they have a large impact on a lot of people. I'm sure you guys remember you've been around long enough to remember the days when you would have cold rooms in your office building that were stacked with computing equipment to deliver your website and the things that you sold over the Internet. We don't have those anymore. We outsource that to Google or Microsoft or Amazon, and we don't have to worry about keeping it working or keeping the equipment cold. Yes, it's one of the disadvantages of our current system, but we've gained many other advantages, I think is the way to think about it.

Tim Beyers: I love that you called server rooms, cold rooms.

Tom King: When I lived in Africa, you would go into the server room to cool down because of one of the only cold rooms in the office building.

Tim Beyers: Nice. I love that you called it that because you go into a server room, and, yes, it's intending to be cold. Get too close to a server, you stop being cold real fast. They run they run really hot. But I love that you called them that. Let's talk about risk here, though, Tom, because I think you're both right. So let's call this the cost of doing business. But how would you assess risk? Because AI is here. The AI buildout is going to continue, and it's going to largely depend on these cloud infrastructure companies quite a lot. Let's call it GPU hoarding. They're going to do the most GPU hoarding. They are going to provide the infrastructure. When the digital assistance you need is housed someplace else that we can't see and it's controlled by somebody else, how should we think about companies like Coinbase, Robinhood, and Roblox? Do they need to be valued with a higher risk premium because of their dependence on something like AWS? Tom, I'll go to you first on this.

Tom King: I don't think so. I think that companies will take the option that costs them the least, both in terms of money and time and headaches and so on. So if you really care about resiliency of always having your services available over the Internet, you would invest the money and the time and everything else to have your own servers. The reason they choose to go with one of the Cloud providers is because they gain many different advantages. It's probably cheaper for them. It's less work for them. It expands to fill the needs. There is no limit to how much demand their service can supply so that you don't, there's no limit on the hardware. Amazon handles that problem or Google or Microsoft. So I think companies are always going to make a decision that is best for them in their own unique circumstances. No, I don't think that any kind of additional risk needs to be considered here.

Tim Beyers: Dave, what do you think?

David Meier: I agree with what Tom was saying. I don't think there's any additional risk premium that's needed. Part of the reason is there's actually a market that's determining where people will spend their money. We don't have a single point of failure. This isn't a monopoly. If this behavior or if these incidences grew in frequency, what would happen? People would switch.

Tim Beyers: Yeah.

David Meier: So there's actually a massive incentive for these companies to make sure that the systems are reliable. Again, it happened early in the morning. There was an immediate response. It actually got back online pretty quickly. If it was that serious, they responded quickly, and I'm sure that they've learned, and we'll see what happens going forward. But I actually think the market is what controls the risk? Yeah, I realize it would be a pain to switch, but there are options available for substitution.

Tim Beyers: There are companies that are switching. Up next, we're going to talk about some fakers and some breakers. It's the faker or breaker game. You're listening to Motley Fool Money.

Fools, we're back. Thank you for listening to Motley Fool Money. We have a wide range of companies that were affected by the AWS outage today. Let's talk about three, guys. We're going to talk about three, and let's get your take on them. Are these AWS dependent businesses fakers or are they breakers? As a reminder, a faker is a company that shows outstanding growth for a brief but unsustainable period of time. They look good enough. They look enough like breakers to small fool some investors. So, Tom, coming to you first here. Faker or breaker, Coinbase.

Tom King: I would probably say faker because it's heavily dependent on the level of speculative activity in the markets, which goes up and goes down. When it goes down, it really, really hurts the people that are involved. So I would say faker.

Tim Beyers: I mean, this is an interesting one because the cryptomarkets are getting more and more real, more and more interesting. There is more real dollars flowing in particular into Bitcoin. However, this is probably one of the most rampant areas for fraud. With that, I'm going to give you Number 2 here, Dave, because it's a related company, Robinhood, which has made a lot of money on orchestrating crypto. So faker breaker, Robinhood.

David Meier: So Robinhood is not for me as an investor.

Tim Beyers: All right. I like you that you're qualifying already, but keep going.

David Meier: Well, we work for the Motley Fool. We tend to think about things differently. Robinhood is a type of competitor. I will say that this is still a breaker. The reason is is because you do have founders who had their vision. They knew what they wanted to do, and they were executing on it. The other thing is, they're serving a big market and one that's growing. And the other thing that in order to serve those markets, they continue to bring new products and services. So they have some sort of ability that they seem to be turning into an advantage to know, Hey, this is what our consumers want. So more people are coming, more people are staying engaged. Yes, whether it was free trades or now it's crypto or now it's prediction markets. Again, not for me. But that's OK. There's a lot of companies out there that don't serve me. But I see this company as having a lot of the traits of a rule breaker.

Tim Beyers: I'm going to bring up the third one, and I'm going to ask you both for a super quick take on this because I think it's a company some of maybe two of us. I know I'm going to be using this company later this week in order to go to Fool fest in DC, and that is Lyft. So faker or breaker Tom Lyft not Uber, Uber look alike.

Tom King: So Lyft is pretty much concentrated in the United States. Uber took the global expansion option. I think that Uber's size gives it certain advantages over Lyft. I know from being a regular or at least until a few years ago, a regular user of both that Uber was always cheaper and always quicker to get to you. They also probably lose more money than Lyft or were back then at the time. I mean, I think, it's hard to say, but, I mean, there has to be a Lyft because Uber can't have this market entirely to itself. Then we'd really be in trouble.

Tim Beyers: Is what you're saying. It passes the Cola test. It is the Pepsi to Uber.

Tom King: It is, yes.

Tim Beyers: So does that make it a breaker for you or it's still not enough?

Tom King: It's still not enough, but I think it's an important company. It has to exist to compete with Uber.

Tim Beyers: Last word, Dave, faker or breaker.

David Meier: Faker is hard, but I have to push in that direction. It's interesting. If I remember my history correctly, Lyft started before Uber.

Tim Beyers: I don't know.

David Meier: But I think that's correct. I think it was called Zimcar and then it morphed and then Lyft was created out of Zimcar. But it got usurped by Uber. Uber just went on, we're going to gobble up the world. We're going to take this market share strategy, and Uber has done a good job of rebounding. But I don't see it as something that can really take serious market share going forward. It'll still grow. It still provides an excellent service. It's the one I go to. I don't go to Uber. But, no, I don't think I can call it rule breaker. But faker is tough. Faker is harsh, in my opinion.

Tim Beyers: This is the game Dave, you got to make a call.

David Meier: No, I'm going with Faker, but I'm I'm giving you the little caveat.

Tim Beyers: Fair enough. Up next, we preview tomorrow. There's some interesting stuff happening. Speaking of prediction markets, there's your hint. You're listening to Motley Fool Money.

Back with a preview for tomorrow, you will have Emily Flippen and Sanmeet Deo and Jason Hall talking about prediction markets. What are they? How are they approved with regulators? How much money is flowing through them now? How can you invest in them? So, if prediction markets are a legitimate market opportunity, are they based on skill? Is it just another form of gambling? This is everything you're going to hear from Emily, Sanmeet and Jason tomorrow. But for today, I think we've concluded here guys, at AWS massive down today for a little period of time, probably going to be down again in the future. Doesn't mean that this is one that we should get too over hyped about, but a little bit annoying. But that's it for today.

Thanks to David Meier and Tom King for joining me today, guys. As always, people on the program may have interest in the stocks they talk about, and the Motley Fool may have formal recommendations for or against. So don't buy or sell stocks based solely on what you hear. All personal finance content follows Motley Fool. Editorial standards and is not approved by advertisers. Advertisements are sponsored content and provided for informational purposes only. To see our full advertising disclosure, please check out our show notes. That's it for Motley Fool Money Today. Again, please be sure you tune in for Emily, Sanmeet and Jason tomorrow for Dave Meier and Tom King, for our engineer, Dan Boyd and our producer, Anna, Chuck and Lou. I'm Tim Beyers. You've been listening to Motley Fool Money. Thanks and see soon Fools. Move on.

David Meier has no position in any of the stocks mentioned. Thomas King, CFA has positions in Alphabet, Amazon, CrowdStrike, Microsoft, Roblox, and Uber Technologies. Tim Beyers has positions in Alphabet and Amazon. The Motley Fool has positions in and recommends Alphabet, Amazon, Bitcoin, CrowdStrike, Microsoft, Roblox, and Uber Technologies. The Motley Fool recommends Coinbase Global and Lyft and recommends the following options: long January 2026 $395 calls on Microsoft and short January 2026 $405 calls on Microsoft. The Motley Fool has a disclosure policy.