The AWS Outage Was the Oligarch Stranglers Reaching for Substack's Neck

A single point of failure shows that Substack writers and readers aren't prepared for a coordinated attack, direct censorship, or even simple tech mishaps

If you woke up early Monday morning in the United States to play Wordle, you were probably sad. Or maybe you had trouble logging into Substack. Or, if you’re a writer here, maybe you had trouble sending newsletters.¹

If you didn’t wake up early on Monday morning, you may have missed most of it: Huge swaths of the internet went dark, including Wordle and a host of other sites.

The reason for the outage was a single point of failure: Amazon Web Services. This wasn’t just any old single point of failure. It was (likely) a misconfiguration of a teeny, tiny part of AWS, its DynamoDB service, which blew away AWS’s domain name service (DNS) handling.

DNS is a routing mechanism that tells computers where to get information on the internet, whether it’s a website or some other kind of service (like a newsletter posting service!).

Amazon Web Services (AWS) is a cloud-based service created by Amazon when the company was still struggling to make its shopping site profitable. It sort of runs the world.

I haven’t asked Jeff Bezos, so the motives for creating AWS are just a guess on my part, but I always figured that Amazon wanted to hedge its bets against the possibility that the original retail site might fail (viva Webvan!²), so it leveraged its existing data and deployment infrastructures into a profit center.

Whatever the reasons, it was a stroke of genius.

Today, AWS has 1.45 million business customers.

That includes a huge number of websites you use every day. It includes cloud services for smartphone apps you and your kids use — everything from Snapchat to Mailchimp, from Substack to another writers’ platform, Medium.

AWS logo by Amazon.com Inc., Apache License 2.0, via Wikimedia Commons; nuke image licensed from Adobe Stock; image smashup by author

If you’re wondering why some parts of some services you know stopped working and parts of other services for the same company didn’t, it’s because companies often split their services into different providers. For example, Semafor, a media company, reported that the main area of its affected services was newsletters.³ They also use different parts of the vast AWS infrastructure.

Almost every large company uses AWS or a cloud service like it, such as Microsoft Azure or Google Cloud. Many small ones do, too. In fact, small businesses and start-ups make up the fastest-growing part of AWS, and have for years. Last year’s growth rate among these kinds of businesses within the AWS domain was 42.2 %.⁴

Out of the three major cloud providers, AWS is the monster. When parts of it go down, the world shakes.

According to CNN, “Mehdi Daoudi, CEO of internet performance monitoring firm Catchpoint, estimates the total financial impact of the AWS service disruption will be in the billions of dollars.”⁵

See the Notes at the end of this post for some of the companies that use it and/or reported outages.

Why????

Good question. The reasons why AWS has such a large foothold could fill a book, and probably have.

I’ll relate my personal experience to give you an idea of what is going on here. Try to bear with me as I explain the why of it all before I get into the danger zone of it all.

When I was a software engineer for big tech firms, the teams I worked on spent all day coding and programming. We didn’t worry about deployments much. That was someone else’s job. The only time I had to think about deployments was when I was asked to manage production rollouts, but that was mostly a process for merging different versions of the code base from (for example) a successful test environment into the production queue.

The process was arduous because it sometimes involved merging thirty or forty different code base versions (branches) into one.

Hot tip if you’re in the software biz with huge code bases: Don’t get a reputation as an expert at merging other people’s code. You’ll find yourself doing it a lot, often in the wee hours of the morning.

The process of managing things like server loads and other arcane (to most people) tasks related to software deployments was handled by teams of what we called “release engineers.”

For a long time, companies managed all of that internally, but as AWS built its cloud, it succeeded at luring companies into outsourcing an ever-growing number of their production release processes.

At first, this mostly meant server management. If a company encountered a spike in web traffic, all they had to do was click a button (I’m greatly simplifying this process to illustrate a point), and additional servers could be added to handle the load.

This process works well for startups, too. If you can focus on code development, and your release team can be whittled down to some people who know how to work with and deploy to AWS, you can cut your costs significantly because you don’t need to worry about buying computers (which, compared to people, are almost free, but bean counters have their principles).

AWS makes it easy to start small, and if you’re lucky enough to hit the big time, it’s easy to scale up.

The number of your AWS servers might start at two, but quickly expand to thirty or even a hundred once your new business idea starts turning into a gold mine.

As time went on, a cottage industry was born. Technical specialists with expertise in AWS filled the skies, eager to help the next company that wanted to deploy using AWS. AWS grew in complexity. It became a behemoth that no single person could master.

One of the classic case studies of a company transitioning to AWS is that of Netflix, which in 2008 suffered catastrophic database failures. Yury Izrailevsky, who was Vice President of Cloud Computing and Platform Engineering when the eight-year transition to AWS was completed, explains what prompted the move:⁶

Our journey to the cloud at Netflix began in August of 2008, when we experienced a major database corruption and for three days could not ship DVDs to our members. That is when we realized that we had to move away from vertically scaled single points of failure, like relational databases in our datacenter, towards highly reliable, horizontally scalable, distributed systems in the cloud. We chose Amazon Web Services (AWS) as our cloud provider because it provided us with the greatest scale and the broadest set of services and features.

There’s that phrase again: “single point of failure.” There’s another phrase there that matters: “distributed systems.” That sounds like tech bro talk, but all it means is that Netflix’s databases are no longer hosted at a single Netflix datacenter in Campbell, California (its HQ city) or wherever they were before the transition to AWS.

When I worked at eBay a million years ago, we actually had a server room. It was a dark, scary place with a lot of blinking lights guarded by hooded dudes in sunglasses. Don’t tell anybody, but I even hosted a small public-facing test application from a single Linux box in a cube a few steps away from me. That doesn’t happen these days.

These days, everything for most companies, including Netflix, is hosted in several different locations within the AWS infrastructure. Data corruption is considerably less likely to ruin the company because several more sites, in multiple locations, host the data than before the transition.

Izrailevsky continues:

Leveraging multiple AWS cloud regions, spread all over the world, enables us to dynamically shift around and expand our global infrastructure capacity, creating a better and more enjoyable streaming experience for Netflix members wherever they are.

We rely on the cloud for all of our scalable computing and storage needs — our business logic, distributed databases and big data processing/analytics, recommendations, transcoding, and hundreds of other functions that make up the Netflix application. Video is delivered through Netflix Open Connect, our content delivery network that is distributed globally to efficiently deliver our bits to members’ devices.

Yay, another tech bro phrase: “multiple AWS cloud regions.” Another one he uses (not quoted) is “redundancy,” which simply refers to the fact that there are multiple instances of things like databases that duplicate the same information or functionality as other instances.

The AWS cloud is distributed at several different points throughout the globe. For years, as the AWS infrastructure grew, new Amazon data centers seemed to pop up at least every month around the U.S, then the globe.

In theory, this distribution methodology should lessen the impact of outages. Obviously, that wasn’t the case on Monday.

AWS does much more than provide web service deployment options. You can store all types of databases there, you can host websites, you can leverage their AI services, to name just a few.

I use something called S3 to store a PDF of short stories that my paying subscribers receive a link to when they subscribe (if you’re a subscriber to my free tier, you may have received that link, too, but that is changing).

S3 is pretty neat. I can store a file there, then make it directly available to folks using an HTTPS address, and not worry that the file will be corrupted or hijacked somehow. It’s like having a hard drive out in the wild and a web server all at once, and because I use it so infrequently, it’s free (AWS lures you in with a free tier).

There’s almost no work involved for me here. It’s plug and play. I upload the file, set the permissions, then copy the URL from the AWS listing for my file and distribute the link.

Here’s a list of some of the most popular AWS services (or the ones they are pushing the hardest), in their words:

• Amazon Quick Suite
  Digital workspace that unifies AI agents for research, business insights, and automation

• Transform
  Agentic AI to accelerate modernization of .NET, mainframe, and VMware workloads

• Aurora
  Serverless relational database service for PostgreSQL, MySQL, and DSQL

• Amazon Bedrock
  Managed service for building and scaling generative AI apps with foundation models

• Amazon Connect
  AI-native omnichannel cloud contact center

• EC2
  Secure and resizable compute capacity for virtually any workload

• Nova
  Foundation models delivering frontier intelligence and top price performance

• SageMaker
  The center for all your data, analytics, and AI

• S3
  Virtually unlimited secure object storage for AI, analytics, and archives

In all, there are 264 AWS product offerings.

The one in that list we’re interested in is EC2

At the time I began writing this, Amazon was saying that its EC2 service was the source of the outages. By the time I finished writing this, the source of trouble seemed to be centered around a bad update to AWS’s DynamoDB service, which then corrupted a Domain Name System (DNS) component, which then blocked access to what seemed to most of us to be about half the internet.

EC2 was heavily affected by the misconfiguration, which is probably why some initial reports focused on it.

EC2 is the service most of us use when we want to build and deploy websites and scale them up to higher traffic when we need to. Here’s the AWS product description:

Amazon Elastic Compute Cloud (Amazon EC2) offers the broadest and deepest compute platform, with over 750 instances and choice of the latest processor, storage, networking, operating system, and purchase model to help you best match the needs of your workload. We are the first major cloud provider that supports Intel, AMD, and Arm processors, the only cloud with on-demand EC2 Mac instances, and the only cloud with 400 Gbps Ethernet networking. We offer the best price performance for machine learning training, as well as the lowest cost per inference instances in the cloud. More SAP, high performance computing (HPC), ML, and Windows workloads run on AWS than any other cloud.

In other words, we can use EC2 to spin up just about any kind of offering, using whatever operating system we’re comfortable with, and AWS handles the traffic.

This is catnip to someone like me. Whenever I’ve worked for small startups or even built my own website, I’ve used EC2. It’s been around for years. I want to say “We’ve all used it,” but that’s an overstatement. Most of you aren’t tech nerds. Still, it’s incredibly popular.⁷

So popular that it has the potential to bring down the world.

Or, worse, maybe just the resistance and opposition infrastructure against Trump, because we now have a new single point of failure.

Lina Khan and the threat against the oligarchs

Lina Khan, Joe Biden’s Federal Trade Commission (FTC) chairman from 2021 to 2025, wrote a seminal work about antitrust and monopolies when she was a Yale Law Student. The paper she wrote, titled Amazon’s Antitrust Paradox,⁸ scared enough crap out of monopolists and oligarchs everywhere to fill up the mad clown’s AI jet fighter and send it off to dump it all on all the world’s protesters.

The core principle behind her paper was that pricing issues are a poor measurement of corporate dominance and monopoly. Amazon, she observed, owned major pieces of disparate businesses that dominate the economy in ways most people don’t think about.

From logistics (Amazon Freight Services)⁹ to medicine and pharmaceuticals, Amazon’s tentacles are in almost every aspect of American life. AWS is another example of this.

And now that Amazon’s main beneficiary, Jeff Bezos, has freed himself from everyday decision-making, he has moved on to other projects, such as spaceships and the Washington Post, which he is in the process of trying to mold into his image, with some welcome resistance from the few determined, independently-minded journalists who still work there.

Amazon, Khan wrote in her paper, became masters of their customers’ future:

Amazon’s willingness to sustain losses has allowed it to engage in below-cost pricing in order to establish dominance as an online retailer. Amazon has translated its dominance as an online retailer into significant bargaining power in the delivery sector, using it to secure favorable conditions from third-party delivery companies. This in turn has enabled Amazon to extend its dominance over other retailers by creating the Fulfillment-by-Amazon service and establishing its own physical delivery capacity. This illustrates how a company can leverage its dominant platform to successfully integrate into other sectors, creating anticompetitive dynamics. Retail competitors are left with two undesirable choices: either try to compete with Amazon at a disadvantage or become reliant on a competitor to handle delivery and logistics.

The thousands of retailers and independent businesses that must ride Amazon’s rails to reach market are increasingly dependent on their biggest competitor.

Khan’s dream, before Kamala’s curious (some might say suspicious¹⁰) loss in the election, was to begin restoring “traditional antitrust and competition policy principles or applying common carrier obligations and duties.”

In other words, break up some of the damn conglomerates that are eating the world.

One of the ideas Khan floated was, for example, to break up Amazon into its various constituent parts such that AWS would become a separate company.

It’s debatable that she would have been successful, given the makeup of the John Roberts court, which has basically been purchased by oligarchs through the Federalist Society,¹¹ but she wanted to give it a shot.

The reason for the improbability of her success is that she would have needed to rely on federal lawsuits. Even successful ones would have probably been struck down by the oligarchy’s friends at the Supreme Court.

But if she had been successful, the ramifications might have stretched across the conglomerate world and cut their ravaging canines down to size by forcing divestiture from companies like multinational agribusinesses, who steal small farmers’ land through bankruptcy proceedings and distress sales, sow the newly acquired land with commercial seed, and sell highly processed foods from the harvest, thus, ultimately, giving birth to obesity miracle drugs like Ozempic.

Wow, what a supply chain. What a world.

The oligarchs strike back

The oligarchs weren’t taking their chances. They outsourced attack vectors against her through the outrage media machine. The fact that Khan was a woman with Pakistani heritage lured a deluge of angry fanatics to the cause, many of whom didn’t even really care or know anything about why they were mad at her aside from the fact that she was a woman of color in a position of enough power to be a bit of a pest to the world’s corporations.

Lina Khan, public domain photo

The oligarchs spent hundreds of millions of dollars in campaign contributions to defeat Kamala and rid the world of the threat posed by Lina Khan, led by Elon Musk’s $200 million.

In normal times, business titans just roll with the flow. But the oligarchy sensed a true enemy in Khan and set out to destroy the path she was on.

Now, that path has been obliterated and replaced by one paved in gold.

If she were in office now, she’d be jumping all over the most recent outage. Anti-trust lawyers would have scurried to their laptops to draw up antitrust lawsuit papers, and maybe, at the least, a discussion about the future of Amazon hegemony would have filled business journals for a few days.

Instead, the world will shrug it off until darkness falls again.

We can’t even dismiss the possibility of direct interference by the Trump regime. It isn’t hard to imagine the mad clown demanding that Amazon cease providing services to sites like Substack and other pests.

If you ever bang your head against the wall, crying in wonderment why American business titans lined up behind Trump, and still do, there’s your answer. They smelled an existential threat in Lina Khan, one that was completely overblown given the complexity of antitrust law.

If Substack goes dark sometime in the near future, now that you’ve read this, you’ll know why.

Notes

If you’re wondering where U.S. government servers are located, I’m pretty sure it’s a mish-mash, but I think the government tends to lean more towards Microsoft Azure. Maybe one or two of my readers have some insight there. If so, drop a comment!

It’s fair to guess that the courts will impact Amazon’s bottom line without Lina Khan’s help, if Mehdi Daoudi is correct that the outage cost billions of dollars in losses. They’ll be hammered with some nasty class action lawsuits.

Meanwhile, at least one AWS engineer is out there looking for a new job. Maybe DOGE should hire them.

Here is just a small list of companies that use AWS

These are not comprehensive lists.¹²

1. Leading spenders:
   Sony $11M

   Adobe: $7.5M

   Facebook: $5.6M

   Johnson & Johnson: $5M

   3M: $5M

   Coca-Cola: $5M

   Twitter: $3.7M

   Netflix: $2.4M

   ESPN: $2M

   Reddit $2M

   Twitch: $1.3M

   Samsung: $1.2M

   Walt Disney: $1.2M

   Okta: $1M

   Canva: $981K

• Some major companies that use AWS:

  3M

  Airbnb

  Canva

  Coca-Cola

  Formula 1

  GoDaddy

  Goldman Sachs

  Heineken

  Johnson & Johnson

  LaunchDarkly

  MAN

  McAfee

  Moderna

  Netflix

  Okta

  Panasonic

  Pinterest

  Salesforce

  Samsung Electronics

  Siemens

  ŠKODA

  Snap

  Starbucks

  State Farm

  Thomson Reuters

  Toyota

  TUI

  Vanguard

  Verizon

  Warner Bros. Discovery

  Wealthfront

  Zalando

  Zomato

• Some websites and/or apps affected by the outage (various sources)

  Amazon.com

  Prime Video

  Alexa

  Robinhood

  Snapchat

  Reddit

  Perplexity AI

  Venmo

  Canvas by Instructure

  Crunchyroll

  Roblox

  Whatnot

  Rainbow Six Siege

  Coinbase

  Canva

  Duolingo

  Goodreads

  Ring

  The New York Times

  Life360

  Fortnite

  Apple TV

  Verizon

  Chime

  McDonald’s app

  CollegeBoard

  Wordle

  PUBG Battlegrounds

  OpenAI

  Vimeo

  Twitch

  Shopify

  Google Maps

  Claude (Anthropic)

  Cursor

  Dialpad

  reCAPTCHA

  YouTube

  Khan Academy

  NPM

  Dragon Ball

  AT&T

  DoorDash

  Spotify

  Google Cloud

  Discord

  Google

  Google Meet

  Character.AI

  Rocket League

  Cloudflare

  Google Nest

  Pokémon Trading Card Game

  FuboTV

  HighLevel

  Box

  Etsy

  Google Drive

  Mailchimp

  Lyft

  Signal

  WhatsApp

  Airbnb

  Disney+

  Lloyds Bank

  Bank of Scotland

  Vodafone

  BT

  HMRC (UK tax authority)

  UK government websites

  Medium

Why is Google Cloud in that list? I have no idea.

Temporary Deal Alert!

Can’t deal with subscriptions? I don’t blame you. Perhaps you’ll consider this:

Or, share:

Thanks for reading, liking, commenting, sharing, and all the other things you do to keep me motivated.

Footnotes

1. I read reports of login issues, but the Substack status page makes no mention of it. I didn’t have any issues.

https://substack.statuspage.io/

2. Contributors. 2004. “Defunct Ecommerce Business.” Wikipedia.org. Wikimedia Foundation, Inc. May 8, 2004. https://en.wikipedia.org/wiki/Webvan.

3. J.D. Capelouto. 2025. “🟡 Semafor Flagship: A Single Point of Failure | Semafor.” Semafor.com. October 20, 2025. https://www.semafor.com/newsletter/10/20/2025/semafor-flagship-a-single-point-of-failure

4. Michalowski, Mariusz. 2023. “Who’s Using Amazon Web Services? [2025].” Spacelift. July 24, 2023. https://spacelift.io/blog/who-is-using-aws.

5. Olesya Dmitracova, Charlotte Reck, and Ana Nicolaci. 2025. “Amazon Says Systems Are Back Online after Global Internet Outage.” CNN. October 20, 2025. https://www.cnn.com/business/live-news/amazon-tech-outage-10-20-25-intl?post-id=cmgzdye9q00003b6ngdv69nyh.

6. “Completing the Netflix Cloud Migration - about Netflix.” 2016. About Netflix. 2016. https://about.netflix.com/en/news/completing-the-netflix-cloud-migration

7. Microsoft Azure has broken through the competitive barrier significantly over the years. Many release engineers prefer it. I honestly don’t know much about Google Cloud, aside from the fact that I always see it “powering” things like statcast operations for sports broadcasts.

8. Khan, Lina M. 2016. “Amazon’s Antitrust Paradox.” Yalelawjournal.org. 2016. https://www.yalelawjournal.org/note/amazons-antitrust-paradox.

9. https://freight.amazon.com/

10. She Won. They Didn't Just Change the Machines. They Rewired the Election.

11. The Federalist Society’s War Against Americans

12. Sources:

Michalowski, Mariusz. 2023. “Who’s Using Amazon Web Services? [2025].” Spacelift. July 24, 2023. https://spacelift.io/blog/who-is-using-aws.

‌“AWS Outage: A Complete List of Every Site and App That Went Down.” 2025. Dataconomy. October 20, 2025. https://dataconomy.com/2025/10/20/aws-outage-a-complete-list-of-every-site-and-app-that-went-down/.

‌HT News Desk. 2025. “AWS Outage: Full List of Sites and Apps Affected by Amazon Cloud Service Issue | World News.” Hindustan Times. October 20, 2025. https://www.hindustantimes.com/world-news/aws-outage-full-list-of-sites-apps-impacted-amazon-canva-down-roblox-perplexity-robinhood-101760948068768.html.

“Status Overview.” 2025. Downdetector.com. 2025. https://downdetector.com/status/aws-amazon-web-services/map/.

‌