Monday’s Amazon Web Services outage — and the global disruption it caused — underscored just how reliant the internet has become on a small number of core infrastructure providers.

The ramifications of such outages could only get worse if artificial intelligence becomes as central to work and daily life as tech giants suggest it will in the coming years.

Monday’s outage briefly blocked some people from scheduling doctor’s appointments and accessing banking services. But what if an outage took down the AI tools that doctors were using to help diagnose patients, or that companies used to help facilitate financial transactions?

It may be a hypothetical scenario today, but the tech industry is promising a rapid shift toward AI “agents” doing more work on behalf of humans in the near future – and that could make businesses, schools, hospitals and financial institutions even more reliant on cloud-based services. A global survey of nearly 1,500 firms published by McKinsey & Company in May found that 78% of respondents already use AI in at least one business function, up 55% from a year earlier.

“If there’s an outage and you rely on AI to make your decisions and you can’t access it, that’s going to have an effect on performance,” said Tim DeStefano, associate research professor at Georgetown’s McDonough School of Business.

Monday’s outage had such a widespread impact because many companies rely on cloud providers for the backend functions that support their businesses, such as virtual server space, storage or developer tools. Typically, this set up is more affordable, flexible and secure for those customers, except when AWS experiences an outage. Then it’s effectively a single point of failure for a huge swath of the internet.

To be fair, these services are remarkably sturdy considering the scale of their operations. But outages like Monday’s raise questions abouthow crucial tech services could be made even more reliable.

AWS serves millions of customers, from retailers and restaurants to financial services firms and government agencies; it holds around 37% of the cloud computing market, according to Gartner. Together with Microsoft and Google, the three companies service around 70% of the market.

And the consolidation of the internet’s backbone is continuing in the age of AI. While there’s some grappling between the big three, Amazon, Microsoft and Google remain by far the prominent cloud computing providers for AI applications, according toEmarketer senior analyst Jacob Bourne — and their futures depend at least in part on serving AI demand.

While websites and apps can still technically function using their companies’ own less powerful on-premises servers, “cloud computing represents a technological prerequisite for using AI,” DeStefano said. That’s because the computers needed to run AI tools are powerful and expensive, and on-site hardware isn’t as easy to modify as business needs change. It just makes more sense to rent that computer space and pay for it only as needed.

And as AI becomes more widespread, data center outages could happen more frequently since AI models are so power-hungry, Bourne said. Major cloud providers, including Amazon, Microsoft and Google, are spending billions on data centers to address this growing need.

The risk of serious disruption from an outage rises considerably the more companies rely on AI agents to do critical tasks and automate the work of humans, a transition that’s already in progress despite disagreement about just how far it will go.

Tech companies are relying on AI to do much of their coding; big banks are hiring fewer workers as they lean more on AI; even Amazon is looking at how AI-enabled robots could automate 75% of its warehouse operations, the New York Times reported Tuesday, citing leaked internal documents. (Amazon says the documents paint a misleading picture of its plans.)

“This is the dream, but if something goes wrong and you don’t have that human intelligence that’s up to speed,” Bourne said, “then we’re really offloading all of these critical tasks to AI and putting a lot of trust in the technology.”

But that threat isn’t inevitable: The shift to AI presents an opportunity to build more resilient internet architecture. Smaller cloud computing competitors like Oracle and CoreWeave are gaining market share with AI-specific offerings. And companies are increasingly relying on multiple cloud providers to create a backstop if one goes down.

Major large language model makers, including Meta and OpenAI, are also investing billions to build their own data centers, which could reduce the strain on shared systems. The tech industry is also pushing to make some AI models smaller and more power-efficient, so they can run locally on smartphones and laptops rather than relying on the cloud so much.

And AI could help find and fix security flaws to prevent outages like Monday’s — if companies invest in those capabilities as much as buzzy, popular tools like AI chatbots and video generation apps.

“There is a pathway to make AI serve us in the best possible ways,” Bourne said. “It doesn’t necessarily seem like we’re on that pathway, though.”

Monday’s Amazon Web Services outage — and the global disruption it caused — underscored just how reliant the internet has become on a small number of core infrastructure providers.

The ramifications of such outages could only get worse if artificial intelligence becomes as central to work and daily life as tech giants suggest it will in the coming years.

Monday’s outage briefly blocked some people from scheduling doctor’s appointments and accessing banking services. But what if an outage took down the AI tools that doctors were using to help diagnose patients, or that companies used to help facilitate financial transactions?

It may be a hypothetical scenario today, but the tech industry is promising a rapid shift toward AI “agents” doing more work on behalf of humans in the near future – and that could make businesses, schools, hospitals and financial institutions even more reliant on cloud-based services. A global survey of nearly 1,500 firms published by McKinsey & Company in May found that 78% of respondents already use AI in at least one business function, up 55% from a year earlier.

“If there’s an outage and you rely on AI to make your decisions and you can’t access it, that’s going to have an effect on performance,” said Tim DeStefano, associate research professor at Georgetown’s McDonough School of Business.

Monday’s outage had such a widespread impact because many companies rely on cloud providers for the backend functions that support their businesses, such as virtual server space, storage or developer tools. Typically, this set up is more affordable, flexible and secure for those customers, except when AWS experiences an outage. Then it’s effectively a single point of failure for a huge swath of the internet.

To be fair, these services are remarkably sturdy considering the scale of their operations. But outages like Monday’s raise questions abouthow crucial tech services could be made even more reliable.

AWS serves millions of customers, from retailers and restaurants to financial services firms and government agencies; it holds around 37% of the cloud computing market, according to Gartner. Together with Microsoft and Google, the three companies service around 70% of the market.

And the consolidation of the internet’s backbone is continuing in the age of AI. While there’s some grappling between the big three, Amazon, Microsoft and Google remain by far the prominent cloud computing providers for AI applications, according toEmarketer senior analyst Jacob Bourne — and their futures depend at least in part on serving AI demand.

While websites and apps can still technically function using their companies’ own less powerful on-premises servers, “cloud computing represents a technological prerequisite for using AI,” DeStefano said. That’s because the computers needed to run AI tools are powerful and expensive, and on-site hardware isn’t as easy to modify as business needs change. It just makes more sense to rent that computer space and pay for it only as needed.

And as AI becomes more widespread, data center outages could happen more frequently since AI models are so power-hungry, Bourne said. Major cloud providers, including Amazon, Microsoft and Google, are spending billions on data centers to address this growing need.

The risk of serious disruption from an outage rises considerably the more companies rely on AI agents to do critical tasks and automate the work of humans, a transition that’s already in progress despite disagreement about just how far it will go.

Tech companies are relying on AI to do much of their coding; big banks are hiring fewer workers as they lean more on AI; even Amazon is looking at how AI-enabled robots could automate 75% of its warehouse operations, the New York Times reported Tuesday, citing leaked internal documents. (Amazon says the documents paint a misleading picture of its plans.)

“This is the dream, but if something goes wrong and you don’t have that human intelligence that’s up to speed,” Bourne said, “then we’re really offloading all of these critical tasks to AI and putting a lot of trust in the technology.”

But that threat isn’t inevitable: The shift to AI presents an opportunity to build more resilient internet architecture. Smaller cloud computing competitors like Oracle and CoreWeave are gaining market share with AI-specific offerings. And companies are increasingly relying on multiple cloud providers to create a backstop if one goes down.

Major large language model makers, including Meta and OpenAI, are also investing billions to build their own data centers, which could reduce the strain on shared systems. The tech industry is also pushing to make some AI models smaller and more power-efficient, so they can run locally on smartphones and laptops rather than relying on the cloud so much.

And AI could help find and fix security flaws to prevent outages like Monday’s — if companies invest in those capabilities as much as buzzy, popular tools like AI chatbots and video generation apps.

“There is a pathway to make AI serve us in the best possible ways,” Bourne said. “It doesn’t necessarily seem like we’re on that pathway, though.”