The emergence of generative AI opens up a massive new market for large cloud providers, but it also represents a reset of sorts. Unlike the rise of containers, for example, generative AI is an entirely new market. Some players who have long lagged behind the overall cloud market – such as Google – are clearly betting on this to change their fortunes. But for AWS, it means it is now in a position where it potentially has to defend its leadership, something the company is not used to and something that became abundantly clear in AWS CEO Adam Selipsky’s keynote today: Invent.
AWS made a habit of mocking Oracle from time to time during its early keynotes. Today, during what was a relatively quiet affair by mainstream AWS standards, there were numerous mentions of Google and Microsoft (and its close relationship with OpenAI) — and it very much felt as if Selipsky was pushing the envelope and taking some cheap shots at the expense of his competitors. This is not something a company does when it feels like it is at the top of its game.
When talking about AWS’s data center footprint for example, Selipski noted – as he did in previous keynotes – that AWS ensures its availability zones are geographically distributed – not just three regions in the same data center. He then specifically explained that if AWS had a setup like this in France and a fire had broken out there, the region would have been down. At best, this was a veiled reference to Google An interruption of several weeks at its data center in Paris earlier this year. Naturally, AWS has also had its share of outages – just like any other cloud provider. Although, to be fair, Google’s outage was excessive.
Image credits: Frederic Lardinois/TechCrunch
He said: “Others may think that all clouds are the same, but this is not true.” “I mean, imagine if you had a region backed by a single data center, or if you thought your service provider had multiple Availability Zones in France, for example, and you thought they were resilient, but it turned out they were actually in the same location. I mean, A single incident such as a water leak followed by a fire can devastate an entire area for days.
Then, at the launch of the new Graviton chips, Selipski also emphasized: “We are now in our fourth generation in just five years. Other cloud providers haven’t even delivered their first server processors yet.
But it’s generative AI that makes Amazon the most vulnerable right now. For the first time in a long time, it does not have a leadership advantage in the cloud. Unsurprisingly, the recent OpenAI drama has provided plenty of material for the AWS CEO to take swipes at the competition, including news that Microsoft briefly suspended its employees from using OpenAI’s ChatGPT over security concerns.
“You wouldn’t actually use any of this in a serious way if it wasn’t secure and private,” Selipski said. “When I think about it, I can’t imagine other cloud providers trying to provide general AI services or customers using general AI applications. They’re not confident in their models and that their data is secure.”
While Microsoft Azure provides access to dozens of proprietary and open source models as well, Selipski tried to capitalize on the fact that the company is closely associated with OpenAI: “Things move very quickly — and in this kind of environment, the ability to adapt is the most valuable ability you can have.” You own them. There’s not going to be one model that rules them all. And there’s certainly not going to be one company that provides the models that everyone uses. So you don’t want a cloud provider that’s primarily beholden to one model provider.
![](https://techcrunch.com/wp-content/uploads/2023/11/IMG_3838.jpg)
Image credits: Frederic Lardinois/TechCrunch
During the keynote, Selipski brought Anthropic CEO Dario Amodei on stage to talk about the close partnership between the two companies and noted how AWS users will get exclusive early access to some features in Anthropic’s Claude model. AWS also plans to invest up to $4 billion in Anthropic, making this somewhat similar to the relationship between Microsoft and OpenAI. But I guess it depends on your point of view.
Speaking about CodeWhisperer, Amazon’s competitor to GitHub’s more successful Copilot, the AWS CEO noted that “other providers that have released tools have launched them without the data privacy and security capabilities that almost every organization requires.”
I think AWS competitors would disagree.
The problem is that AI is causing companies to rethink their cloud strategies. Microsoft’s close relationship with OpenAI gave it an early advantage. Google has long had a good reputation in the world of artificial intelligence and machine learning in general. With tools like SageMaker, Bedrock, and now Q Assistant, AWS is arguably on par, but it’s not in its element — and its rhetoric shows it.