The Battle of Artificial Intelligence: Amazon is Catching Up with Microsoft and Google

Amazon's AWS re:Invent developer conference this week announced a series of announcements, indicating that Amazon is desperately catching up with the AI field's leader and main competitor, Microsoft.

re:Invent is Amazon's annual showcase for its important Amazon Web Services (AWS) business. This year, the company is focusing on generative AI, trying to compete with Microsoft and Google's early market advantage.

One of the biggest announcements at this conference includes Amazon's new Q chatbot, upgraded AI chips, and deeper collaboration with powerful chip manufacturer NVIDIA. These should help Amazon and its AWS business catch up with industry leaders.

Morgan Stanley analyst Doug Anmuth wrote in a research report, "Our discussions indicate that AWS has made solid progress in building its [generative] AI strategy and capabilities, with a focus on providing a wide range of customizable FMs in a secure environment." He added, "We are confident that AWS is starting to narrow some of the early gaps in generative AI, chip development, the breadth of [large language models], and the significant amount of customer data that has already been moved to the AWS cloud, all of which are key differentiators over time."

However, Amazon's moves do not mean that it is on par with Microsoft or Google, at least not yet.

Amazon's Upcoming Q

The hottest news from re:Invent is Amazon's Q chatbot. Q is designed for enterprise customers rather than the general public, aiming to serve as a general assistant to help users do things like drafting emails, proposing business ideas, summarizing reports, and planning workshops.

Amazon claims that Q is trained on 17 years of AWS data, which is a key advantage of the platform. Currently, the platform is in preview. In one example, Amazon explains how to use Q to transform information from corporate blogs into tagged social media posts. In another example, the company explains how to use Q to create useful applications for employees.

This service is similar to what Microsoft and Google currently offer on their generative AI platforms. Microsoft's Copilot and Azure AI Studio help employees summarize content, write emails, and create custom AI copilots. Copilots are essentially AI-driven applications.

Meanwhile, Google's Duet for Workspaces serves as a generative AI assistant for brainstorming and summarizing content, while its Vertex AI service helps enterprise customers build generative AI applications.

The key point here is that Amazon's Q puts the company on a more direct competition path with Microsoft and Google, the latter two ranking second and third in the cloud market.

Amazon's Chip Moves

In addition to Q, Amazon also unveiled its latest AI chip: Trainium2. Trainium2 is designed for training AI models, and Amazon says the chip will offer four times the training speed of the first-generation Trainium, with twice the energy efficiency.

Importantly, AI developer Anthropic, the company behind the Claude chatbot, says it will develop future base models on Amazon's Trainium2 chip. In September, Amazon announced a $4 billion investment in this startup.

Amazon has been developing and using its own AI chips for the past few years, which puts it ahead of Microsoft in custom AI chip development, as the latter only recently announced its research in this area. More importantly, despite investing more in its own silicon, Amazon is also deepening its relationship with AI chip leader NVIDIA.

During re:Invent, NVIDIA CEO Jensen Huang took the stage to praise Amazon's commitment to NVIDIA chips and announced that the company will receive NVIDIA's latest hardware.

Amazon's push for generative AI comes as the company continues to address the slowdown in AWS revenue growth, which has raised concerns on Wall Street. In the third quarter, the company reported $23.06 billion in net sales for AWS. Analysts had expected $23.13 billion.