The artificial intelligence (AI) revolution kicked into a higher gear this year. In January, Microsoft added $10 billion to its $3 billion stake in ChatGPT developer OpenAI, and that kicked off an arms race among tech giants to embed AI technologies into their businesses.
Google parent Alphabet raced to release its Bard chatbot so as not to fall behind its rival, and it's also rapidly building out its portfolio of cloud-based AI tools for businesses.
But Amazon (AMZN 0.87%) -- which operates the world's largest cloud platform -- initially seemed to be missing from the AI conversation. Below, I'll explain how the company is making up ground at a lightning-fast pace, with a little help from what could end up being a $4 billion September bet on AI start-up Anthropic.
NASDAQ: AMZN
Key Data Points
Amazon is tackling three core areas of AI
End users like consumers and businesses are mostly familiar with relatively polished AI platforms like ChatGPT, but they don't typically see the technologies underneath those applications. It's important for investors to understand them because Amazon CEO Andy Jassy says they have just as much financial potential as those finished products.
Most organizations will access AI products and services in the cloud, because it's where they already store their valuable data, applications, and digital assets. Amazon Web Services (AWS) is Amazon's cloud platform, and Jassy says there are three layers of AI it wants to conquer.
First, like most cloud providers, AWS is partnered with Nvidia to power its data centers because that company produces the most advanced AI chips in the world. But Amazon knows demand for AI hardware is skyrocketing, and it has developed its own chips called Trainium and Inferentia, which are specifically designed to accelerate the development of AI models.
Nvidia CEO Jensen Huang says there is $1 trillion worth of existing data center infrastructure that needs upgrading to support accelerated computing and AI, so hardware is a huge opportunity for Amazon.
Second, AWS continues to expand the number of large language models (LLMs) on its platform. Developing LLMs from scratch can take monumental amounts of data and substantial financial resources, so Jassy says business customers on AWS prefer to adopt third-party models and build on top of them instead. AWS offers its own model called Titan, which was built in-house, in addition to models provided by leading AI developers like Anthropic, which I'll talk about more in a moment.
Finally, AWS has developed customer-facing applications like CodeWhisperer that can help developers accelerate software programming. These tools fall in the same category as ChatGPT.
Amazon's $4 billion Anthropic bet will help AWS across all three layers
Anthropic is a start-up founded by seven ex-OpenAI employees, and it describes itself as an AI research and products company with a focus on safety. Many tech experts, including Tesla CEO Elon Musk, have expressed concerns that AI could be used for harmful purposes -- or that developers could lose control of the technology entirely.
While Anthropic is building advanced generative AI models just like OpenAI, the start-up echoes similar concerns on its website, and it admits developers don't fully understand how to train AI models to behave well. Therefore, the company is committed to only training AI models up to a threshold that remains within the bounds of its safety standards.
This approach is the reason tech giants like Zoom Video Communications, Google, Salesforce, and now, Amazon, have invested in Anthropic. Amazon's plan to invest up to $4 billion in Anthropic is the largest so far, and it comes with some significant strategic benefits for AWS.
Anthropic will use AWS as its primary cloud provider, and it will train its future AI models on Trainium and Inferentia chips. This is a big vote of confidence from one of the industry's leading AI start-ups, and it could lead to more customers perceiving Amazon's chips as viable alternatives to those designed by Nvidia.
Anthropic has also agreed to make its models available on AWS so millions of business customers will be able to reap the benefits of AI. This could be a major drawcard for AWS because, over the long term, organizations will likely decide which cloud providers to use based on the depth of their AI offerings.
Image source: Getty Images.
What the Anthropic deal could mean for Amazon in the short and long term
Since AWS is already the world's leading cloud provider, the Anthropic deal might not move the needle straight away. However, AWS delivered a modest 12% year-over-year revenue growth in the recent second quarter of 2023 (ended June 30), which was far slower than competitors like Microsoft Azure, which grew by 26% year over year, and Google Cloud, which grew by 28%.
Therefore, it's crucial for AWS to move to the front of the pack as a provider of AI products and services, as it's clearly losing market share in the cloud right now.
Microsoft's deal with OpenAI might offer clues as to how impactful the Anthropic deal will be. Since launching Azure OpenAI Service in January, which gives businesses access to the ChatGPT creator's latest AI technology, the segment has acquired 11,000 business customers. They include global furniture retailer Ikea and automotive giant Mercedes-Benz.
The amount of money those customers are spending remains unknown, but it's clear some of the largest organizations in the world are seeking the latest AI technology to help boost productivity and serve their customers more efficiently.
Since AWS has developed its own hardware, LLMs, and AI applications in-house, it offers customers a unique set of services that will always be somewhat different from its competitors.
Therefore, Amazon might stand to make a significant amount of money from its (up to) $4 billion bet on Anthropic, but the pedigree that the start-up brings to the portfolio of AWS AI services will likely be worth far more. It could help attract scores of customers with the potential to contribute to the cloud platform's revenue for the long term.
After seemingly falling behind in the AI race earlier this year, Amazon is certainly making up ground. But more importantly, it has a crystal-clear strategy for the future, and with its stock trading 32% below its all-time high amid the broader market sell-off, this could be an ideal entry point for investors.
