Adam Selipsky, CEO of Amazon Web Services (AWS), stated, “Artificial intelligence (AI) is almost certain to reinvent nearly every application used by hundreds of thousands of our customers.”
AWS is a leading provider of cloud services, holding a 32% market share, ahead of Microsoft at 22% and Google at 11%.
Currently, AWS serves 100,000 customers using artificial intelligence and is gearing up for a world where most of its customer base does the same.
AWS aims to provide its customers with a wide range of options for synthetic AI applications, whether they are Amazon’s own offerings or from partners like Anthropic, Cohere, Stability AI, and others.
The company builds its dedicated chips, trains its models, and develops software to train other AI models while offering straightforward front-end interfaces for developers to create code instructions to harness AI services, selipsky believes this comprehensive approach is key to making synthetic AI productive.
AWS relies heavily on its Trainium machine learning chips to train AI models, which are seen as an alternative to the market-leading graphics processing units from NVIDIA.
Although Trainium chips are an additional option rather than a replacement for GPUs, AWS hopes they will eventually outperform them in terms of both performance and price.
AWS’s Titan foundational models are offered, but the company expects most customers to use other foundational models, prioritizing flexibility and choice, this approach allows AWS to democratize IT by providing its services to the largest possible customer base.
AWS also owns Bedrock, a service for building foundational models, along with applications like Code Whisperer, an AI-powered code instruction generator.
Despite the overall slowdown in hiring, AWS’s Synthetic AI division continues to actively recruit new employees.
Enthusiasm has caused some delays in adopting synthetic AI, according to AWS executives, including Selipsky.
Some customers have started with about 200 ideas or more on how to create synthetic AI features or new products, these ideas took months to narrow down to trial versions and adoption.
AWS executives are optimistic about enterprises’ ability to quickly adopt synthetic AI, some customers, such as Snap and Canva, have adopted AI features within weeks, the company expects broader adoption of synthetic AI in the coming months.
