Dive Brief:
- Amazon President and CEO Andy Jassy sees generative AI as possibly the "largest technology transformation since the cloud" and the company’s next pillar of growth, he said in the company’s annual shareholder letter Thursday.
- “While we’re building a substantial number of generative AI applications ourselves, the vast majority will ultimately be built by other companies,” Jassy said. In line with AWS’ efforts to optimize customers’ cloud bills last year, the cloud giant is looking to help ease costs as generative AI workloads grow.
- Chips that support compute-intensive workloads are still a key part of its strategy, too. “Customers have asked us to push the envelope on price-performance for AI chips, just as we have with Graviton for generalized CPU chips,” Jassy said.
Dive Insight:
To meet the moment, Amazon is investing heavily in the technology to help customers, no matter the AWS adoption route, from turnkey tools to heavier lifts, Jassy said.
“Much of the early public attention has focused on GenAI applications, with the remarkable 2022 launch of ChatGPT,” Jassy said. “But … there are three distinct layers in the GenAI stack, each of which is gigantic, and each of which we’re deeply investing.”
The three layers are separated by how customers approach adoption, such as enterprises opting to build foundational models; those looking to customize an existing model with internal data on Bedrock; and customers seeking generative AI-powered tools.
“What we’re building in AWS is not just a compelling app or foundation model,” Jassy said. AWS services "will empower internal and external builders to transform virtually every customer experience that we know.”
To get customers on board, AWS is trying to clear the path to entry, from costs to chip supply issues.
While AWS customers began to show signs of increased spending and lengthier deals at the end of 2023, Jassy said, cost optimization is still a priority.
AWS drove down prices for general cloud computing with specialized chips and is now doing the same for generative AI with Trainium and Inferentia to differentiate itself amid an AI arms race amongst providers.
The company has also targeted cost reduction by relaxing discount pricing restrictions and waiving egress fees.
