Amazon’s cloud computing division, AWS, is shifting its focus to large-scale language models (LLM) and generative AI-based offerings as its overall revenue growth continues its downward spiral.
Amazon Web Services (AWS) posted 16% year-over-year growth in the first quarter of fiscal 2023 on revenue of $21.4 billion. However, this revenue growth is slower than the 20%, 27.5% and 33% growth seen in Q4, Q3 and Q2 respectively.
The slowdown in growth could be attributed to companies optimizing their cloud spending due to uncertain macroeconomic conditions, according to company management.
“Given continued economic uncertainty, customers of all sizes in all industries continue to seek cost savings across their businesses, much like we do at Amazon. Customers continue to look at ways to optimize their cloud spending in response to the challenging economic conditions of the first quarter,” said Brian Olsavsky, AWS Chief Financial Officer, on the earnings call.
“These optimizations will continue into the second quarter, with revenue growth in April approximately 500 basis points lower than in the first quarter,” added Olsavsky.
Last quarter, the company warned that the trend for companies to optimize cloud spending will continue to be a headwind over the next two quarters.
In response to this trend, Olsavsky said AWS sales and support hours continue to spend a lot of time optimizing spending to help customers “better navigate this uncertain economy.” said.
However, AWS’ top executives remained bullish about the sector’s growth prospects, citing opportunities around transforming on-premises workloads.
“The new customer pipeline looks strong. The ongoing series of migrations of workloads to AWS is strong. Product innovation and delivery are rapid and compelling. It’s easy to forget that it’s still on-premises,” Amazon CEO Andy Jassy said on a conference call.
AWS shifts focus to generative AI
Additionally, the CEO hinted that much of the company’s cloud business stems from machine learning requirements.
“And in my opinion, few people realize how much new cloud business will emerge from the machine learning deluge over the next few years,” said Jassy. .
The company has already adjusted its capital expenditures, directing funds toward improving its large-scale language models and generative AI capabilities.
Olsavsky said Amazon has been cutting spending in its fulfillment and transportation divisions year-over-year, and has decided to pass the savings on to AWS to invest in infrastructure and large language models.
According to Jassy, AWS’ strategy is to target revenue generation from generative AI or large-scale language models by providing computing resources, training capabilities, and applications.
“I would say there are three macro areas in this space. Now thinking about the probably bottom layer is that all large language models run on compute. And the key to that compute is will be the chips that power that computing,” Jassy said, adding that the company has already launched Tranium chips and accelerators for memory-intensive tasks, ideal for AI-intensive workloads. added.
According to the CEO, the second tier is the training of the underlying models, and AWS offers multiple AI applications designed to allow companies to customize and create their own generative AI applications, including programs for general commercial use. We have just launched the Amazon Bedrock service, which provides the underlying model for .
Jassy cites Amazon CodeWhisperer to provide a third macro area or layer that delivers applications for developers, such as ChatGPT-enabled Copilot on Microsoft-owned GitHub.
“Every business within Amazon is building on a language model at scale to reinvent the customer experience. added the CEO. .
Other investments by the Cloud Computing division in the first quarter included a new region in Malaysia and a second region in Australia.
