Dive Brief:
- Amazon plans to keep quenching AI’s insatiable thirst for compute and capital in order to reap the business opportunities ahead, Amazon CEO Andy Jassy said in a Thursday letter to shareholders.
- “We continue to believe AI is a once-in-a-lifetime reinvention of everything we know,” Jassy said in the letter. “The demand is unlike anything we’ve seen before and our customers, shareholders, and business will be well-served by our investing aggressively now.”
- Citing skyrocketing demand for AWS services and the potential customer experience gains from AI deployment, Jassy said the company has invested to deliver AI fundamentals to the market. This effort includes the Amazon Nova models released in December and homegrown computing chips like Trainium2, which will help lower compute costs, Jassy said.
Dive Insight:
The leading hyperscalers — AWS, Microsoft and Google Cloud — have pledged to invest billions in capital expenditures to meet AI demand, citing infrastructure buildouts as a key driver of spend.
Amazon CFO Brian Olsavsky said the company plans to deploy $100 billion in capital expenditures this year alone.
“Similar to 2024, the majority of the spend will be to support the growing need for technology infrastructure,” said Olsavsky in February, during the Q4 2024 earnings call. “This primarily relates to AWS, including to support demand for our AI services as well as tech infrastructure to support our North America and international segments.”
In addition to standing up more compute capacity, Amazon has worked to deploy foundational models, expand access to specialized training and create its own line of AI-optimized chips. Amazon plans to unveil its latest chipset, Trainium3, later this year.
Chips are the biggest driver of costs for AI services, according to Jassy.
“Most AI to date has been built on one chip provider,” Jassy said. “It’s pricey. Trainium should help, as our new Trainium2 chips offer 30-40% better price-performance than the current GPU-powered compute instances generally available today.”
As providers increase their capital investments geared at boosting compute, governments have also laid out plans to expand domestic access to AI-ready infrastructure.
In January, the U.S. unveiled a $500 billion plan to grow the national data center footprint with support from the private sector. On Wednesday, the European Union unveiled its own action plan to increase compute capacity and specialized skills in the region.
