Amazon (AMZN) is going to attempt to use its Trainium and Inferentia processors to develop its artificial intelligence models in an effort to cut costs, The Wall Street Journal reported.
“If we can build our models on our chips, we can build them at a fraction of the cost of a pure-play AI model provider,” Amazon’s new artificial intelligence czar, Peter DeSantis, told the news outlet in an interview.
“AI has a cost problem,” DeSantia added. “If we ultimately want AI to transform everything, the costs have to be different.”
Nvidia (NVDA) is widely believed to have the vast majority of the market share for AI processors. The Jensen Huang-led company just reported fiscal fourth-quarter results and guidance that both topped estimates, but shares fell 5% on Thursday.
Earlier this month, Amazon said it would spend $200B in capital spending in 2026, far higher than what Wall Street was expecting. Part of that outlay will be on chips, Amazon CEO Andy Jassy said.
“With such strong demand for our existing offerings and seminal opportunities like AI, chips, robotics, and low earth orbit satellites, we expect to invest about $200 billion in capital expenditures across Amazon in 2026 and anticipate strong long-term return on invested capital,” Jassy said.
Amazon shares fell fractionally in premarket trading on Friday.