Amazon CEO Highlights Chip Momentum and Satellite Ambitions in 2026 Letter
Amazon CEO Andy Jassy outlined the company’s growing focus on artificial intelligence chips, broadband satellites, and large-scale infrastructure spending in his 2026 shareholder letter, published on Amazon’s website.
Jassy said the company’s custom silicon business, which includes its Trainium and Graviton chips, has reached an annual revenue run rate exceeding $20 billion. Demand for the newest AI chip, Trainium3, is nearly sold out, while future capacity for Trainium4 is already heavily booked despite being 18 months from launch. He emphasized that these chips offer improved price-performance compared to competitors such as Nvidia and Intel.
The letter also highlighted the progress of Amazon Leo, the company’s low Earth orbit satellite network set to begin service in mid-2026. Jassy said Leo has secured contracts with Delta Air Lines, AT&T, Vodafone, Australia’s National Broadband Network, and NASA. The system aims to provide lower-cost, higher-performance connectivity than existing services and will integrate directly with AWS for enterprise and government use.
Jassy confirmed that Amazon plans to invest about $200 billion in capital expenditures during 2026, primarily to expand AWS data centers. He said the spending is backed by customer commitments, including a $100 billion deal with OpenAI and several additional agreements in progress.
He also mentioned potential future ventures in robotics, noting that data from over one million warehouse robots could form the basis for new commercial robotics solutions.
We hope you enjoyed this article
Consider subscribing to one of our newsletters like Silicon Brief or Daily AI Brief.
Also, consider following us on social media:
More from Data Centers
Sep 29 Efficient Computer Raises $97 Million to Scale Its Processors Sep 29 Hikvision Adds HIKO AI Engine to Hik-Connect 7 Sep 29 Compal to Show NVIDIA AI Factory Infrastructure at OCP Summit Sep 29 Samsung Invests $1 Billion in KKR's Helix Data Center Platform Sep 29 Cerebras to Supply 100 Megawatts of AI Systems to Gimlet LabsSilicon Brief
Weekly coverage of AI hardware developments including chips, GPUs, cloud platforms, and data center technology.
Whitepaper
Tensordyne Napier: What If One Rack Could Do the Work of Nine?
Tensordyne
This Tensordyne whitepaper presents Napier, an inference-focused AI processor and rack-scale system based on the company’s TDN Math logarithmic number system. It examines infrastructure requirements for large mixture-of-experts and agentic models, compares major inference architecture approaches, and details the TDN AIP processor, TDN72 pod, TDN Link fabric, and Napier Ultra configuration. The paper reports simulation-based performance, cost, and accuracy-validation results, including Tensordyne’s projected comparison of one Napier rack with a nine-rack Nvidia Rubin plus Groq deployment; the chip is reported as taped out and in fabrication.
Read moreYou may also like
Anthropic signs $517 billion in compute agreements over 11 months
Thread AI Finds 5% of Enterprise AI Buyers Commit to Ship Dates
New Constructs Warns AI Compute Spending Is Not Profit
Qualcomm and Amazon to Develop Custom Silicon for AI Data Centers
Cerebras to Supply 100 Megawatts of AI Systems to Gimlet Labs
Daily AI Brief: the AI news that matters, in your inbox.