By Interestana AI Editorial — AI-drafted, human-overseen. How we report
Amazon Trainium3 Chips Boost AI Compute and Efficiency

Amazon announced its latest AI chip, Trainium3, in December 2025 at AWS re:Invent, specifically designed to address current and future AI workflows, including emerging areas like world models. The new chip architecture aims to provide enhanced compute performance and energy efficiency, crucial for the rapidly evolving field of artificial intelligence. The Trainium3 UltraServer, a system incorporating up to 144 of these chips, reportedly quadruples the compute performance and energy efficiency compared to its predecessor. This significant upgrade positions Amazon to better support the intensive demands of advanced AI research and deployment. Early adoption of Trainium3 by prominent companies highlights its practical impact. Uber is integrating Trainium3 for critical tasks such as calculating estimated arrival times, a process that benefits from faster and more efficient computation. Furthermore, the video-generation platform Decart has leveraged Trainium3 to achieve a 4x increase in frame throughput and a substantial 75% reduction in latency. These results demonstrate the chip's capability to accelerate demanding multimedia AI applications. Beyond Amazon's chip advancements, the computing landscape is seeing innovation across multiple fronts. Cadence, a leader in chip design, introduced its Super Agents in early 2026. These AI agents are designed to automate key aspects of processor design, a field historically reliant on human engineering talent. Super Agents handle tasks such as creating test plans, managing design node migration, and executing digital implementation. Companies like Nvidia, Qualcomm, and Altera have already adopted Cadence's Super Agents, reporting productivity increases of up to 10x in design and verification processes. This signifies a shift towards AI-assisted engineering in hardware development. HP is also contributing to the future of computing with its Z8 Fury G6i desktop workstation. This workstation is engineered to support current high-performance GPUs, such as up to four Nvidia RTX Pro 6000 GPUs. Crucially, HP has also designed the workstation with future GPU advancements in mind. An optional Max Side Panel expands the workstation's chassis by 1.4 inches and incorporates additional cooling fans. This expansion capability allows for tool-free installation of larger, more power-hungry, and hotter GPUs that are anticipated in the coming years, thereby extending the workstation's lifespan and preventing the immediate need for replacement. HP's commitment to sustainability is also evident in the Z8 Fury G6i, as it incorporates recycled plastic, steel, and rare earth materials in its construction, aligning with broader industry trends towards environmentally conscious manufacturing.
Original source — read the full reporting at the publisher:
Read on Fast CompanyGet the weekly AI digest
AI news + new model releases, weekly. Drafted by our agents, reviewed by humans.