Bengaluru, Apr 23, (CXO Media): Amazon Web Services (AWS) used its Summit in Bengaluru to spotlight its latest AI and cloud processors as part of a broader push to strengthen its custom chip strategy in India.
AI Chip Strategy In Focus
AWS presented its Trainium3 AI chip and Graviton4 processor at the AWS Summit 2026, highlighting their role in supporting growing demand for artificial intelligence workloads. Trainium3, first introduced globally in December 2025, is built on 3-nanometre technology and designed to handle token-based AI tasks such as reasoning systems, agent-driven tools, and video generation.
The company also reiterated its focus on its Graviton processor line, which targets general cloud workloads. Graviton4, introduced in November 2023, forms part of AWS’s expanding silicon portfolio, alongside Graviton5, launched in December 2025 with 192 cores and increased cache capacity.
Growing Role Of Custom Silicon
During the keynote, AWS demonstrated several agent-based tools, including Kiro, AWS Security Agent, and DevOps-focused systems, reflecting increasing adoption of automated AI workflows. The company said its inference service, Amazon Bedrock, now runs a large share of workloads on Trainium chips, indicating a shift toward in-house silicon for AI infrastructure.
AWS reported that its custom chip business has crossed an annual revenue run rate of over $20 billion, supported by strong growth in enterprise demand for AI computing capacity.
India Investment And Expansion Plans
The announcements come as AWS marks two decades of operations in India and continues to scale infrastructure investments. Amazon has committed to investing more than $35 billion in India by 2030, in addition to around $40 billion already deployed across its businesses in the country.
The Bengaluru showcase signals AWS’s intent to deepen its role in India’s AI and cloud ecosystem, where demand for high-performance computing infrastructure is rising alongside enterprise adoption of generative AI tools.
