Amazon.com Inc. (AMZN) rolled out its Graviton5 processors on Thursday, aiming to help cloud customers handle increasingly complex workloads without breaking the bank or the planet. The chip promises faster performance, lower energy consumption, and cost savings—a trifecta that matters when you're running applications at scale.
This latest release sharpens Amazon's competitive edge against traditional chipmakers Intel Corp (INTC) and Advanced Micro Devices (AMD). The battle isn't just about raw computing power anymore. Hyperscalers like Amazon, Microsoft Corp (MSFT), Alphabet Inc (GOOGL), and Meta Platforms Inc (META) have been designing their own custom chips to cut costs, solve compatibility headaches, and sidestep supply chain bottlenecks.
What Makes Graviton5 Different
According to Amazon, Graviton5 delivers up to 25% better performance than its predecessor while using less energy. That combination helps organizations meet both budget targets and sustainability commitments—increasingly important as data centers consume massive amounts of electricity.
The new Graviton5-powered EC2 M9g instances are engineered to handle large-scale tasks more smoothly by speeding up internal data movement. This matters for demanding applications like gaming platforms, big data analytics tools, and online services that need reliable, fast responses.
Amazon also upgraded how quickly the chip accesses critical information and improved its memory handling capabilities. These enhancements boost performance for applications working with large datasets or constantly changing information streams.
Network and storage speeds got a boost too, enabling faster data transfers, backups, and better support for systems requiring real-time data processing. All these improvements come while using energy more efficiently—not always an easy balance to strike.
Amazon says the performance gains stem from updating the chip's underlying architecture and tailoring it specifically for how Amazon Web Services customers actually use the cloud. It's custom silicon designed around real-world usage patterns, not generic computing.
On the security front, Amazon added a new verification layer using a more rigorous method to ensure systems behave as intended. The company plans to let customers review this security work directly—a transparency move that should appeal to enterprises with strict compliance requirements.
Amazon opened Graviton5-based M9g instances for preview, with additional versions designed for compute-intensive and memory-heavy workloads planned for 2026.
Wall Street Sees AWS Momentum Building
The $2.5 trillion tech giant has gained just 4% year-to-date, but Amazon recently won fresh bullish support from Wall Street. Analysts praised its agent-driven AI strategy and new custom chips unveiled at AWS re: Invent.
Bank of America's Justin Post highlighted Amazon's new autonomous AI agents, deeper collaboration with Nvidia, and the Trainium4 roadmap as factors strengthening AWS's long-term competitive position. He expects AWS revenue growth to approach 25% in 2026.
JP Morgan's Doug Anmuth argued that Amazon is closing the competitive gap through stronger Trainium performance, Bedrock momentum, and partnerships with Anthropic and OpenAI. He projected 23% AWS growth in both the fourth quarter and 2026, driven by Trainium 3, Trainium 4 development, and new AI Factories.
Wedbush's Scott Devitt noted that Trainium3, the Nova 2 models, and new autonomous agents position 2026 as a breakout year for Amazon. He pointed to rising demand, higher capital expenditures, and AWS's expanding capacity as catalysts for accelerating growth.
Amazon.com shares were down 1.58% at $228.72 at the time of publication on Thursday.