Amazon Web Services (AWS) is making strides in artificial intelligence (AI) with investments in AI chips and software that save customers money while speeding up the migration of their legacy Windows and VMware workloads to the cloud.
At the Re:Invent conference in Las Vegas, AWS CEO Matt Garman highlighted the transformative role of generative AI (GenAI) for businesses. He discussed how AWS is ramping up its cloud infrastructure to meet the escalating demand for the computing power required for AI projects. Recently, AWS reported a 19.1% increase in annual revenue, largely attributed to the surge in demand for GenAI workloads.
Garman emphasized AWS’s 14-year partnership with Nvidia, which has enabled AWS to introduce more powerful graphics processing unit (GPU) instances to keep pace with customer needs. Additionally, AWS is focusing on its own AI silicon, specifically its Tranium chip family, to enhance performance and reduce costs for compute-intensive tasks. Garman announced the launch of the second generation of Tranium instances, claiming that they can deliver 30-40% better price performance compared to current GPU-powered instances. Early adopters like Adobe and the AI startup Poolside have experienced promising outcomes with this technology, with Poolside forecasting a 40% cost saving.
The discussion also revolved around how GenAI is improving AWS operations. Amazon Q, the company’s generative AI chatbot, assists in refactoring legacy on-premise workloads for cloud migration. With many enterprises still reliant on on-premise systems, Garman noted that Amazon Q is a game-changer. He explained how Amazon Q Developer, tailored for developers, speeds up the CodeDeploy process, allowing for faster, more secure software updates.
Garman highlighted the chatbot’s ability to automate Java version upgrades, transforming applications in a fraction of the usual time. This capability saved AWS an astonishing 4,500 developer years and significantly reduced operating costs. He also addressed the common challenge of migrating from Windows to Linux, expressing confidence that the updated Q Developer makes this process easier and quicker.
He referenced Signature IT, which managed to move its legacy .Net applications from Windows to Linux in just days—a project originally estimated to take months. Garman pointed out that VMware workloads also pose modernisation challenges. With Q Developer, AWS aims to simplify the migration process by identifying application dependencies and creating migration plans.
Looking ahead, AWS is also targeting mainframe migrations, notoriously complex and time-consuming. Garman shared that Amazon Q can help analyze and document mainframe code, potentially slashing the time required for these projects from several years to just a few quarters. Customers are showing excitement at the prospect of accelerating their modernization efforts with Amazon Q.