Revolutionizing Application Modernization with Generative AI Foundation Models, Data Analytics Breakthroughs, Next-Gen Silicon and Amazon Q Index at AWS re:Invent 2024
Analyst(s): Mitch Ashley, Camberly Bates, Daniel Newman, Don Hinchcliffe, Nick Patience, and Keith Townsend
Publication Date: December 13, 2024
Among the plethora of product announcements at AWS re:Invent 2024, AWS announced the GA of Trainium2 while hinting about its next generation of silicon, launched new Nova AI foundation models and model distillation and training in Amazon Bedrock, generative AI to modernize applications, Amazon Q Index taps into enterprise data across apps, and increased data analytics workload performance with S3 enhancements.
Key Points:
- AWS’s Generative AI Advancements for Application Modernization: This article and our full report highlight AWS’s new generative AI features and tools (e.g., Amazon Q Developer) aimed at modernizing legacy VMware, .NET, and mainframe applications, making it easier to migrate, refactor, and optimize workloads in the AWS cloud.
- New Generative AI Models and Supporting Infrastructure: AWS introduced Amazon Nova (a new family of generative AI foundation models) and enhancements to Amazon Bedrock, including model distillation for cost-effective fine-tuning. These announcements are backed by improved in-house silicon (Trainium2 is now GA, and Trainium3 is coming) and expanded data management capabilities in S3.
- Broader AI Ecosystem and Competitive Landscape: Our analysis acknowledges the robust, growing AI market, in which alternatives to NVIDIA’s GPUs are emerging but unlikely to dethrone it. AWS’s moves—such as integrating AI-driven indexing (Amazon Q Index), refining AI-based developer tools, and evolving its infrastructure—demonstrate its intent to foster a diverse, competitive, and thriving AI ecosystem.
Insights from Our Expert Analysts
Which Company Will Take On NVIDIA?
Daniel Newman, The Futurum Group CEO
The question arises about who will challenge NVIDIA with each new AI chip. At AWS re:Invent, Amazon’s Trainum chips sparked this conversation. However, no one is truly displacing NVIDIA. Demand for AI is huge, encouraging alternatives and competition. NVIDIA remains strong, powered by its innovation pipeline and software stack.
Meanwhile, AWS, Google, Microsoft, AMD, Intel, and others will also grow, increasing diversity and customer choice. AWS’s silicon successes (Graviton, Nitro) suggest Trainum’s future impact. This isn’t a zero-sum scenario; as the market expands, multiple players, including NVIDIA, will thrive, benefiting the entire AI ecosystem. All parties stand to gain overall.
Trainium2 now GA, Trainium3 on the way
Nick Patience, VP & Practice Lead, AI
AWS’s Trainium chips gained the spotlight at re:Invent despite the company’s longstanding partnership with NVIDIA. Trainium2, now GA, claims 30-40% better price performance than current GPU-based EC2 instances, and customers are testing its potential. Apple, a Trainium1 user, expects a 50% efficiency boost in pre-training. Adobe tested it on Firefly models, Databricks sees up to 30% cost savings, Poolside plans to train models, and Qualcomm will handle training and inference at the edge. AWS and Anthropic’s Project Rainer will leverage thousands of Trainium2 chips. AWS revealed Trainium3, coming in 2025, with double the performance and 40% more efficiency.
Application Modernization
Keith Townsend, Chief Technology Advisor and President of the CTO Advisor
Mitch Ashley, VP and Practice Lead, DevOps and Application Development
AWS is shifting focus from multi-cloud to cloud migration. At re:Invent, it introduced Amazon Q Developer features to ease moving data and workloads into AWS. CapitalOne demonstrated how it’s operationalizing observability. With VMware halting new VMware on AWS Cloud accounts, AWS offers tools to reduce refactoring costs, enticing customers to adopt native AWS services. For .NET applications, Amazon Q Developer helps organizations move off Windows-based environments, cutting licensing costs. For mainframes, AWS uses AI to document COBOL code and streamline modernization. These changes highlight AWS’s commitment to simplifying migration and reducing barriers to embracing its cloud-native infrastructure and services.
Plus – in the full report (for Subscribers):
Amazon Nova and Amazon Bedrock Improvements
Nick Patience, VP & Practice Lead, AI
AI In Development
Mitch Ashley, VP and Practice Lead, DevOps and Application Development
AWS Expands S3 as the Data Lake
Camberley Bates, Chief Technology Advisor, Analyst and Data Infrastructure Lead
Amazon Q Index Can Help CIOs Unleash Generative AI Insights Across Enterprise Apps
Dion Hinchcliffe, VP, Practice Lead,CIO / CTO
AWS is going after on-premises workloads with the value proposition of a well-thought-out platform that integrates identity management, data center infrastructure primitives, data management, and AI. No single enterprise customer can build such an integrated platform for enterprise applications. However, enterprise IT inertia and data gravity factor into the decision to migrate at the scale of customers such as CapitalOne, AWS’s premier reference customer.
While these tools help with many tactical challenges enterprises face, migrating to the public cloud remains high in risk and effort. Customers are also questioning the ROI, given the trend of repatriating legacy cloud workloads to the private data center.
The full report is available via Futurum’s IQ service subscription – click here for inquiry and access.
Futurum clients can read more about it in the DevOps & Application Development Portal. Nonclients can learn more here: DevOps & Application Development Practice.
About the Futurum DevOps & Application Development Practice
The Futurum DevOps & Application Development Practice provides actionable, objective insights for market leaders and their teams so they can respond to emerging opportunities and innovate. Public access to our coverage can be seen here. Follow news and updates from the Futurum Practice on LinkedIn and X. Visit the Futurum Newsroom for more information and insights.
