Amazon Web Services (AWS) has significantly expanded its artificial intelligence (AI) offerings with the launch of Nova 2, a new suite of four AI models, and Nova Forge, a service allowing enterprise customers to build custom AI versions. The announcement was made during AWS CEO Matt Garman’s keynote at the AWS re:Invent conference on Tuesday.
New Nova 2 Models: A Deep Dive
The Nova family, first introduced last year, now includes four new models designed for a range of applications:
- Nova 2 Lite : A cost-effective reasoning model for everyday tasks. These models “think” before responding, processing text, images, and videos to generate coherent outputs.
- Nova 2 Pro : A high-performance reasoning agent tailored for complex tasks, including coding, that can process text, images, videos and speech.
- Nova 2 Sonic : A new speech-to-speech model specifically designed for conversational AI applications.
- Nova 2 Omni : A multimodal model capable of processing and generating both text and images from a combination of inputs (images, text, video, and speech).
The introduction of Nova 2 builds on the rapid adoption of the initial Nova models, already used by tens of thousands of customers including major players like Infosys, Blue Origin, and Robinhood.
Nova Forge: AI Customization for Enterprises
AWS is also introducing Nova Forge, a service enabling enterprises to create proprietary AI models—dubbed “Novellas”—based on AWS’s existing Nova architecture. The service is priced at $100,000 per year and provides access to pre-trained, mid-trained, or post-trained models that companies can further refine with their own data.
This addresses a common challenge in AI: model drift. As enterprises customize models with additional data, the core reasoning capabilities of the original model can degrade over time. Garman likened this to the human learning process, where acquiring new skills becomes more difficult with age.
Companies like Reddit, Sony, and Booking.com are among the first customers to test Nova Forge.
Why This Matters
The expansion of AWS’s AI suite signals a broader trend: cloud providers are not just offering AI tools, but also the infrastructure for businesses to build bespoke AI solutions. This is critical because many enterprises need AI tailored to their specific data and use cases, rather than relying on generic models. The launch of Nova Forge positions AWS as a key player in this emerging market.
The service is an attempt to solve the problem of model drift, a phenomenon where the performance of AI models degrades over time when fine-tuned on proprietary data. This is a significant hurdle for enterprises looking to integrate AI into their existing workflows.
The release of Nova 2 and Nova Forge solidifies AWS’s position as a major competitor in the rapidly evolving AI landscape. The move towards customizable AI services is likely to drive further innovation in the industry.
