LAS VEGAS — Amazon CEO Andy Jassy took the stage at AWS re:Invent on Tuesday morning to reveal one of the day’s most significant announcements: a new suite of AI foundation models named “Nova.”
These Nova models will be available alongside popular third-party models through the company’s Bedrock service. This move represents one of Amazon’s most substantial efforts to establish a presence in AI, aiming to address perceptions of initial delays in the tech giant’s involvement in the early stages of the generative AI wave.
Jassy, the former AWS CEO, noted that the company initially developed the models for its internal use before opting to release them to the public.
“You guys want a lot. There’s a lot to do,” Jassy said. “It’s one of the reasons why we have continued to work on our own frontier models, and those frontier models have made a tremendous amount of progress in the last four to five months. And we figured if we were finding value out of them, you would probably find value out of them.”
According to Amazon’s announcement, the Nova family introduces several primary models.
- Amazon Nova Micro, a text-only model delivering low-latency responses at a low cost.
- Amazon Nova Lite, a budget-friendly multimodal model for handling images, videos, and text.
- Amazon Nova Pro, a model balancing accuracy, speed, and cost for various tasks.
- Amazon Nova Premier, described by Amazon as “the most capable of Amazon’s multimodal models for complex reasoning tasks and for use as the best teacher for distilling custom models.” Premier is expected to be available in the first quarter of next year.
- Amazon Nova Canva for generating images.
- Amazon Nova Reel for generating videos.
Additionally, Jassy hinted at two more models under development for next year as part of the Nova initiative — a speech-to-speech model and an “any to any” model.
“This is really multimodal to multimodal,” Jassy said. “So you’ll be able to input text, speech, images, video and output text, speech, images and video. This is the future of how frontier models can be built and consumed.”
This initiative is part of Amazon’s broader strategy to equip AWS customers with new AI capabilities while using AI to reinvigorate its cloud business. Jassy conveyed that insights from internal developers indicate customers desire a wide array of AI models, a trend he expects to continue.
“We’re going to give you the broadest and best functionality you can find anywhere,” Jassy said, emphasizing that this promise “is going to mean choice.”
At re:Invent this week, the company also showcased advancements in its Trainium and Inferentia AI chips, its Amazon Q AI assistant, and its collaboration with Anthropic, the creator of Claude chatbot, among numerous other announcements.
