Amazon just unveiled Nova, a new family of AI models that puts the tech giant squarely in the race to build frontier AI systems. And the timing is no coincidence—it comes right as founder Jeff Bezos returns to help guide the company's AI strategy.
But in an increasingly crowded field of tech giants building their own AI models, what makes Nova significant?
I got the scoop from Marketing AI Institute founder and CEO Paul Roetzer on Episode 126 of The Artificial Intelligence Show.
Inside the Nova Family of Models
Announced at Amazon's re:Invent conference, the Nova suite includes:
- Four text-generating models (Micro, Light, Pro, and Premiere)
- Nova Canvas for image generation
- Nova Reel for video creation
The models boast impressive capabilities.
Even the smallest model, Micro, can handle up to 100,000 words, while larger models can process up to 225,000 words or 30 minutes of footage. Amazon plans to expand this to over 2 million tokens for some models in early 2025.
On the media generation front, Nova Canvas creates and edits images with control over color schemes and layouts, while Nova Reel can generate videos up to six seconds long (with two-minute capabilities promised soon). Both include built-in safeguards including watermarking and content moderation systems.
Looking ahead, AWS has announced plans for two more models: a speech-to-speech model in Q1 2025 and an "any-to-any" model in mid-2025 that promises to handle multiple types of input and output.
The Big Picture Gets Complicated
Amazon's move comes at an interesting time. The company recently invested billions in Anthropic while simultaneously working to reduce its reliance on external AI providers by building in-house capabilities.
"This whole space is confusing right now," says Roetzer. "Everybody's building their own models and doing deals with other people who are building models.”
There also might be more to this strategy than meets the eye—at least for Amazon.
“I would imagine they’re building a lot of these models because they see the opportunity to transform Amazon internally,” he says.
“My guess is that they’re largely focusing on internal applications and then they’re, as a by product, able to also build and open up these models and maybe drive different uses of AWS.”
And a Crowded Field Gets More Crowded
With Nova's launch, the field of companies building frontier AI models continues to expand.
OpenAI, Anthropic, Google, Microsoft, xAI, Meta, Nvidia, and Amazon are all building their own frontier models. You can also throw in Mistral and Cohere as model players, too.
What does that mean?
Likely, it means buckle up for some serious AI acceleration.
“You have the biggest companies in the world who have tens of billions of dollars of R&D money every year to build these massive frontier models,” says Roetzer.
As this acceleration plays out, Roetzer is paying close attention to how the frontier model companies themselves are hiring in order to gauge the impact of increasingly smarter models on the economy at large.
"I keep trying to pay attention to: Are they still hiring?,” says Roetzer.
“Are these companies still hiring marketing people and sales people and HR people? The way we're going to know when we're starting to see the impact on the economy is when the hiring practices of the frontier model companies change because the tech they've built has enabled them to change the way they hire and promote and retain workers."
Mike Kaput
As Chief Content Officer, Mike Kaput uses content marketing, marketing strategy, and marketing technology to grow and scale traffic, leads, and revenue for Marketing AI Institute. Mike is the co-author of Marketing Artificial Intelligence: AI, Marketing and the Future of Business (Matt Holt Books, 2022). See Mike's full bio.