Amazon Launches ‘Nova’ AI to Compete with Rivals
LAS VEGAS, Dec 3 (Reuters) – Amazon (AMZN.O) has entered the AI arms race with its newly announced “Nova” models. Unveiled at the company’s annual AWS conference, these new platforms will offer text, image, and video generation capabilities. The move directly challenges competitors like Adobe (ADBE.O) and Meta (META.O), who are all vying to provide customers with increased automation capabilities.
Amazon CEO Andy Jassy highlighted developer needs when announcing the new “Nova” models on Tuesday. He stated developers were looking for improved latency, cost reductions, and fine-tuning capabilities, driving the development of these new services.
Amazon Aims to Catch Up in the AI Race
The move represents Amazon’s most significant response to concerns that it was falling behind in the AI revolution. Rohit Prasad, Amazon’s head of artificial general intelligence, indicated the company would compete on both price and functionality, emphasizing the faster speeds offered by the new models. “If I have something better to offer, then customers will come and use it,” Prasad said in an interview.
New AI Features
The market for video generation from a single image or text prompt has become particularly competitive, with Adobe (ADBE.O), Meta (META.O), OpenAI, and TikTok parent ByteDance all unveiling similar AI programs. Amazon announced its Nova Reel software which helps users make six-second videos, which can be utilized to show products on the Amazon website. Videos up to two minutes will be available in the coming months, Amazon reported.
Amazon also announced Canvas, a method for generating images based on short text prompts. Jassy stressed that Amazon will incorporate watermarking to ensure responsible software usage and prevent the spread of harmful content.
Amazon plans to introduce an AI model soon that will accept text, images, speech, and video and generate any of these formats.
Further, Jassy announced a revamped, AI-powered version of its Alexa speech assistant in the coming months. Known internally as Banyan, this project has experienced delays due to concerns about the accuracy and speed of its responses, according to Reuters.
Prasad, who previously oversaw Banyan, expressed overall optimism for voice assistants’ future while declining to provide more details.
Reporting by Greg Bensinger.
Editing by Will Dunham and Jonathan Oatis
