Amazon Nova AI models now available in AWS Sydney Region

Amazon has announced the availability of its Amazon Nova foundation models in the AWS Asia Pacific (Sydney) Region.

AWS customers in this region can now utilise Amazon Nova models using cross-region inference profiles, aimed at reducing latency and bringing cost efficiency. This new option supports routing requests primarily in the source region without incurring additional costs, thereby charging based solely on the source region of requests. The models are also available in other regions, including Europe and other parts of Asia Pacific, such as Tokyo and Seoul.

The Amazon Nova range includes several models, each catering to different needs. Amazon Nova Micro offers low latency responses at minimal cost for text-only models, while Amazon Nova Lite supports multimodal capabilities for text, image and video inputs to text outputs at low cost and speed. Amazon Nova Pro, the most comprehensive of the three, provides a balanced approach in terms of accuracy, speed, and cost for various tasks. These models support over 200 languages and offer features for easy integration with proprietary databases through Amazon Bedrock, which include tools like Amazon Bedrock Knowledge Bases.

Rohit Prasad, Senior Vice President of Amazon Artificial General Intelligence, explained, "Inside Amazon, we have about 1,000 generative AI applications in motion, and we've had a bird's-eye view of what application builders are still grappling with. Our new Amazon Nova models are intended to help with these challenges for internal and external builders, and provide compelling intelligence and content generation while also delivering meaningful progress on latency, cost-effectiveness, customisation, Retrieval Augmented Generation (RAG), and agentic capabilities."

AWS partners and customers in Australia are already engaging with Amazon Nova to enhance their operations. Jesse Clark, Chief Technology Officer and co-founder of Marqo, commented, "We look forward to using Amazon Nova models, which we anticipate can deliver a 50%+ reduction in costs through batch inference processing. Based on our initial testing with over a million product images and catalog data, Amazon Nova has demonstrated potential to reduce our model training time from weeks to days. Its multimodal capabilities and state-of-the-art performance improve product metadata coverage by 120% while being 50%+ more accurate compared to our previous methods. Amazon Nova's advanced text and image processing capabilities and the support from our AWS account team, ultimately enables us to enhance our platform and offer even more sophisticated and personalized search experiences to our retail and e-commerce customers."

Baradhazhvar Balaji, Chief Executive Officer of Crystal Delta, also shared his experience, "Amazon Nova has been instrumental in our global launch of SkillsMax.Ai, an assessment platform used with more than 7,000 AI-driven interview assessments. Amazon Nova's advanced inference capabilities provide real-time, precise insights into candidate responses, enhancing the quality and efficiency of our automated assessments. By leveraging AWS's robust and scalable AI infrastructure, including today's launch of the models in the AWS Asia Pacific Region, we have accelerated innovation, optimised operational efficiency, and delivered concept to production in a quick turnaround. We've achieved a 40% increase in productivity and efficiency, as well as reduced AI-related costs by 90% compared to the previous models we explored. Additionally, Amazon Bedrock's ability to seamlessly integrate with different AI models allows us to customise our solutions to meet the diverse needs of our customers."

Stefan Buchman, Head of Solutions at Slalom Australia, noted the potential impact on local customers, "The availability of Amazon Nova models through cross region inference in the AWS Sydney Region marks a perfect intersection of global innovation and local delivery capabilities. We've been testing Amazon Nova as part of our Slalom GenAI Testbed for AWS, intended to help customers who want to quickly, efficiently, and safely experiment with GenAI. As a global consulting firm focused on strategy, technology, and business transformation, we're particularly excited about Amazon Nova's multimodal capabilities and its potential to help our local clients reimagine their business and technology processes. This combination of world-class AI technology with our local market expertise positions us to better serve Australian organisations as they accelerate their AI transformation journeys."

Adam Durbin, Chief Technology Officer at Mantel Group, emphasised the importance of cost and efficiency improvements, "We are excited about the ability to now leverage the new Amazon Nova models through cross region inference in the AWS Sydney Region. We're most looking forward to Amazon Nova Pro's speed and cost-efficiency. It would mean we can develop and deploy AI solutions that are both faster and more affordable for our clients. Its strength in multimodal understanding, instruction following, and agentic workflows will enable us to build sophisticated AI agents for complex tasks. Additionally, the availability of Amazon Nova models enhances our ability to provide more innovative solutions to Australian and New Zealand businesses to suit their specific needs, helping them achieve their business goals faster."

AWS has also expressed their commitment to supporting local businesses in their AI endeavours, as highlighted by Rianne Van Veldhuizen, Managing Director, ANZ at AWS, "There is tremendous appetite for ambitious thinking and creativity across Australia and New Zealand, and we're seeing exciting examples of generative AI being used to drive productivity and innovation across organisations. Today, businesses are evolving from 'what is possible' to asking 'how to deliver real business outcomes and business value' using AI and more recently agentic AI. We're making significant investments in local infrastructure and AI to support these aspirations, spending more than $9 billion dollars in the last decade, and investing a further $13.2 billion dollars by 2027 in local infrastructure. We're also investing to make the latest models available on Amazon Bedrock in the Sydney Region, such as Claude 3.5 and our own Amazon Nova models."