Summary Bullets:

• Amazon Web Services (AWS) has launched Amazon Bedrock in Australia, and it is helping customers develop generative AI (GenAI) applications using different foundation models to meet their specific needs.

• Australia is an attractive market for AWS since the company has a significant presence in the country, and it can support GenAI workloads sustainably as the country is seeing investments in clean energy and more efficient data centers.

AWS highlighted, at the recent AWS Summit in Sydney (Australia), that Australia was an important market for the company and that it was optimistic about the potential of GenAI adoption in the country. At the event, AWS also discussed its GenAI strategy. Rivals Microsoft and Google Cloud have been actively promoting their GenAI development through OpenAI and with Google Gemini, respectively, as well as building foundational models to help customers tap GenAI capabilities without building their own. While AWS has its models under the umbrella of Amazon Titan, it emphasizes giving customers their model of choice, including its own and partner models hosted within its cloud environment. This is through Amazon Bedrock, which is a fully managed service. It gives customers the ability to choose from a range of models based on their needs and preferences. They can develop GenAI capabilities and use cases without leaving the AWS environment. So far, AWS has agreements with several AI companies, natural language processing (NLP), and large language model (LLM) specialists, including AI21 Labs, Anthropic, Cohere, Meta, Mistral AI, and Stability AI. The advancement of LLM is happening by leaps and bounds with new startups entering the fray and new versions being announced at an ever-faster pace. The introduction of Anthropic’s Claude 3, Mistral-Large, and Stable Diffusion 3, for example, have drawn global attention, offering even more powerful models and creating new possibilities such as self-awareness and actualization, tier reasoning capabilities, and video generation.

With a variety of models designed for different needs, AWS’ strategy is to offer Amazon Bedrock as a way for customers to access the latest models on its platform and build AI applications with guardrails to meet their responsible AI framework. The solution also enables users to customize their models with retrieval augmented generation by using their own systems and data. Instead of competing with others to develop models for different needs, AWS is therefore playing to its strength by developing an AI ecosystem within its cloud and offering customers the tools and infrastructure to achieve their business objectives. To support AI training and inference, AWS is investing in infrastructure comprising new hardware (e.g., GPUs, AWS Trainium ML chips), AWS Inferentia accelerators, and Amazon EC2 Capacity Blocks) to deliver the performance and scale required. Moreover, it is developing GenAI-enabled applications that customers can readily deploy. This includes Amazon Q, a GenAI assistant that can be deployed within Amazon Connect (contact center solution) for enhanced customer experience; or within Amazon QuickSight (business intelligence tool) to produce faster, meaningful insights.

At AWS Summit Sydney, the company announced the launch of AWS Bedrock in the Sydney region. This is crucial to drive the adoption of AWS’ GenAI solutions locally by addressing the requirements around latency, data security, and industry-specific regulations. GlobalData believes that GenAI model training will either be performed in the cloud or in a hybrid environment since it is very costly to operate a dedicated AI infrastructure in-house. AWS is therefore well-positioned to capture the GenAI opportunities, particularly with its ability to support computing resources at scale. Australia is also an important market for AWS to prove its GenAI solutions. Australian enterprises have been early adopters of technology and it was ahead of many countries in cloud and mobile adoption over a decade ago. AWS has, for several years, been supporting many startups and cloud-native firms in Australia; and these digital businesses are more ready to explore and implement GenAI-enabled solutions. For example, at AWS Summit, Leonardo AI shared its experience working with AWS to power its AI image generator, allowing it to scale its business quickly – since the launch in December 2022, the company’s users (over 14.7 million) from 189 countries have generated over one billion images for gaming, media, entertainment, and marketing use cases. There is also a strong technology ecosystem in Australia as hyperscalers, global system integrators, application vendors, data center providers, etc. have significant presence in the country.

Moreover, a major challenge for supporting GenAI workloads is the power consumption and the need for efficient, large-scale data centers. As the Australian government is pushing the development of renewable energy, combined with the extensive renewable resources and vast landscape, the country will be able to operate these data centers without compromising on its goals to achieve its net-zero targets. AWS already operates regions in Sydney and Melbourne (Australia) as well as local zones in other cities, and it has an extensive partner ecosystem and customer base. With the launch of Amazon Bedrock in Sydney, the competition for GenAI solutions will start to intensify as many major integrators (also channel partners of AWS and competitors) are building multiple use cases transitioning to some of the first scaled deployments.