Hyperion Research’s AI Beacon:

Illuminating AI at Scale

The world changed in the fall of 2022 with the introduction of highly visible AI capabilities via ChatGPT. Since then, both the AI and traditional HPC markets have undergone a revolution and are exhibiting growth rates never seen before.

Over the last decade, Hyperion Research has been tracking the AI and big data markets in greater depth and is now launching an updated view & tracking of the AI market, technologies and use cases for AI applications that use advanced computational resources.

In its continuing efforts to cover the most compelling trends in advanced computing, Hyperion Research is expanding its research centered on AI at Scale, gathering data and providing analysis on the most important changes, challenges and opportunities in this rapidly evolving sector.

Our coverage targets the software, hardware, data and models used to leverage AI at all levels of science and engineering computing as well as in new enterprises and organizations moving to apply AI across a broad set of application areas. Our data-focused analysis is informed by a continual series of surveys/interviews of computing developers and end users from around the world, spanning academia, government, and industry organizations.

Key AI-focused Studies

Updated AI/HPC Five Year Forecast Database with Verticals: currently available

AI ROI Study: currently available

This study examines how high-performance computing users and those in the advanced scientific fields are investing in artificial intelligence with a special emphasis on the evolving expectations and realities of return on investment. It explores user sentiment around costs and benefits, discusses the challenges involved in achieving lasting value, and highlights the shift from rapid adoption towards a more measured integration. Readers will gain insight into how industry leaders are balancing innovation with the pursuit of tangible ROI in a rapidly changing landscape.

Click here for more information. To purchase this document contact Mike Thorp at mthorp@hyperionres.com

AI in the Cloud Study: currently available

This study captures insights at the intersection of advanced AI usage and cloud computing resources. Respondents answered questions and offered insights regarding the relationship between on-premises resources and cloud computing. It also provides information on various stages of integration into production environments, which application spaces are being most often accelerated or enabled, and budgetary considerations for provisioning compute and infrastructure resources for growing AI/LLM capabilities. The report and study also provide an overview of the cloud computing landscape of the major providers and lists specific cloud resources most often used for AI at scale.

Click here for more information. To purchase this document contact jeansorensen@hyperionres.com.

LLM Study: currently available

This study was conducted across industry verticals, governmental organizations, and academic institutions to capture the current activity, intent, and details regarding experimental and production LLM applications. The respondents provided insights on model types, integration techniques, support requirements, and budgetary considerations. This report brings into focus the emergence of AI and LLM capabilities in a wide range of application spaces as well as the rate at which this technology is being integrated into key sectors and industry verticals. It also contains information on the landscape of LLM and AI development from proprietary to open-source offerings.

Click here for more information. To purchase this document contact jeansorensen@hyperionres.com.

End User AI Inferencing: currently available

The study is targeted towards the inferencing side of production and near-production integration of advanced AI/LLM. The associated survey delves into the hardware and software requirements of user groups and organizations managing high inference demands, as well as related budgetary and infrastructure requirements. Survey respondents will provide insights on their specific inference types, level of integration and experimentation, and other details of their advanced AI usage including plans and methods of scaling.

For information about any of the above studies, contact Jean Sorensen at jeansorensen@hyperionres.com . Please put “AI Studies” in the subject header.

AI Advisory Committee

Expanded AI Program [Click Here]

The mission of the AI Advisory Committee is to support the worldwide AI/HPC community by helping answer key questions about how AI is evolving. Members are from all over the world and from various areas of the AI ecosystem (users, vendors, CSPs, academic experts, etc.). Members will receive the ability to help guide Hyperion Research studies with questions are important to the Committee members.

For more information, please click on the link above or contact Earl Joseph at ejoseph@hyperionres.com

2024 AI End-User Study (also known as Multi-Client Study):

Hyperion Research’s High-Performance Computing (HPC) End-User Multi-Client Study 2024 (MCS) is the seventh edition of a comprehensive study that surveys a large number of HPC customer sites worldwide to create a detailed profile of HPC workloads, plans, and spending on various aspects of HPC, including AI. The most recent version of the study profiled 3,830 HPC systems at 181 sites worldwide, with an emphasis on industrial/commercial sites (60.2% of the sample) to allow confident “drill downs” into specific industry segments. In all categories, sites were chosen to reflect a wide range of users within the global HPC community.

For information about the AI End-User Study, contact Jean Sorensen at jeansorensen@hyperionres.com . Please put “AI End-User Study” in the subject header.

AI Subscription Service

Hyperion Research’s AI rolling 12-month subscription service (previously called AI-HPDA) has a 10 year history of targeted documents summarizing and providing analysis on recent studies, news, recent technology developments, products, and HPC User Forum AI-related presentations.