Graphic Processing Units (GPUs) are now very much at the forefront of computation. They’re helping us to meet our ever-increasing demand for artificial intelligence (AI), accelerating research and development, and boosting our data-driven workflows with powerful visualization capabilities.
Professional GPUs and their parallel processing capabilities and higher RAM configurations are used for different types of simulations, such as computational fluid dynamics (CFD). But what the UK’s workforce really needs is purpose-built machines that accelerate AI workloads locally.
The UK government has been keen to build up the country’s AI capabilities. Not long after the initial Labour cabinet was formed, it launched the ‘AI Opportunities Action Plan’ to identify ways in which AI can help drive economic growth and support infrastructure, talent, and data access. Its plans were backed by the International Monetary Fund, which estimated that by exploiting AI, the UK could ultimately see productivity gains of up to 1.5% per year.
"Growing our economy and rebuilding Britain to make everyone better off is our number one mission - and artificial intelligence has the potential to raise productivity and help us do that," Chancellor of the Exchequer Rachel Reeves said at the time.
Delivering on that potential requires the right tools and infrastructure. One example is the ASUS Radeon AI PRO R9700, a workstation-class GPU designed for developers, engineers, and professionals running data-intensive AI workloads locally.
ASUS has a strong presence in the UK with its GPU capabilities. Integrated security service provider Certis, for example, has used ASUS GPUs within its research hub, the Certis Centre for Applied Intelligence (CCAI). The hub is equipped with software and hardware capabilities for machine learning and deep learning in AI, robotics, and data analytics. And it’s powered by ASUS Cloud Infra, made up of ASUS ESC8000 G4 servers that use GPU architectures.
Introducing the ASUS Turbo Radeon™ AI Pro R9700
Developers and professionals need hardware built specifically for them and their data-intensive workloads. This is essentially what the ASUS Turbo Radeon AI Pro R9700 32GB is – a workstation-class graphics card for those who need powerful local performance in a compact system.
The Turbo Radeon AI Pro can run large memory-intensive AI models locally, and its two-slot profile is designed specifically for multi-GPU setups and space-constrained cases.
With its cutting-edge RDNA 4 Architecture and 2nd-gen AI Accelerators, the Radeon AI PRO R9700 can deliver as much as double the AI performance of the previous generation (Radeon PRO W7800), according to AMD tests using the DeepSeek R1 Distill Llama 8B benchmark. What’s more, it's equipped with 32GB of dedicated video memory, allowing developers to tackle large and complex projects without limits.
One of the most impressive elements of the Radeon, however, is RAM. The 32GB of GDDR6 (Graphics double data rate 6) memory provides significant benefits for handling large datasets, mainly through its high capacity and high bandwidth. It is that combination that allows for faster processing and fewer bottlenecks. Giving developers the ability to work with more complex models in applications like AI, machine learning, and high-performance computing (HPC).
This capacity of RAM also accommodates larger datasets and models, such as high-resolution medical images or 4K video footage. And it eliminates slow data transfers between system RAM and the GPU – a very common bottleneck with limited memory systems.
For machine learning, a larger video random-access memory (VRAM) capacity allows for the use of bigger batch sizes during model training. And, larger batches improve hardware usage, speed up the time it takes for a model to learn, and lead to more efficient training.
GDDR6 is designed for high bandwidth. It’s all about rapid data transfers between a GPU’s processing cores and the memory itself. A crucial factor for data-intensive workloads where the GPU needs to access and process massive amounts of data in parallel quickly.
Enterprise-class AI performance and cooling
Processing speed and scalability are two key demands from AI applications and their AI accelerators that deliver the near-instantaneous results that make these applications valuable. AI accelerators lower the latency of the time it takes to come up with an answer, and this low latency is especially important in safety-critical applications where every second counts.
While most new laptops can peak at around 40 TOPS of AI performance under ideal conditions, the Radeon AI Pro R9700 delivers a sustained 1531 TOPS, enabling significantly more demanding AI workloads. That amount of power can run large models locally and power the high-end workloads or an enterprise business.
However, all that processing – whether it’s in a data center or locally, on-device – generates immense heat. That is very much the case for all AI use cases – processing Large Language Models (LLMs), simulations, modeling for CAD software, data analysis, visualization, and even immersive VR experiences – advanced cooling systems are paramount.
The Radeon is equipped with a special blower-style fan with a dual-ball bearing. The bearing is up to twice the lifespan of a more traditional sleeve bearing, and it enables the fan to generate high static pressure. The fan is paired with a diecast metal shroud and backplate that efficiently dissipates heat. In fact, the design expels most heat directly out of the case, helping to keep system temperatures stable, which is needed to maintain consistently high performance, even under the most demanding AI workloads.
For more information on ASUS and its AI supporting hardware, you can visit the website here.
