What becomes possible when researchers are no longer constrained by compute?

That may be one of the most important questions emerging across Singapore's research and innovation ecosystem today.

Singapore's long-term investments in research, innovation and enterprise have always been focused on a practical goal: translating research excellence into real-world impact. Research, Innovation and Enterprise 2030 (RIE2030) led by The National Research Foundation, Singapore, continues that ambition, with a strong emphasis on scientific excellence, economic competitiveness, resilience, and outcomes that benefit society.

At the same time, Singapore's National AI Strategy 2.0 recognises the growing importance of AI capability and adoption across the economy.

What is becoming increasingly clear is that advances in AI are not only creating new applications. They are also changing how research itself gets done.

At RE:AI (Singtel), we see this as a shift from AI as a tool to AI as a capability multiplier.

The value is not simply faster processing.

The value lies in enabling researchers to explore more scenarios, analyse larger datasets, run more sophisticated simulations, and tackle problems that would otherwise be difficult, costly, or time-consuming to pursue.The recent collaboration between Nanyang Technological University Singapore (NTU Singapore) and RE:AI offers an early example of this shift in practice.

Researchers across urban sustainability, healthcare, robotics, and earthquake modelling are using RE:AI, Singtel Digital InfraCo's on-demand GPU-as-a-Service platform powered by NVIDIA H200 GPUs, to support computationally intensive workloads.

According to Singtel and NTU, some simulations that previously took more than a day on conventional CPU-based systems can now be completed in approximately 10 minutes. Certain earthquake simulations that previously took up to a week can now be completed within a day.

Different research domains. Different societal challenges. Yet a common pattern is emerging.

In urban environmental modelling, researchers are accelerating simulations that can help improve understanding of urban heat, ventilation, pollution dispersion, energy efficiency, and climate resilience.

In healthcare, large pathology imaging datasets are being processed to train deep learning models that could support more efficient tissue assessment workflows and more timely clinical information.

In robotics, researchers are developing AI models designed to help robots make better decisions in factories, laboratories, and service environments.

In earthquake research, larger and more realistic simulations are helping scientists better understand earthquake and fault behaviour and improve future seismic hazard modelling.

The common thread is not the technology itself.

It is the ability to push research further.

Across these projects, access to advanced AI and GPU capabilities is helping researchers work at greater scale, test more complex hypotheses, and accelerate computationally intensive experimentation.

For us, that is the more significant story. As Singapore advances the ambitions of RIE2030 and NAIS 2.0, the conversation should not be limited to AI models or compute performance alone.

It should also focus on what researchers, institutions, and innovation ecosystems can achieve when advanced AI capabilities become operationally accessible.

The NTU and RE:AI collaboration offers a practical example of that future beginning to take shape. 

Sources

1. National Research Foundation Singapore, RIE2030 https://www.nrf.gov.sg/rie2030/
2. Singtel, Singtel Digital InfraCo's RE:AI powers faster, scalable research at NTU Singapore https://www.singtel.com/about-us/media-centre/news-releases/singtel-digital-infracos-reai-powers-faster-scalable-research-at-ntu-singapore
3. NTU Singapore, NTU scientists power up research with Singtel AI cloud https://www.ntu.edu.sg/news/detail/scientists-power-up-research-with-singtel-ai-cloud
4. Singapore National AI Strategy 2.0 https://file.go.gov.sg/nais2023.pdf