Explore this series: Crazy Rich Singapore article directory · Crazy Rich World — all countries and learning routes.
Crazy Rich Singapore | Supercomputing, ASPIRE 2B and the High Performance Computing Economy begins with a machine that is difficult to imagine because ordinary language runs out of scale.
Did you know Singapore launched ASPIRE 2B on 8 June 2026 as its newest and most powerful national supercomputer? NSCC says it can deliver up to 115 petaFLOPS of compute capability—more than 100 quadrillion calculations per second—with 184,320 CPU cores, 1,536 NVIDIA H200 GPUs, more than 1 petabyte of system memory and 63.5 petabytes of storage.
That makes “supercomputer Singapore”, “NSCC Singapore”, “ASPIRE 2B”, “HPC Singapore”, “AI computing Singapore” and “high performance computing Singapore” powerful search themes. The Crazy Rich story is not a giant computer sitting idle for prestige. It is a national research machine designed to let scientists model weather, discover materials, train AI, simulate engineering and explore problems too large for ordinary computers.
Crazy Rich Singapore can compress months of computation into something closer to hours or days.
Did You Know? A Supercomputer Is Not Just One Very Fast PC
High Performance Computing, or HPC, combines large numbers of processors so they can work together on difficult computational problems.
The system also needs:
- high-speed interconnects;
- large-scale memory;
- parallel storage;
- schedulers;
- specialised software; and
- expert users who know how to divide problems efficiently.
The machine becomes powerful because thousands of components cooperate.
ASPIRE 2B: Up to 115 PetaFLOPS
NSCC describes ASPIRE 2B as capable of up to 115 petaFLOPS.
A petaFLOP represents one quadrillion floating-point operations per second.
That means the machine can perform more than 100 quadrillion numerical operations every second under the relevant theoretical workload.
The important word is not only “fast”.
It is parallel.
Four Times the Previous Combined National Capacity
NSCC says ASPIRE 2B provides about four times the combined computing capacity of ASPIRE 2A and ASPIRE 2A+.
It also more than doubles Singapore’s national GPU capacity.
That increase matters because AI and scientific simulation are both becoming more computationally demanding.
Research ambition rises when compute stops being the bottleneck.
One Hundred Times More Compute Than a Decade Ago
NSCC’s launch material says ASPIRE 2B represents roughly a hundred-fold increase in compute power compared with ASPIRE 1, Singapore’s first national supercomputer launched a decade earlier.
That growth illustrates how quickly the research-computing frontier moves.
A machine considered extraordinary ten years ago can become a baseline reference within one research generation.
Compute capability itself has an upgrade cycle.
The 184,320 CPU Cores
ASPIRE 2B includes 184,320 AMD EPYC CPU cores.
CPUs remain important for workloads such as:
- large scientific simulations;
- engineering models;
- fluid dynamics;
- computational chemistry;
- climate models; and
- data preprocessing.
Not every problem is a GPU problem.
A heterogeneous supercomputer matters because different tasks prefer different processors.
The 1,536 NVIDIA H200 GPUs
ASPIRE 2B also includes 1,536 NVIDIA H200 GPUs.
GPUs are particularly useful for massively parallel workloads such as:
- AI training;
- deep learning;
- scientific machine learning;
- large numerical models; and
- accelerated simulation.
The modern supercomputer increasingly combines traditional HPC and AI acceleration inside the same research workflow.
Why Heterogeneous Computing Matters
NSCC explicitly describes ASPIRE 2B as a heterogeneous platform.
That means researchers can combine CPU-heavy and GPU-heavy parts of a problem instead of forcing everything onto one type of hardware.
For example:
a physics simulation may run on CPUs, produce a giant dataset, then use GPUs to train an AI surrogate model from the results.
The workflow becomes hybrid rather than separate.
The 63.5 Petabytes of Storage
A supercomputer does not only calculate.
It creates and consumes enormous datasets.
ASPIRE 2B includes 63.5 petabytes of storage.
That matters because scientific computing can fail operationally if the system computes faster than data can be stored or moved.
Storage becomes part of performance.
The 400 Gbps Slingshot Interconnect
NSCC lists a 400-gigabit-per-second Slingshot interconnect.
The interconnect allows compute nodes to exchange data rapidly.
That is critical because a parallel job may involve thousands of processors continuously communicating.
A supercomputer with slow internal networking would behave like a team of brilliant people forced to communicate by post.
Why Cooling Is Part of Computing
High-performance processors convert enormous amounts of electrical energy into heat.
That means computing performance depends on:
- power delivery;
- cooling;
- chilled-water systems;
- data-centre design; and
- operational resilience.
NSCC’s August 2026 ASPIRE 2A service incident, caused by a cooling and chilled-water issue at the hosting data centre, is a useful reminder.
Compute infrastructure is physical infrastructure.
The Data Centre Connection
Supercomputers live inside specialised data-centre environments.
That connects directly to Crazy Rich Singapore | Data Centres, Cloud and AI Infrastructure.
The supercomputer is the visible brain.
Power, cooling, networks and storage are the organs keeping the brain alive.
Climate Modelling Needs Supercomputers
Weather and climate models divide the atmosphere and oceans into computational grids and repeatedly solve mathematical equations across them.
Higher resolution and richer physics demand more computing power.
NSCC specifically identifies climate and weather science as an ASPIRE 2B use case.
That connects directly to Crazy Rich Singapore | Weather Radar, Climate Services and the Meteorological Economy.
The radar observes the atmosphere.
The supercomputer helps simulate what the atmosphere may do next.
Materials Discovery Needs Massive Search
Materials scientists may need to explore enormous combinations of atoms, structures and process conditions.
Simulation can reduce the number of physical experiments required before promising candidates emerge.
That connects to Crazy Rich Singapore | Advanced Materials, Polymers and the Materials Science Economy.
Compute becomes a laboratory where some experiments happen digitally before matter is ever manufactured.
The AI Connection
ASPIRE 2B’s large GPU capacity makes it a national AI research asset.
That connects directly to Crazy Rich Singapore | National AI Strategy, SEA-LION and the AI Assurance Economy.
Models become more capable as researchers can train on larger datasets, test more architectures and run more experiments.
Cheap intelligence still requires expensive compute somewhere underneath.
Why National Compute Matters When Commercial Cloud Exists
Commercial cloud is excellent for many workloads.
National research infrastructure solves a different problem.
It can provide strategically allocated compute for:
- publicly funded research;
- long-running scientific programmes;
- national priorities;
- sensitive research environments; and
- capability development across universities and institutes.
The question is not cloud versus supercomputer.
It is which infrastructure best fits the workload and national objective.
The July 2026 Strategic Resource Allocation Policy
NSCC revised its Strategic Resource Allocation policy in June 2026, effective from July, to incorporate ASPIRE 2B and rising GPU demand.
The policy prioritises national R&D and research projects according to categories and strategic value.
That matters because compute is scarce.
Even a 115-petaFLOPS machine cannot run every possible project simultaneously.
Allocation becomes science policy.
Why GPU Hours Become a Research Currency
Researchers request computing resources over time.
A project may require thousands or millions of CPU- or GPU-hours.
That makes compute allocation similar to granting access to an expensive laboratory instrument.
The scientist does not need to own the machine.
They need enough scheduled access to complete meaningful work.
The University Connection
NSCC serves researchers across institutions including NUS, NTU, A*STAR and other national research organisations.
That connects to Crazy Rich Singapore | Universities, Research and the Global Talent Economy.
A national supercomputer becomes shared research capital too expensive and specialised for every department to reproduce independently.
The Precision Medicine Connection
Genomics and precision medicine can generate enormous datasets requiring high-performance analysis.
That connects to Crazy Rich Singapore | Precision Medicine, Genomics and the Future of Healthcare.
The genome is biological.
Understanding millions of genomes becomes computational.
The Advanced Manufacturing Connection
Engineering simulations can test structures, airflow, thermal behaviour and production processes before expensive physical prototypes are built.
That connects to Crazy Rich Singapore | Jurong Innovation District and the Advanced Manufacturing Cluster Economy.
High-performance computing moves some trial-and-error from factory floor to virtual model.
The Quantum Connection
NSCC also highlights readiness for emerging hybrid quantum–classical computing approaches.
That connects to Crazy Rich Singapore | Quantum Computing, Quantum Networks and Deep Tech.
Future researchers may use classical supercomputers and quantum processors together rather than imagining one completely replacing the other.
Why HPC Talent Matters as Much as Hardware
A researcher can waste thousands of GPU-hours with poorly designed code.
High-performance computing therefore needs specialists who understand:
- parallel programming;
- job scheduling;
- memory;
- data movement;
- GPU optimisation; and
- scientific software.
NSCC’s role increasingly includes user enablement and capability development, not only operating machines.
The richest compute is compute people know how to use.
What Students Can Learn from Supercomputing
Mathematics
Large numerical models turn equations into billions of coordinated calculations.
Computing
Parallelism, memory and networking determine whether more processors actually make a program faster.
Physics
Cooling, electricity and heat remind us that computing remains a physical process.
Science
Simulation allows researchers to test hypotheses before or alongside physical experiments.
Economics
National research infrastructure pools expensive capital so many institutions can use it.
Ten Vocabulary Words for the HPC Economy
1. Supercomputer
A very high-performance computing system built for large-scale scientific, engineering or AI workloads.
2. HPC
High Performance Computing, the use of parallel computing systems for computationally demanding problems.
3. FLOPS
Floating-point operations per second, a measure of numerical computing performance.
4. PetaFLOPS
Quadrillions of floating-point operations per second.
5. GPU
Graphics Processing Unit, a processor well suited to many massively parallel workloads.
6. CPU core
An independent processing unit inside a central processor.
7. Interconnect
The high-speed network connecting compute nodes inside a supercomputer.
8. Parallel computing
Splitting work across many processors operating at the same time.
9. Scheduler
Software allocating computing resources and deciding when jobs run.
10. Heterogeneous computing
Using different processor types together within one computing environment.
Frequently Asked Questions
When was ASPIRE 2B launched?
8 June 2026.
How powerful is it?
NSCC says it delivers up to 115 petaFLOPS of compute power.
How many GPUs does it have?
1,536 NVIDIA H200 GPUs.
How much storage does ASPIRE 2B have?
63.5 petabytes.
What is it used for?
NSCC highlights climate and weather science, healthcare, materials discovery, advanced manufacturing, AI and emerging hybrid quantum-classical research among its intended applications.
Helpful Reading Across the Singapore Graph
- Crazy Rich Singapore | Data Centres, Cloud and AI Infrastructure
- Crazy Rich Singapore | National AI Strategy, SEA-LION and the AI Assurance Economy
- Crazy Rich Singapore | Weather Radar, Climate Services and the Meteorological Economy
- Crazy Rich Singapore | Quantum Computing, Quantum Networks and Deep Tech
References and Current Sources
- National Supercomputing Centre Singapore, ASPIRE 2B, current system specifications.
- National Supercomputing Centre Singapore, NSCC Singapore Launches ASPIRE 2B, 8 June 2026.
- National Supercomputing Centre Singapore, Strategic Resource Allocation Policy, updated 1 July 2026.
- NSCC Help Centre, ASPIRE 2A FAQs, current background on Singapore’s prior national HPC system.
Crazy Rich Singapore Buys Time with Compute
Scientific research is often limited by how many experiments can be run, how finely a system can be modelled and how much data can be analysed.
Supercomputing changes those limits.
Did you know? One of Singapore’s richest machines is valuable not because it thinks for us, but because it lets researchers ask larger questions before time runs out.