Nvidia Simulates 104-Qubit Fault-Tolerant Quantum Computer Using JUPITER Supercomputer
Nvidia used the JUPITER supercomputer to run the largest full simulation of a fault-tolerant quantum machine, reducing qubit needs by 90 percent.
Key points
- Nvidia simulated a 104-qubit error-correcting quantum computer on Germany’s JUPITER system.
- The simulation utilized magic state cultivation to reduce required qubits by 90 percent.
- The run accounted for real-world hardware noise, unlike previous idealized tests.
Nvidia has completed a full simulation of a 104-qubit fault-tolerant quantum computer using the JUPITER supercomputer located at Germany’s Jülich Supercomputing Centre. According to The Next Web, the company states this represents the largest complete simulation of such a machine to date. The exercise demonstrates how conventional supercomputing resources can model complex quantum architectures before physical hardware is constructed.
The simulation incorporated realistic environmental noise that typically degrades qubit performance in actual quantum processors. Qubits are highly sensitive to interference from control equipment, which introduces errors during calculations. By modeling these imperfections, the team validated a fault-tolerant design that can detect and correct mistakes while operating. Most experts view error correction as a critical prerequisite for making quantum computers practically useful for real-world tasks.
Technical Methodology
Nvidia tested a technique known as magic state cultivation to facilitate specific quantum calculations. These special quantum states are necessary for many operations but traditionally demand a large number of additional qubits to create. The company reported that its method uses 90 percent fewer qubits than earlier approaches, significantly improving resource efficiency. This optimization addresses a major bottleneck in scaling quantum systems without requiring exponentially more physical components.
The entire simulation ran on conventional computers rather than a quantum processor. Nvidia first developed and tested the model on a smaller system at its own quantum lab, which reduced the computational workload from weeks to days. For the final large-scale run, the team utilized Nvidia GH200 chips across 8,192 GPUs. This massive parallel processing power allowed for the detailed modeling of the 104-qubit system’s behavior under realistic conditions.
Infrastructure and Context
JUPITER, which opened in September 2025, was Europe’s first exascale supercomputer capable of performing more than a billion billion calculations per second. Researchers at Jülich previously simulated 50 qubits on the same system in November 2025 under ideal, noise-free conditions. The latest work builds on that foundation by introducing the complexity of real-world hardware noise, providing a more accurate preview of future quantum machine performance.
Nvidia CEO Jensen Huang noted that useful quantum computing will require fantastically complex, fault-tolerant systems. Kristel Michielsen, director of the Jülich Supercomputing Centre, stated that Europe’s investment in JUPITER gives researchers the ability to investigate questions that were previously beyond reach. The announcement coincides with Huang preparing to deliver his keynote at Nvidia’s GTC conference in Berlin on Wednesday.
Industry Implications
Germany signed an AI partnership with Nvidia at an investment summit in the capital earlier this week. The simulation results highlight the growing synergy between European supercomputing infrastructure and US hardware innovation. By enabling researchers to test quantum designs virtually, the work aims to accelerate the development timeline for physical quantum computers. This approach allows for iterative design improvements without the high costs and delays associated with building experimental hardware prototypes.
What should you do?
Security and technology professionals should monitor advancements in quantum simulation for implications on cryptographic standards. Organizations should audit their current encryption protocols to ensure they align with future quantum-resistant guidelines. Teams should track vendor updates regarding quantum-safe algorithms as the industry moves toward fault-tolerant systems. IT leaders should plan for long-term infrastructure transitions that account for emerging quantum capabilities.
What to do and how to stay safe: Nvidia
- Review current encryption standards to identify vulnerabilities to future quantum attacks.
- Monitor industry updates on quantum-resistant algorithms and migration timelines.
- Assess organizational readiness for post-quantum cryptography implementation.
- Track developments in quantum simulation for early indicators of hardware maturity.
Step-by-step guide: Mobile Device Security: How to Protect Data Beyond Passwords
General security guidance from the Threat Times newsroom. It is not confirmed advice from the organisations named in this story.
Frequently asked questions
How many qubits did Nvidia simulate in this test?
Nvidia simulated a 104-qubit quantum computer that can correct its own errors.
What supercomputer was used for the simulation?
The simulation ran on JUPITER, the supercomputer at Germany’s Jülich Supercomputing Centre.
How does this simulation differ from previous work?
This test accounted for real-world hardware noise, unlike a previous 50-qubit simulation that used ideal, noise-free conditions.
