Most quantum computers today use superconducting circuits or trapped ions.
Jiuzhang 4.0 is different — it runs on photons, the tiny particles that make up light.
The team led by Chao-Yang Lu at the University of Science and Technology of China has pushed the limits of this approach using a method called Gaussian boson sampling. This is a special kind of problem that is nearly impossible for classical supercomputers to solve, but it’s something a photonic quantum machine like Jiuzhang 4.0 can handle.
Jiuzhang 4.0 shows us that the future of computing is not just faster, but completely different — and it’s being built with light.
A Leap Beyond Classical Power
Jiuzhang 4.0 recently set a new record by finishing a Gaussian boson sampling task in just a few microseconds. To compare, the world’s fastest supercomputers would take an unimaginably long time — estimates go beyond trillions of years. For perspective, the universe itself is only about 13.8 billion years old.
This isn’t just a matter of speed. Jiuzhang 4.0 is showing a completely new kind of computing power — one that classical computers simply cannot match.
How Jiuzhang 4.0 Works
The way Jiuzhang 4.0 works is surprisingly simple at its core.
Light particles (photons) are sent into a network of mirrors and beam splitters.
These photons travel along many possible paths at the same time.
When they are measured at the end, the result is like picking from a very complex set of possibilities.
This process is called Gaussian boson sampling. On its own, it doesn’t have direct everyday uses yet. But it’s extremely valuable as a test, because classical computers simply cannot handle these kinds of calculations once the number of photons becomes very large.
Why It Matters for Data Science
For data scientists, Jiuzhang 4.0 gives a peek into the future:
•Many tasks in machine learning and optimization depend on sampling.
•Generative AI models also work by sampling from complex probability distributions.
•Fields like network analysis and drug discovery involve heavy computations that could one day be sped up by quantum machines.
Right now, Gaussian boson sampling is mainly a demonstration. But it shows where things are heading: toward computers that don’t just run faster, but open up new kinds of problems we’ve never been able to solve before.
Looking Ahead
Jiuzhang 4.0 is not yet a computer we can use for everyday tasks. But with each new version — from Jiuzhang 1.0 in 2020 to today — photonic quantum computing is becoming more powerful and stable. The key message is clear: quantum advantage is no longer just theory, it’s starting to happen.
For data scientists, this raises new questions:
Which problems that seem impossible today might one day be easy for quantum machines?
How can we start building algorithms and workflows that mix classical and quantum computing?
The future of computing won’t just mean faster machines — it will change what kinds of problems we can solve. Jiuzhang 4.0 shows that this future of light-speed computing is, quite literally, powered by light.
And as the community looks ahead, DSC Next 2026 will be the stage where these questions evolve into strategies. By then, we may not just be discussing benchmarks like boson sampling, but real-world applications where photonic quantum machines begin to intersect with AI, optimization, and large-scale data challenges. If 2025 was about proving that light can outperform silicon, 2026 will be about showing how this quantum leap can illuminate the path for data science itself.
Reach us on WhatsApp