Thread by @wmhuo168

Harvard built a paper-thin quantum chip. Media cheered. But China already left this field in the dust. Here's what they won't tell you.

Harvard’s ultra-thin chip could revolutionize quantum computing | ScienceDaily share.google/VZhvaR7N8knLqc…

🚨 Harvard’s metasurface chip is making headlines. But stop comparing it to Google. The real benchmark is China’s quantum supremacy. A thread. 🧵
(1/14)

Harvard just demoed a flat, nanoscale metasurface chip that can generate and control entangled photons. It could replace bulky optics used in quantum setups. It's a huge simplification play. But don’t confuse it for a leap in compute power.
(2/14)

This chip uses graph-theory design to control photon interactions on a wafer-thin surface. The ambition is a scalable, room-temperature, photonic quantum platform with less cost and fewer alignment headaches.
(3/14)

Now let’s talk real benchmarks. Forget Google’s Willow. Compare it to China’s Zuchongzhi-3.0, which achieved 10¹⁵ times the speed of the world’s fastest supercomputer on random circuit sampling.
(4/14)

Zuchongzhi-3.0 runs on 105 superconducting qubits with 182 couplers. It outperforms Google's Willow by more than a million times on core benchmarking tasks. Its fidelity levels are nearing production-grade.
(5/14)

Then there’s Origin Wukong, China’s 72-qubit superconducting platform. It’s already live on the cloud. Over 20 million remote visits and 300,000 tasks have been completed by users around the world.
(6/14)

Now back to Harvard’s metasurface chip. It's a sleek lab demo. It’s photonic and elegant. But it’s not a full-stack quantum computer. No error correction. No remote access. No benchmark results.
(7/14)

Harvard’s chip offers room-temperature operation and ultra-compact design. China’s Zuchongzhi and Wukong run at cryogenic temperatures. These are different trade-offs with different strengths. But only one side is running live compute tasks.
(8/14)

China’s scaling trajectory is aggressive. QuantumCTek’s latest control system already supports over 10,000 error-corrected qubits. That’s not a white paper. It’s in the commercial pipeline.
(9/14)

Photonic metasurfaces might become dominant in the future. But right now, they’re early-stage technologies. Harvard’s chip is a proof of concept. China’s chips are operating in production settings.
(10/14)

If you’re keeping score, this isn’t a contest between countries. It’s a contest between experimental simplicity and industrial-grade execution. And only one side is delivering real results at scale.
(11/14)

The press keeps spotlighting Google’s Willow. But the actual heavyweight in 2025 is USTC’s Zuchongzhi-3.0. No photonic chip, including Harvard’s, is even in the same league right now.
(12/14)

Bottom line: Harvard’s metasurface chip is an exciting step in quantum engineering. But the gold standard for quantum performance is being set in China.
(13/14)

Don’t let press releases and polished animations mislead you. One lab demo doesn’t overtake a quantum system already clocking one million times the performance of its Western rivals.
(14/14)