Machine learning just cracked quantum error correction in under a microsecond — and governments are betting billions on what comes next. From IBM's real-time AI decoder to China's cryogenic-free neutral-atom systems, the quantum race just shifted from physics to engineering.
Audio is available on Spreaker — see link below.
Machine learning just solved a problem that's been blocking quantum computing for decades. Real-time AI error correction, running in under one microsecond, has cracked the scaling bottleneck that kept quantum systems too fragile to be useful.
On the same day this was becoming clear, the US Department of Commerce moved in a way that signals how seriously Washington is reading this moment. Two point zero one three billion dollars distributed across nine quantum firms through the CHIPS Act, including one billion to IBM and three hundred and seventy-five million to GlobalFoundries.
China's response is already in hardware. Origin Wukong-180, a fourth-generation one hundred and eighty qubit superconducting system, went online in early May with ninety-nine percent accuracy.
The clearest signal of maturation came from USTC, where Jiuzhang 4.0 was published in Nature. The photonic system manipulated three thousand and fifty photons, compared to two hundred and fifty-five in its predecessor, and solved a benchmark problem ten to the power of fifty-four times faster than the world's most powerful supercomputer.
On the capital side, Quantinuum, the Honeywell-backed trapped-ion firm, filed for an IPO at a twelve point seven billion dollar valuation. Swiss firm Terra Quantum reached a three point five billion dollar SPAC agreement.
The uncertainty that matters most is whether AI-driven error correction scales reliably across all four hardware approaches at volume. Superconducting, trapped-ion, photonic, and neutral-atom systems each have different error profiles.
Chapter summary auto-generated from the verified script. Listen to the full episode for the complete content.