The Uncomfortable Truth Behind Quantinuum’s Earnings and Its “Near Five-Nines” Claim. Disclaimer: This post is not intended to discredit Quantinuum or favor any particular company. My concern is simply that many investors make investment decisions—or repeat claims in discussions—without understanding the underlying facts and technical context. The purpose here is to look beyond the headlines and clarify what was actually demonstrated. Quantinuum has been at essentially the same physical-gate fidelity level for years. In 2023, its H-Series was already around 99.997% 1Q fidelity and roughly 99.8% 2Q fidelity. By late 2025, Helios reported 99.9975% 1Q and 99.921% 2Q. The improvement is real, but incremental. Helios’ published hardware benchmark reports those latter figures directly. The bigger headline has been logical qubits: Quantinuum has promoted 94 error-detected LQs (it doesn't mean the error is corrected) and 48 error-corrected LQs from just 98 physical qubits. Now it is advertising “near five-nines” logical fidelity. But the underlying result is about 99.996% logical QEC-cycle fidelity, and the experiment still uses post-selection—detected bad outcomes are excluded, and the fidelity is evaluated on the accepted shots. That is still a meaningful QEC result. But it is not the same thing as demonstrating 99.999% universal logical-gate fidelity, nor is it equivalent to 48 utility-scale fault-tolerant logical qubits. The headline sounds much stronger than what was actually demonstrated.