The quantum endgame isn't whose qubits are cleaner, it's who can mass-produce them on a phone-chip line
The endgame of quantum computing isn't whose qubits are cleaner, it's who can mass-produce them on existing semiconductor lines; silicon spin qubits reuse TSMC's lithography, while trapped ions are stuck at a hundred thousand laser beams and superconductors at warehouse-scale refrigerators.
The video won't play here. Listen to the audio instead:
The argument · tap a timestamp to hear it
Trapped ions and superconductors must invent a miracle first; silicon doesn't
To build a hundred-thousand-qubit trapped-ion machine, you first have to invent the optical miracle of a hundred thousand laser beams; for a million superconducting qubits, you have to build a warehouse-sized cryostat; a million spin qubits only need a standard 300mm silicon wafer, run through the same lithography equipment TSMC, ASML and Intel use to make phone processors. Krishna uses this to set the tone for the whole set: the key to the quantum race isn't single-qubit fidelity, it's whether a path to scale already exists inside the industrial system.
— KrishnaQuantum computing has stopped comparing principles and started comparing engineering
In 2000 DiVincenzo proposed five criteria, with the aim of ruling out liquid-state NMR quantum computing. In this episode FFP proposes three criteria of its own: qubit quality, control, and scalability and economics. The rest of the session is a point-by-point scoring of the four routes — superconducting, trapped ion, neutral atom, silicon spin. Laying out the evaluation framework is itself telling you: quantum computing has entered the engineering-competition stage, and you can't just talk about principles.
— KrishnaSuperconducting isn't a failure of technology, it's an engineering dead end
Estimates suggest a thousand logical qubits would need ten million transmon physical qubits. The number of control lines is limited by the cooling capacity of the dilution refrigerator, frequencies are crowded, and the helium-3 supply becomes a new resource constraint. FFP gives superconducting just 1/10 on scalability and economics, for a total of 11/30. This is the first time in the session that "technically strong" and "engineering-infeasible" are scored separately; a warehouse-sized cryostat means even a data center may not be able to house it.
— HostTrapped ions win on coherence time and lose on gate speed
Trapped-ion coherence times run as long as ten hours, but gate operations rely on phonons and are limited by the Paul trap's oscillation frequency, making them about a thousand times slower than superconducting. The result is counterintuitive: even if trapped ions could scale to cryptobreaking size, breaking 2048-bit RSA would take about a year, whereas superconducting would need only about eight hours. Long coherence time doesn't equal more done per unit time; for practical computation, slow gate times are a throughput catastrophe overall.
— HostThe quantum threat is neither tomorrow nor forever
A March 2026 paper from Harvard, Caltech and Qera says ten thousand qubits might break internet encryption, but in practice it would take years; even with a hundred thousand physical qubits, breaking RSA under the mathematically optimal assumptions would take about three months. This number strikes at both extreme narratives at once: the quantum threat isn't tomorrow, but it is by no means forever distant. To achieve "breakable within months," qubit count, gate speed and error correction all have to hit the mark at the same time.
— GuestThe life or death of quantum companies rests in government contracts
HRL lost key project funding in early 2026 and laid off 376 people. Government funding shifted from basic research toward applied research demanding rapid commercialisation; IBM and Intel received support under the CHIPS Act, and eventually IBM acquired HRL. Krishna calls this the debut of the quantum computing MVP: a quantum company's survival depends on government contracts, and being acquired isn't failure but rather letting the technology enter an industrial system with manufacturing capability.
A million qubits can't be tuned by hand, only handed to image recognition
Automated tuning becomes the key to scaling: a 2023 paper had only six electron spins that could be tuned by hand, and hand-tuning a million qubits would take a human lifetime. Tuning is essentially image recognition — when scanning voltages, tunnelling events show up as fringes, and Krishna's team used DETR, combining ResNet and a Transformer, to automatically recognise the fringes and generate parameters. This is also the hidden premise of why silicon spin can reuse semiconductor lines: no matter how many qubits there are, software can align them automatically.
— KrishnaSilicon may not be first to be made, but it will be the last to win
FFP's final scores: silicon spin gets 10/10 on qubit quality, 10/10 on control (provided the isotope enrichment and valley state problems are solved), 10000/10 on scalability and economics, for a total of 10020/30; by contrast neutral atoms score minus 3000. Krishna concedes silicon may not be the first route to achieve fault tolerance, but once it does, being cheap and easy to manufacture will make it the eventual standard, just as the transistor replaced the vacuum tube.
— KrishnaIn their own words · checked verbatim
But if you want a million spin qubits, you just put it on a standard 300 millimeter silicon wafer. You run it through the exact same photolithography machines at TSMC, ASML, Intel.
Krishna0:00
A quantum computer is not a machine that tries every single answer at the same time because a qubit is a zero and a one simultaneously.
Krishna23:39
This is where the startup pitch decks collide with reality. They collide with the laws of thermodynamics and the laws of economics, which are not real laws, which is why the economics Nobel Prize is not a real Nobel Prize.
For a thousand plus logical qubits. Which is not the million we talked about earlier. No. It's just a thousand. Thousand logical qubits. You need 10 million physical qubits of these transplants.
Host1:24:33
Here, even if we achieve scale, because the gate times are so slow, it might take like on the order of a year to decrypt 2048 RSA.
Host1:47:08
the current evidence is not enough. And you need to show, like, you know, I would expect, for example, a statistical analysis of valley splitting across devices.
reviewer3:13:48
I don't think it's going to be the future quantum computer. You know, once it's achieved, I think silicon is going to be the one that actually gets scaled because it's cheap and it's easy to make.
Krishna3:33:40
Figures
| Nature paper co-authors | 250 people | 3:07 |
| Nature paper publication date | July 30, 2026 | 3:07 |
| Superconducting qubit FFP score | 11/30 | 1:25:33 |
| Trapped-ion coherence time | 10 hours | 1:35:45 |
| Time for 100,000 physical qubits to break RSA | about three months | 2:04:35 |
| Share of 28-silicon in natural silicon | 92% | 2:38:43 |
| HRL layoffs | 376 people | 2:50:02 |
| CNOT error rate in the HRL paper | 0.09% | 3:07:30 |
| Silicon spin FFP total score | 10020/30 | 3:32:31 |
| Neutral atom FFP total score | -3000 | 3:32:31 |
Glossary
- transmon
- A type of superconducting qubit that uses a Josephson junction to introduce anharmonicity, making the energy levels unequal and avoiding mistaken operations.
- valley splitting
- The energy-level difference after the degeneracy of silicon's multiple electron valleys is lifted, which determines silicon spin qubit quality.
- 2DEG
- Confines electrons to move within a plane; it is the physical basis of silicon quantum dots.
- DETR
- A detection transformer, an object-detection model used here to recognise tunnelling fringes in tuning images.
- exchange-only qubit
- A qubit that can implement quantum gates using only the exchange interaction between electrons, with no microwaves, but each qubit needs three electrons.
How to listen
Investors watching the choice of quantum computing route, semiconductor process engineers, and security practitioners who want to calibrate the quantum-cryptobreaking timeline.
Listeners already familiar with the principles of superconductors and trapped ions can skip 57:56 to 1:28:00.