The Hardware Race: Four Ways to Build a Qubit
Superconducting, trapped-ion, neutral-atom, photonic — each builds a qubit from something different, and each wins on some axes while losing on others. No clear champion yet.
Lesson 13 mentioned that qubits can be built in different physical ways; now that you understand what a good qubit must do — stay coherent (lesson 14), connect to others for checks (lesson 31), and scale to millions (lesson 32) — we can finally ask the question the whole industry is betting billions on: which physical platform will win? There are several serious contenders, each making a qubit out of something completely different, and — crucially — none is clearly ahead, because each is strong on some axes and weak on others. Hold the question: what would you actually compare these machines on to judge which is winning?
The four axes that decide the race
Before the contenders, the scorecard. Every platform is judged on roughly four things (each grounded in earlier lessons). Coherence: how long a qubit holds its state before decohering (lesson 14) — longer is better. Gate speed: how fast you can perform operations — faster means more computation before errors pile up. Connectivity: how easily a qubit can interact with others (which the surface code’s checks demand, lesson 31) — more connections make error correction and algorithms easier. And manufacturability/scalability: how realistically you can build millions of them (lesson 32). Here’s the catch that defines the whole field: no platform wins on all four — each trades some against others, so the race is about which tradeoff pans out, not which is flatly best.
The four contenders, in brief
Superconducting qubits are tiny circuits chilled near absolute zero. Very fast gates and built with chip-fabrication techniques (good for scaling) — but relatively short coherence and limited to nearest-neighbor connectivity. Trapped-ion qubits are individual charged atoms held by electromagnetic fields. Excellent coherence and high connectivity (ions can be made to interact flexibly) with very precise operations — but slower gates and harder to scale to huge numbers. Neutral-atom qubits (atoms held by laser “tweezers”) are a fast-rising middle path: good coherence, flexible connectivity, and promising scaling to many atoms. Photonic qubits use particles of light — they barely decohere and fly at light speed (great for networking), but making photons reliably interact to do gates is notoriously hard. Four bets, four different strengths.
Same scorecard, different profiles: • Superconducting: speed ✔, scalable-fab ✔, coherence ✘ (shorter), connectivity ✘ (nearest-neighbor). • Trapped-ion: coherence ✔, connectivity ✔, precision ✔, speed ✘, scaling ✘. • Neutral-atom: coherence ✔, connectivity ✔, scaling promising; still maturing. • Photonic: coherence ✔ (barely decoheres), networking ✔, but making photons interact for gates ✘. • Every row is a different mix of strengths — which is exactly why there’s no obvious winner.
Why there’s no winner yet — and how to read the race
Put the scorecard together and the honest picture emerges: because each platform is strong on some axes and weak on others, the winner depends on which weaknesses get solved first — and that’s genuinely unknown. A fast-but-flaky platform wins if error correction (this module!) tames its errors; a pristine-but-slow one wins if it can be scaled; a barely-decohering one wins if its gates get solved. This is why the field hedges across all of them. The practical takeaway, echoing lesson 17’s hype-vs-reality discipline: when you read that a platform hit “the most qubits” or “the longest coherence,” ask about the other axes — a record on one axis says little if another is failing. The right mental model isn’t “who’s ahead?” but “which tradeoff is closest to crossing every threshold at once?” — and, honestly, we don’t yet know.
It’s like the early days of a transportation race with cars, trains, planes, and boats all being invented at once. The car is fast and easy to mass-produce but limited to roads; the train carries huge loads reliably but needs tracks laid everywhere; the plane is blazingly fast but brutally hard to build; the boat barely wears out but is slow to maneuver. Which “wins” isn’t obvious from any single stat — top speed, capacity, range, cost of manufacture all differ — and the champion depends on which weakness gets engineered away first. The qubit platforms are exactly that: four fundamentally different vehicles, each leading on a different gauge, no clear overall winner yet.
Reading a hardware headline honestly: 1. “Company X hits record qubit count!” → which platform, and at what coherence and error rate? A big count of flaky, poorly-connected qubits may be worth less than fewer excellent ones. 2. “Company Y sets a coherence record!” → great on that axis — but how fast are its gates and can it scale? 3. Each claim is one row of the scorecard; the machine’s real promise is the whole row. 4. The winner will be whichever platform first gets all four axes past the bar at once — which no one has done, so the race is open.
This is the reading. The interactive version — active-recall quiz, a hands-on experiment you run in your own AI, and an earned mastery check — is free in the app.
Start this lesson free →