IBM Condor: What Its 1,121-Qubit Processor Means for Quantum Computing

IBM's Condor processor, unveiled in late 2023, packs 1,121 superconducting qubits onto a single chip. It is the most quantum bits any company has assembled in one device, and it still cannot run most useful algorithms at useful length. That gap—between impressive hardware and practical payoff—defines the field in 2025.

Four Routes to a Qubit

Classical computers encode data as bits that are either 0 or 1. Qubits use superposition to hold a mix of states and entanglement to correlate with other qubits, which lets a quantum machine explore certain problems exponentially faster than any classical computer. Factoring the large numbers behind RSA encryption, simulating molecular interactions, and optimizing systems with millions of variables are the headline targets.

But qubits are extraordinarily fragile. A stray photon, a vibration, or a thousandth-of-a-degree temperature shift can collapse their state, which is why most machines run near absolute zero at -273.15 degrees Celsius. Companies are pursuing four different physical implementations, each with a different trade-off.

IBM and Google use superconducting circuits—fast gate speeds and compatibility with chip fabrication, but coherence times measured in hundreds of microseconds. IonQ and Quantinuum trap ions with lasers, achieving coherence times of seconds and gate fidelity above 99.9 percent at the cost of speed. PsiQuantum and Xanadu use photonic qubits that operate at room temperature but struggle with deterministic two-qubit gates. Intel and UNSW are betting on silicon spin qubits, which could one day use existing chip fabs to scale to millions.

NISQ Reality Check

The industry is in what researchers call the NISQ era: Noisy Intermediate-Scale Quantum. Today's machines have enough qubits to do things classical computers cannot easily simulate, but too much noise to run long, complex algorithms reliably.

Progress is real, though. Google demonstrated in 2024 that adding physical qubits to form a logical qubit actually lowers error rates, confirming a prediction that had been debated for decades and removing a key doubt about whether fault-tolerant quantum computing is physically possible. IBM has a roadmap to more than 100,000 qubits by 2033. Most estimates put useful fault-tolerant machines somewhere between 2029 and 2035.

Even within NISQ constraints, narrow applications are producing value. In 2024, a Harvard-MIT-Quantinuum collaboration used a 56-logical-qubit system to simulate a lithium hydride molecule with chemical accuracy, a result classical machines can only approximate. Better molecular simulation matters because the Haber-Bosch fertilizer process alone consumes about 2 percent of global energy. D-Wave's annealing machines, specialized for optimization rather than general computation, are used commercially by Volkswagen for traffic routing and by JPMorgan Chase for portfolio risk analysis.

The Cryptography Clock Is Ticking

The most immediate practical effect of quantum computing is defensive. A sufficiently large machine running Shor's algorithm could break the RSA and elliptic-curve encryption that secures most internet traffic. NIST finalized its first post-quantum cryptography standards in 2024, and organizations are beginning the long migration to quantum-resistant algorithms—a process expected to take a decade or more given the scale of legacy infrastructure.

Capital has flowed in fast: cumulative funding exceeded $28 billion globally through 2025, with the U.S. committing $5.4 billion through the National Quantum Initiative, China an estimated $15 billion, the EU 7.8 billion euros through its Quantum Flagship, and the UK 2.5 billion pounds. Unlike the semiconductor industry, which consolidated around silicon decades ago, quantum has no obvious winning hardware approach, and that genuine openness is why IBM, Google, IonQ, PsiQuantum, and a dozen startups are all still in the race.

For technology leaders, the action items today are boring but concrete: inventory cryptographic systems for quantum vulnerability, build some in-house quantum literacy, and run small experiments with hybrid quantum-classical workflows. Quantum will not replace classical computers; it will handle specific problem classes that classical machines cannot. The companies that start mapping where that boundary sits now—not in 2030—xill be the ones positioned to use it.