IBM has officially unveiled its roadmap for building what will be the world’s first large-scale fault-tolerant quantum computer. This ambitious project, named IBM Quantum Starling, will be developed at the company’s new quantum data center in Poughkeepsie, New York, and is expected to be operational by 2029.
The magnitude of this technological leap is such that, according to IBM, representing the complete computational state of Starling would require the memory of more than a "quindecillion" (10⁴⁸) of the most powerful supercomputers on the planet. Starling will be able to perform 20,000 times more operations than current quantum computers, opening the door to exploring the full complexity of quantum states, far beyond the capabilities of existing systems.
A Milestone in the Quantum Race
IBM’s announcement marks a key stage in the transition of quantum computing from laboratory experiments to practical, scalable systems with real-world industry impact. According to Arvind Krishna, IBM’s Chairman and CEO, “our cross-disciplinary knowledge in mathematics, physics, and engineering enables us to pave the way for a large-scale fault-tolerant quantum computer capable of solving real-world challenges and unlocking immense possibilities for businesses.”
Such systems will allow, for example, the execution of hundreds of millions or billions of quantum operations, accelerating processes in pharmaceutical development, material discovery, computational chemistry, and industrial optimization.
From Physical to Logical Qubits: The Challenge of Fault Tolerance
At the heart of IBM’s announced advance is the architecture of logical qubits. A “logical qubit” is a quantum information unit corrected for errors, constructed from several physical qubits that work together. The significant challenge in quantum computing has always been the instability of physical qubits and the accumulation of errors; fault tolerance requires suppressing these errors exponentially as the size of the cluster of physical qubits increases.
Until now, error correction codes needed such a vast number of physical qubits that they were unfeasible for large systems. However, IBM announces it will utilize qLDPC (Quantum Low-Density Parity Check) codes that reduce the number of physical qubits needed by up to 90% compared to other leading methods. This paves the way for much more efficient and scalable systems.
IBM Quantum Roadmap: From Laboratory to Industry
IBM’s new roadmap outlines the technological milestones for achieving large-scale fault tolerance:
IBM Quantum Loon (2025): Will test architectural components for qLDPC codes, including long-range couplers within the chip.
IBM Quantum Kookaburra (2026): The first modular processor from IBM designed to store and process encoded information, combining quantum memory and logic.
- IBM Quantum Cockatoo (2027): Will enable the entanglement of two Kookaburra modules, unifying several quantum chips as nodes in a larger system.
All of this will culminate in Starling in 2029, which will be capable of executing 100 million quantum operations with 200 logical qubits. Subsequently, IBM Quantum Blue Jay will extend the architecture to 2,000 logical qubits and 1 billion operations.
Impact on Industry and the Future of Quantum Computing
The development of fault-tolerant and scalable quantum computers promises to transform sectors such as healthcare, chemistry, banking, and logistics, addressing intractable problems for classical computers.
IBM emphasizes that its modular and efficient approach is key to avoiding massive infrastructures and prohibitive costs, a challenge that has so far slowed industry progress. With Starling, IBM aims to position itself at the forefront of the global quantum race, establishing an unprecedented technological and infrastructural standard.
This announcement also solidifies IBM’s commitment to quantum computing as a driver of industrial innovation for the next decade, signaling a clear transition from experimentation to mass commercial adoption.