IBM has unveiled its next-generation quantum processor, the IBM Nighthawk quantum chip. The new Nighthawk r3, announced on Wednesday, is designed to accurately execute 10,000 quantum operations across 520 programmable qubits, and IBM plans to link two of them into a single system with 1,040 programmable qubits in 2027, its first machine past the 1,000-qubit mark and a major step toward fault-tolerant quantum computing.

The announcement, made on November 11, 2026, in Yorktown Heights, New York, sets out IBM's clearest timetable yet for scaling quantum computers beyond a single chip. The processor carries 2,034 total physical qubits and uses 10 chip-to-chip interconnects, called l-couplers, to connect processors into larger systems.

Key Points

• IBM Quantum Nighthawk r3 is designed to accurately execute 10,000 quantum operations across 520 programmable qubits, with 2,034 physical qubits.
• In 2027, IBM plans to link two Nighthawk r3 processors into one 1,040-qubit system, its first machine past the 1,000-qubit mark.
• The chip uses 10 interconnects called l-couplers; IBM has linked Nighthawk r1 processors with four l-couplers at 99.3% state-transfer fidelity, short of its 99.9% target.
• The predecessor, Nighthawk r2, introduced in September 2026 and available now through the cloud, has run 34 billion circuits, with speedups of up to 19 times.
• The r3 figures are design targets, not measured performance; IBM disclosed no release date and no benchmarks for the completed architecture.

What Happened

IBM announced the Nighthawk r3 on Wednesday, November 11, 2026, as the next step in a modular approach to building bigger quantum computers. Rather than fitting ever more qubits onto one chip, IBM is designing processors that can be linked together, as classical supercomputers are.

The r3 keeps the square-lattice connectivity and fast operations of the current Nighthawk r2 while adding the 10 l-couplers that make linking possible. IBM has already demonstrated two key technologies: this year it connected Nighthawk r1 processors using four l-couplers, reaching 99.3% state-transfer fidelity toward a 99.9% target, and it connected two cryogenic modules into a single system for the first time, cooling it below 15 millikelvin.

Background and Context

Quantum computing has long promised breakthroughs in drug discovery, materials science, finance and logistics, but practical use has been held back: quantum machines are error-prone, and they are slow at the scale real problems demand. Each run of a quantum circuit returns a single sample, and useful applications increasingly need circuits executing thousands of operations on about 100 qubits or more. Scale and quality make bigger computations possible, but speed decides whether they can be finished in a practical time.

IBM's answer has been to improve all three at once. The Nighthawk r2, introduced in September 2026, delivers up to 100,000 samples per second and supports accurate computations with more than 7,500 quantum operations on 120 qubits. Clients and partners have executed 34 billion circuits on Nighthawk r2, reporting improvements of up to 19 times, and E.ON reported a 15-times speedup in quantum machine learning for energy grid optimisation.

IBM Nighthawk Quantum Chip: How the 1,040-Qubit System Will Work

The 1,040-qubit system planned for 2027 will join two Nighthawk r3 processors through their l-couplers, the chip-to-chip interconnects at the heart of the new design. Each r3 carries 10 couplers, giving the pair multiple paths to share quantum states. The test so far, linking r1 processors with four couplers at 99.3% state-transfer fidelity, shows the approach works but has not yet reached the 99.9% fidelity IBM wants, and that gap is one of the main hurdles still to clear.

Modularity matters because a single quantum chip can only grow so far before defects and noise make it impractical. By linking smaller, well-tested processors, IBM can keep adding qubits without betting everything on one giant chip. The approach also extends to cooling: connecting two cryogenic modules into one system, as IBM has now done, is a necessary step toward housing multi-chip machines in the extreme cold these processors require.

Reactions and Quotes

Jay Gambetta, director of IBM Research, said the test for the technology is whether it can deliver results accurately and quickly. "Useful quantum computing hinges on whether results can be delivered with accuracy and speed, both are fundamental," Gambetta said. "Nighthawk is already showing us what that looks like today. As we scale this architecture to larger and more powerful quantum computers, the capabilities of quantum computation will reach new levels as we integrate more closely with classical resources and march ahead to fault tolerance."

Why It Matters

The race to useful quantum computing is also a race between very different engineering bets, and IBM's modular plan is now its clearest statement of how it intends to win it. Rivals have pursued larger single chips, different qubit technologies and cloud-first strategies, but all face the same wall: without error correction at scale, quantum machines stay confined to demonstrations. A 1,040-qubit system that holds its speed and quality would be a major step toward the thousands of reliable qubits practical applications are expected to need.

The announcement also lands in a turbulent week for the chip industry. Nvidia's board is probing chief executive Jensen Huang over China spy-plot claims, and Nvidia shares plunged 43%, wiping $192 billion, as the FBI probes the Huang spy claims. For the United States, the stakes are strategic as well as commercial, and IBM is positioning Nighthawk as proof that the US can keep scaling the technology at home.

What Happens Next

The immediate test is the 2027 system. IBM must show that two r3 processors can be linked into a single 1,040-qubit machine that keeps the speed and quality of the current generation, and close the gap between the 99.3% state-transfer fidelity demonstrated so far and its 99.9% target. Until the hardware exists, the 10,000-operation and 1,040-qubit figures remain design goals.

Beyond the chip, IBM is betting on closer integration of quantum and classical computing. Its enhanced IBM Quantum Platform now brings high-performance classical and quantum computing together in one managed environment, and IBM recently launched a co-design platform for quantum error correction with Oak Ridge National Laboratory and AMD. The longer horizon is Starling, IBM's planned fault-tolerant system targeted for 2029, and whether IBM hits its 2027 date will be the first test of that timetable.

Frequently Asked Questions

What is IBM Quantum Nighthawk r3?

IBM's next-generation quantum processor, announced on November 11, 2026, designed for 10,000 operations across 520 programmable qubits, with 2,034 total physical qubits.

How many qubits will the 2027 system have?

IBM plans to link two Nighthawk r3 processors into one system with 1,040 programmable qubits in 2027, its first machine past the 1,000-qubit mark.

What are l-couplers?

Chip-to-chip interconnects that let processors share quantum states, so several chips can work as one larger system. The Nighthawk r3 has 10 of them.

How fast is the current Nighthawk r2?

Introduced in September 2026, Nighthawk r2 delivers up to 100,000 samples per second and supports more than 7,500 quantum operations on 120 qubits. Clients and partners have run 34 billion circuits on it.

When will Nighthawk r3 be available?

IBM has not disclosed a release date for r3, or benchmarks for the completed architecture. The 2027 date refers to the planned 1,040-qubit system built from two r3 processors.