Quantum computers today operate as isolated machines inside sealed environments, with no practical way to connect to other systems as classical computers do over global networks. IBM and Cisco aim to change that. The two companies have outlined a long-term roadmap to link quantum processors across long distances, with an initial goal of demonstrating a workable quantum network before the end of 2030.
Classical computers exchange data easily over the internet, but an equivalent infrastructure for quantum systems does not yet exist. Building a quantum network is significantly more complex: it requires advances in hardware, software, control systems, and foundational research. Both IBM and Cisco say that collaboration with universities, federal laboratories, and research centers will be essential to advance the work.
The companies describe their effort as a potential foundation for a future “quantum internet.” Their partnership is not new—IBM and Cisco have collaborated since 1999—and Cisco contributes deep networking expertise that complements IBM’s quantum hardware and research capabilities.
Why a quantum network matters
Quantum computers offer the promise of solving problems in chemistry, physics, optimization, and cryptography that would take classical systems impractically long to complete. However, quantum devices are still early-stage and prone to errors. Firms including IBM and Google are racing to build systems that can operate reliably at scale. IBM has publicly targeted a fully error-corrected, production-capable quantum machine by 2029.
The joint roadmap calls for a proof-of-concept quantum network within three to five years, followed by efforts through the early 2030s to extend connectivity across more machines and longer distances. This work aligns with IBM’s broader quantum roadmap that lays out incremental steps toward larger, more capable systems.
IBM argues that a successful quantum network could eventually enable computations requiring up to trillions of quantum operations, unlocking use cases such as massive optimization tasks and the computer-aided design of new medicines and materials.
Turning stationary qubits into “flying” qubits
One fundamental challenge is converting the stationary quantum information stored in cryogenically cooled processors into mobile carriers that can traverse a network. IBM’s quantum processors operate inside large cryogenic systems where qubits are essentially stationary. To exchange quantum information between machines, those stationary qubits must be converted into “flying” qubits—quantum carriers that can move through space and across links. Jay Gambetta, director of IBM Research and an IBM fellow, describes flying qubits as qubits that physically travel rather than remain localized.
Demonstrating this conversion at small scale is only the first step. The partners must then prove they can scale beyond connecting two machines, ultimately linking many quantum computers at different locations over longer distances. Achieving that will require breakthroughs in microwave-to-optical transducers and optical-photon technologies that can faithfully convert and transport quantum states without degrading them.
Hardware and software for a quantum network
IBM is building a quantum networking unit (QNU) to serve as the interface between quantum processing units (QPUs). The QNU’s role is to translate stationary quantum states into flying quantum information that can be sent to other machines and to convert incoming flying qubits back into stationary states for processing.
On the software side, Cisco is developing a high-speed orchestration framework that can reconfigure the quantum network dynamically as computations proceed. The framework aims to deliver entanglement and coordinate state transfers among QNUs once each machine completes its portion of a distributed calculation, enabling the network to pass quantum information without interrupting the overall process.
Both companies are also exploring the design of a “network bridge” that would use multiple Cisco quantum network nodes to connect hundreds of IBM QPUs through QNU interfaces. The initial target is deployment within a single data center. To span longer distances, microwave qubits would need to be converted into optical signals able to travel through fiber-optic infrastructure and switching equipment. The enabling device is the microwave-optical transducer, a technology that today falls short of the fidelity and scalability required; IBM and Cisco plan to work with research partners—including the Superconducting Quantum Materials and Systems Center at Fermilab—to advance this area.
As the project advances, the partners intend to publish open-source software that integrates the system components and allows the research community to build on their work.
A joint vision for an integrated quantum system
“We are looking at this end-to-end as a system … rather than two discrete road maps,” said Vijoy Pandey, senior vice president of Cisco’s Outshift innovation incubator. “We are solving it jointly, which has a much better chance of this thing going in the same direction.”
Pandey added that IBM is pursuing aggressive scaling of quantum processors while Cisco focuses on networking technologies that enable scale-out. Together, they aim to address the full stack: hardware to connect quantum computers, software to orchestrate distributed quantum computations, and the networking intelligence to make those connections reliable and efficient.
If a robust architecture emerges, IBM’s quantum machines could one day handle workloads that outstrip what even the largest classical supercomputing clusters can achieve. IBM envisions networked quantum systems supporting runtimes involving trillions of quantum gates—the elementary operations of quantum algorithms—potentially enabling new approaches to drug discovery, material design, and other advanced scientific problems.
Significant uncertainties remain. The roadmap depends on a series of technological breakthroughs that may not arrive on a predictable schedule. Still, the companies believe a broad quantum internet composed of thousands of distributed systems is achievable by the late 2030s. Such a network could enable quantum sensing, long-distance quantum communications, and distributed quantum computing at global scales.
IBM also highlights practical near-term applications, such as quantum sensor networks capable of detecting subtle environmental changes that classical sensors might miss. Progress toward these goals will depend on refining today’s noisy, error-prone quantum processors into more reliable, error-managed systems.
The race toward fault-tolerant quantum machines
IBM has been transparent about its aim to build a fault-tolerant quantum computer—one that continues to operate correctly even as individual components experience errors. Recent demonstrations of running quantum error-correction algorithms on classical hardware provide useful insights into how errors can be managed while quantum systems mature.
Classical computing achieves practical fault tolerance through established design patterns that allow systems to continue operating despite component failures. Quantum systems face greater fragility: their qubits are highly sensitive to external disturbances and decoherence, which makes error control challenging. The objective is not to eliminate errors entirely, but to build machines that maintain correct operation despite them.
While specialized quantum startups often attract attention, IBM has been active in quantum research for decades. It now operates one of the largest fleets of quantum systems in the world, including more than 25 platforms with over 100 logical qubits across its portfolio. Its long-term ambition is to link these systems into a coherent network capable of solving problems well beyond the reach of today’s tools.
(Photo by JJ Ying)