The history of practical quantum computing begins not in a university laboratory or a government research facility but in a startup founded in 1999 in Vancouver, British Columbia. D-Wave Systems, co-founded by physicist Geordie Rose, set out to build quantum computers using an approach that most of the academic quantum computing community considered a dead end: quantum annealing. While the mainstream research community focused on gate-based quantum computing — building universal quantum processors from individually controlled qubits connected by logic gates — D-Wave pursued a fundamentally different architecture. Quantum annealing exploits quantum tunneling and superposition to find the lowest-energy state of a complex optimization problem, encoding the problem into the energy landscape of a network of superconducting qubits and then allowing the system to evolve toward its ground state. It is not universal quantum computing. It cannot run arbitrary quantum algorithms. But it can, in principle, solve certain optimization problems faster than any classical computer.
D-Wave shipped its first commercial system, the D-Wave One, in 2011. It contained 128 superconducting flux qubits and was purchased by Lockheed Martin for approximately $10 million. The sale was historic — the first commercial quantum computer transaction ever completed — but it was also deeply controversial. Physicists debated whether the D-Wave One was genuinely performing quantum computation or merely classical thermal annealing at low temperatures. A series of benchmarking studies published between 2013 and 2015 showed mixed results: on some problems, D-Wave systems showed speedups consistent with quantum behavior; on others, optimized classical algorithms matched or beat them. Google purchased a D-Wave Two in 2013 for its Quantum AI lab at NASA Ames, using it to explore quantum machine learning before ultimately deciding to build its own gate-based processors. Today, D-Wave operates the Advantage system, featuring over 5,000 qubits connected in a Pegasus topology with 15-way connectivity. The company went public via SPAC merger in 2022 and continues to market its systems for optimization problems in logistics, financial modeling, and materials science. But the quantum computing world has largely moved past D-Wave’s approach, and the architectural decisions that made D-Wave first to market are the same ones that limit its long-term relevance.
The gate-based paradigm that now dominates the field operates on fundamentally different principles. In a gate-based quantum computer, individual qubits are initialized, manipulated through sequences of quantum logic gates (analogous to the AND, OR, and NOT gates of classical computing, but operating on superpositions), and measured to extract results. The power of gate-based systems comes from their universality: given enough qubits and low enough error rates, they can implement any quantum algorithm, from Shor’s algorithm for integer factorization to Grover’s algorithm for database search to variational algorithms for simulating molecular chemistry. The challenge is that qubits are extraordinarily fragile. Every interaction with the environment — stray electromagnetic radiation, thermal vibrations, impurities in the fabrication materials — can cause decoherence, destroying the quantum information before it can be processed. This is why error correction is the central challenge of quantum computing, and why the field has fractured into competing hardware approaches, each offering different trade-offs between qubit count, gate fidelity, connectivity, and coherence time.
IBM has pursued superconducting transmon qubits with a consistency and transparency unmatched in the industry. The company published its quantum computing roadmap in 2020 and has updated it annually, providing specific qubit counts, gate fidelity targets, and architectural milestones for each year through 2033. The progression has been methodical: the 27-qubit Falcon in 2019, the 65-qubit Hummingbird in 2020, the 127-qubit Eagle in 2021, the 133-qubit Heron in late 2024, with plans for the modular 100,000-qubit Starling system that would connect multiple quantum processors through quantum interconnects. IBM’s Qiskit software framework is the most widely used quantum programming platform in the world, and its IBM Quantum Network connects more than 200 organizations — including national laboratories, universities, and Fortune 500 companies — to quantum hardware through the cloud. IBM’s approach emphasizes middleware and software as much as hardware, reflecting its long corporate history of providing enterprise computing solutions rather than raw processing power.
Google’s approach has been more hardware-focused and more dramatic in its milestones. The Sycamore chip in 2019 and the Willow chip in 2024 represent the two most widely publicized quantum computing achievements in history. Google’s Santa Barbara lab fabricates its own superconducting processors and has invested heavily in the cryogenic engineering required to operate large-scale quantum systems at millikelvin temperatures. Microsoft took the most scientifically ambitious and commercially risky path of all, betting on topological qubits that did not exist in any laboratory when the program began. The Majorana 1 chip, announced in February 2025, vindicates that two-decade bet — or at least begins to. IonQ and Quantinuum have pursued trapped-ion architectures, using individual atoms suspended in electromagnetic fields as qubits. Trapped ions offer the highest gate fidelities in the industry and naturally all-to-all connectivity, meaning any qubit can interact directly with any other qubit without the routing constraints that plague superconducting systems. The trade-off is speed: ion trap gate operations are slower than superconducting gates, and scaling trapped-ion systems to thousands of qubits requires engineering solutions that remain in development.
The national security implications of quantum computing have been understood since before any quantum computer existed. In 1994, mathematician Peter Shor published an algorithm showing that a sufficiently large quantum computer could factor large integers exponentially faster than any known classical algorithm. This result struck at the foundation of modern cybersecurity. RSA encryption, the most widely used public-key cryptosystem in the world, derives its security from the practical impossibility of factoring the product of two large prime numbers. A quantum computer running Shor’s algorithm with enough error-corrected qubits could break RSA-2048 in hours or days. The same vulnerability applies to elliptic curve cryptography and the Diffie-Hellman key exchange protocol, which together secure virtually all internet communications, financial transactions, and classified government networks. The NSA acknowledged this threat publicly in 2015, when it announced plans to transition to quantum-resistant cryptographic algorithms.
In July 2024, the National Institute of Standards and Technology (NIST) finalized the first three post-quantum cryptographic standards after an eight-year evaluation process that began in 2016 with 82 candidate algorithms submitted by research teams worldwide. The three selected algorithms are CRYSTALS-Kyber for key encapsulation (renamed ML-KEM), CRYSTALS-Dilithium for digital signatures (renamed ML-DSA), and SPHINCS+ for hash-based signatures (renamed SLH-DSA). A fourth algorithm, FALCON (renamed FN-DSA), was selected as an additional digital signature standard. These algorithms are based on mathematical problems — lattice problems and hash functions — believed to be resistant to both classical and quantum attacks. The NSA followed NIST’s announcement with its Commercial National Security Algorithm Suite 2.0 (CNSA 2.0), mandating the transition of all National Security Systems to post-quantum algorithms by 2035. The timeline is aggressive by government standards, reflecting the urgency of the threat. Intelligence agencies operate on the assumption that adversaries are already harvesting encrypted communications today with the intention of decrypting them once quantum computers become powerful enough — a strategy known as “harvest now, decrypt later.”
China has pursued quantum technology with a focus and investment level that matches or exceeds the United States. In 2016, China launched Micius, the world’s first quantum communication satellite, named after the ancient Chinese philosopher. Micius demonstrated quantum key distribution (QKD) between ground stations separated by over 1,200 kilometers, using entangled photons transmitted from orbit to establish encryption keys guaranteed secure by the laws of physics. In 2017, China completed the Beijing-Shanghai quantum communication backbone, a 2,000-kilometer fiber-optic network connecting financial institutions and government agencies in four cities through QKD nodes. Pan Jianwei, the physicist leading China’s quantum program at the University of Science and Technology of China in Hefei, has been called the “father of quantum” in Chinese media. His team claimed quantum computational advantage in 2020 with Jiuzhang, a photonic quantum computer that performed Gaussian boson sampling faster than classical supercomputers. In 2021, the team announced Zuchongzhi 2, a 66-qubit superconducting processor that exceeded Sycamore’s performance on random circuit sampling benchmarks. China’s investments in quantum computing, quantum communication, and quantum sensing are estimated at $15 billion, dwarfing the approximately $3.7 billion authorized by the U.S. National Quantum Initiative Act of 2018.
Quantum sensing — the use of quantum effects to measure physical quantities with extreme precision — may ultimately have more immediate national security impact than quantum computing. Quantum magnetometers based on nitrogen-vacancy centers in diamond can detect magnetic fields with sensitivities approaching femtotesla, potentially enabling detection of submarines by their magnetic signatures from distances and depths that render current magnetic anomaly detection obsolete. Quantum gravimeters can measure variations in the local gravitational field with precision sufficient to detect underground tunnels, bunkers, and installations from the surface or from aircraft. Quantum inertial navigation systems, using atom interferometry to measure acceleration and rotation without GPS, could provide navigation for submarines, missiles, and autonomous vehicles in GPS-denied environments. These technologies are not theoretical — they are in advanced development at defense laboratories in the United States, China, and the United Kingdom. The Army Research Laboratory, the Naval Research Laboratory, and DARPA have all funded quantum sensing programs. The implications for the detection and monitoring of underground military installations — or, conversely, for navigating within them without external signals — are significant.
The quantum arms race between the United States and China extends beyond computing and sensing into the fundamental question of who controls the infrastructure through which quantum capabilities are deployed. In the United States, that infrastructure is being built by the JWCC cloud vendors — Microsoft, Amazon, Google, and Oracle — with quantum hardware from their own labs and from companies like IonQ, Rigetti, and Quantinuum. In China, it is being built by state-directed enterprises with direct connections to the People’s Liberation Army. The playing field is not level: China’s fusion of military, academic, and commercial quantum development under state direction gives it coordination advantages that the market-driven U.S. approach cannot match. But the U.S. has its own form of fusion — the JWCC contract structure that binds the world’s most capable technology companies to the Department of Defense through multi-billion-dollar agreements, channeling their quantum computing advances directly into classified military infrastructure.
From D-Wave’s first 128-qubit annealer in 2011 to Microsoft’s eight topological qubits in 2025, quantum computing has traversed an arc that is both shorter and longer than most people appreciate. Shorter, because the fundamental hardware remains primitive by the standards required for the most feared applications — Shor’s algorithm against RSA-2048 would require an estimated four million physical qubits with current error rates, and no machine on Earth has more than a few hundred. Longer, because the supporting infrastructure — the cloud platforms, the classified networks, the government contracts, the post-quantum cryptography standards, the quantum sensing capabilities — is already in place, waiting for the processors to catch up. The arc from D-Wave to Majorana is not just a story of scientific progress. It is a story of infrastructure being built, contracts being signed, and capabilities being positioned inside the most sensitive installations of the world’s most powerful military. The underground network documented in this investigation does not exist in isolation from these technological developments. It is connected to them — through cloud contracts, through classified networks, through fiber-optic backbones and air-gapped server rooms that run on the same technology described in this article. Following the evolution of quantum computing is following the evolution of the infrastructure that serves that network. The full documentation of those connections is available throughout the investigation published on this site.