Quantum Computing in 2025–2026

Breakthroughs, Players, and the Road to Fault Tolerance

February 20, 2026 · Research Report

Executive Summary

Quantum computing has entered a pivotal era. The period from late 2024 through early 2026 has seen a cascade of breakthroughs that collectively mark the transition from laboratory curiosity toward practical, fault-tolerant machines. Google's Willow chip became the first processor to operate "below threshold" for quantum error correction. Microsoft unveiled Majorana 1, the world's first topological qubit processor. AWS debuted its Ocelot chip based on cat qubits. Quantinuum launched a 98-qubit trapped-ion system with record fidelity. And neutral-atom platforms from QuEra and Atom Computing demonstrated fault-tolerant operations at scale.

Meanwhile, investment has surged — the industry raised nearly $4 billion in the first three quarters of 2025 alone, triple the total for all of 2024. Governments worldwide are committing tens of billions of dollars to quantum strategies, while the race to deploy post-quantum cryptography has accelerated with NIST finalizing its first PQC standards.

The consensus is shifting: fault-tolerant quantum computing is no longer a question of "if" but "when" — and the timeline now looks like the late 2020s to early 2030s for the first commercially useful fault-tolerant systems.

Hardware Breakthroughs

Superconducting Qubits

Superconducting qubits remain the most mature platform, with advances coming from every direction. Google's Willow chip (105 qubits, announced December 2024) demonstrated "below threshold" error correction for the first time — as the surface code grid scaled from 3×3 to 5×5 to 7×7 physical qubits, the logical error rate dropped exponentially. The 7×7 logical qubit lived twice as long as its best physical qubit and 20 times longer than previous results on Google's Sycamore. In October 2025, Google separately reported that its quantum computer was 13,000 times faster than the world's fastest classical supercomputer on a specific benchmark task.

IBM delivered its Nighthawk processor (120 qubits, 218 next-generation tunable couplers) in late 2025, achieving 30% more circuit complexity than prior Heron processors while maintaining low error rates. IBM also announced Quantum Loon, an experimental chip demonstrating all key processor components needed for fault-tolerant computing. The company's 2026 roadmap includes Kookaburra — the first processor module to store information in a qLDPC (quantum low-density parity-check) memory with an attached logical processing unit.

China's Zuchongzhi 3.2 (107 qubits) achieved quantum error correction below the fault-tolerance threshold using an all-microwave control approach — the first demonstration of this outside the United States and a technically distinct pathway from Google's approach. The commercial deployment of the Zuchongzhi 3.0 (105 qubits, 182 couplers) marked a notable milestone for China's quantum ambitions.

In April 2025, Fujitsu and RIKEN announced a 256-qubit superconducting quantum computer — four times larger than their 2023 system — with plans for a 1,000-qubit machine by 2026.

Topological Qubits

In February 2025, Microsoft unveiled Majorana 1, the world's first quantum processor built on topological qubits. Using a novel material class called "topoconductors" — devices combining indium arsenide (semiconductor) and aluminum (superconductor) — the chip hosts Majorana Zero Modes at the ends of topological superconducting nanowires. These qubits store information through electron parity, making them inherently more stable than conventional alternatives. Microsoft claims this architecture offers a clear route to one million qubits on a single chip, with custom error-correction codes reducing overhead roughly tenfold. DARPA selected Microsoft to advance to the final phase of its US2QC benchmarking program. The next step is an 8-qubit (4×2 tetron) array to demonstrate entanglement and quantum error detection on two logical qubits.

Cat Qubits

AWS launched Ocelot in February 2025, its first proprietary quantum chip developed at the AWS Center for Quantum Computing at Caltech. Built on "cat qubits" — named after Schrödinger's thought experiment — the chip provides inherent protection against bit-flip errors by encoding information in superpositions of coherent states within superconducting oscillators. The chip comprises 5 data qubits, 5 buffer circuits, and 4 ancilla qubits, achieving bit-flip times approaching one second (over 1,000 times longer than conventional superconducting qubits). AWS claims the design can reduce error-correction costs by up to 90%.

Trapped Ions

Quantinuum launched Helios in November 2025 — a 98-qubit trapped-ion system featuring record two-qubit gate fidelities of 99.921%, a real-time control engine, and integration with NVIDIA GB200 GPUs via NVQLink for real-time error correction. It demonstrated 48 error-corrected logical qubits at a 2:1 physical-to-logical ratio and set a quantum volume record of 33,554,432. Oxford Ionics separately achieved 99.99% two-qubit gate fidelity.

Neutral Atoms

Neutral-atom platforms emerged as a major force. Caltech scientists built a record-breaking 6,100-qubit neutral-atom array. QuEra and partners at Harvard, MIT, and Yale demonstrated a 3,000-qubit array operating continuously for over two hours, executed algorithms with up to 96 logical qubits, and published techniques reducing error-correction overhead by up to 100 times. QuEra raised over $230 million (led by Google Quantum AI, SoftBank, and NVIDIA) and delivered a quantum machine to Japan's AIST.

Microsoft and Atom Computing announced plans to deliver "Magne" — a 50-logical-qubit (1,200 physical qubit) error-corrected neutral-atom computer — to the Novo Nordisk Foundation in Denmark, expected operational by early 2027.

Platform Key Chip / System Qubits Headline Achievement
SuperconductingGoogle Willow105First "below threshold" error correction
SuperconductingIBM Nighthawk12030% more circuit complexity; path to 2026 advantage
SuperconductingZuchongzhi 3.2 (China)107Below-threshold QEC via all-microwave control
TopologicalMicrosoft Majorana 18First topological qubit processor; path to 1M qubits
Cat QubitsAWS Ocelot5 data90% reduction in QEC cost; ~1s bit-flip time
Trapped IonQuantinuum Helios9899.921% 2Q fidelity; 48 logical qubits at 2:1 ratio
Neutral AtomQuEra / Harvard3,000+96 logical qubits; 2+ hours continuous operation
Neutral AtomCaltech Array6,100Largest neutral-atom qubit array to date

Quantum Error Correction: The Defining Challenge

If 2024 was the year the quantum industry began to take error correction seriously, 2025 was the year it became the central battleground. Research output exploded — 120 peer-reviewed QEC papers were published in just the first 10 months of 2025, up from 36 in all of 2024.

The most significant milestone was Google's demonstration that error rates decrease exponentially as the surface code is scaled up — the long-sought "below threshold" result. This proved, after nearly 30 years of theory, that fault-tolerant quantum computing is physically achievable with current technology.

Multiple approaches to error correction are now being pursued in parallel:

Quantinuum demonstrated a remarkable 2:1 physical-to-logical qubit ratio (48 logical qubits from 98 physical), far more efficient than the typical hundreds-to-one ratio expected with surface codes alone. QuEra published techniques reducing error-correction overhead by up to 100 times.

The industry consensus, as articulated by Riverlane's 2025 report, is that error correction is now "the industry's defining challenge" — the single biggest barrier between today's noisy machines and tomorrow's useful quantum computers.

Major Players: Who's Leading the Race

Big Tech

Pure-Play Quantum Companies

China's National Champions

Chinese teams at the University of Science and Technology of China (USTC) have produced the Zuchongzhi (superconducting) and Jiuzhang (photonic) series of processors. The Zuchongzhi 3.0 is now commercially deployed, and Zuchongzhi 3.2 has matched and extended Google's below-threshold error correction results using a distinct technical approach. China's quantum investment, anchored by its $10B National Laboratory for Quantum Information Sciences, is the largest single-country commitment globally.

Software, Algorithms & Cloud Platforms

The quantum software ecosystem has matured significantly, with cloud access now available through five major platforms:

The dominant paradigm is hybrid quantum-classical computing, where quantum processors handle specific subroutines (optimization, simulation) while classical systems manage control, pre-processing, and post-processing. Strategic alliances between hardware makers, cloud providers, and application companies have created integrated platforms combining quantum processors with classical co-processing units.

Quantum-as-a-Service (QaaS) is democratizing access — researchers and enterprises can now experiment with quantum hardware without capital investment, accelerating adoption. Trapped-ion and photonics architectures have emerged as the dominant focus areas for new funding in the hardware layer.

On the algorithm front, IonQ's novel Clifford Noise Reduction (CliNR) technique has drastically reduced overhead for quantum error correction. Hybrid quantum-AI systems are increasingly being developed, with AI assisting in quantum error mitigation and circuit optimization.

Real-World Applications: From Lab to Market

While large-scale quantum advantage for general computation remains years away, several application domains have seen early quantum-classical advantages:

Drug Discovery & Life Sciences

Pharmaceutical company Roche reported in late 2025 that their quantum-powered molecular simulation platform identified three promising Alzheimer's drug candidates in 18 months — a process that typically takes 4–6 years. IonQ and Ansys achieved a milestone in March 2025 by running a medical device simulation on a 36-qubit computer that outperformed classical HPC by 12% — one of the first documented cases of quantum delivering practical advantage. Quantinuum's Helios successfully simulated high-temperature superconductivity and magnetism at unprecedented scales. Quantum computing is increasingly being used for protein hydration analysis and ligand-protein binding studies.

Finance

Goldman Sachs and JPMorgan have deployed quantum algorithms for portfolio optimization and risk analysis, exploring faster pricing models that could offer competitive advantages in the financial sector.

Materials Science & Chemistry

Quantum simulation of molecular systems continues to be the most natural application for quantum computers. IonQ has claimed to surpass classical methods in chemistry simulations, and hybrid quantum-AI approaches are being applied to materials discovery for batteries, catalysts, and superconductors.

Optimization

Logistics, supply chain, and scheduling optimization problems are active areas for quantum-classical hybrid approaches, though definitive advantage over classical solvers has not yet been broadly demonstrated. The expectation is that hybrid quantum-AI systems will begin to show impact in optimization, climate modeling, and similar domains in the near term.

Quantum Networking & Communication

Quantum networking is advancing on two fronts: connecting quantum computers together (distributed quantum computing) and securing communications via quantum key distribution (QKD).

Distributed Quantum Computing

The near-term driver for quantum networking is connecting multiple quantum processors within a data center. IonQ is pioneering this through photonic interconnects enabling entanglement between multiple QPUs — its acquisition of Qubitekk reflects this strategic priority. Industry experts note that 2026 qubit growth will be limited unless paired with network readiness: stable nodes, quantum memory, repeaters, and standards that allow multi-node systems to operate as a single machine.

Satellite QKD

Several satellite-based QKD initiatives are targeting 2026 launches:

Terrestrial Networks

Japan's NICT–NEC–Toshiba 2025 demonstration proved that high-rate quantum and classical signals can co-exist on the same carrier-grade fiber — an essential prerequisite for practical deployment. The University of Oxford announced a new project in February 2026 aimed at building the foundations of a quantum internet.

A note of caution: The U.S. National Security Agency (NSA) reiterated in 2025 that QKD should not be used for securing National Security Systems, citing limitations in range, integration complexity, hardware vulnerabilities, and cost. The path to quantum-secured communications remains the combination of QKD for specific use cases and post-quantum cryptographic algorithms for broad deployment.

Post-Quantum Cryptography

The race to quantum-proof our digital infrastructure has accelerated. In August 2024, NIST finalized its first three post-quantum cryptography (PQC) standards:

Additional standards are in the pipeline: HQC (Hamming Quasi-Cyclic) was selected for standardization in early 2025 to augment the key encapsulation portfolio, with finalization expected by 2026–2027. NIST has 14 digital signature candidates in Round 2.

The NSA's CNSA 2.0 compliance deadlines require all new National Security Systems to be quantum-safe by January 2027. The 2025–2026 period is focused on integrating PQC algorithms into secure elements and hardware security modules (HSMs), with hybrid implementations (classical + PQC) serving as the transitional approach.

NIST has warned that PQC algorithms are not "drop-in replacements" — organizations should expect different performance characteristics, higher computational costs, and new operational complexities during migration.

Policy & Investment

Global Investment Surge

The quantum industry raised $3.77 billion in equity funding in the first nine months of 2025 — nearly triple the $1.3 billion raised in all of 2024. Public funding now accounts for 34% of startup investment, up 19 percentage points from 2023. The total is projected to have exceeded $4.5 billion by year-end 2025.

United States

The Department of Energy Quantum Leadership Act of 2025 proposes $2.5 billion in quantum funding across FY2026–2030. The DOE announced $625 million for the next phase of National Quantum Information Science Research Centers. In January 2026, Senators Young and Cantwell introduced the National Quantum Initiative Reauthorization, authorizing $85 million per year through 2030 for NIST and $25 million for NASA quantum research. The White House is also drafting an executive order to reshape U.S. quantum policy.

European Union

In July 2025, the European Commission unveiled the Quantum Europe Strategy — a five-pillar roadmap (research, infrastructure, ecosystem, space/dual-use, skills) aiming to make Europe a "quantum industrial powerhouse" by 2030. The EU has invested over €11 billion in quantum over the past five years. The EuroQCI is set to become operational from 2026, and a call for evidence is open to shape the EU Quantum Act, scheduled for adoption in 2026.

China

China has made the largest single-country quantum investment globally, committing $10 billion to its National Laboratory for Quantum Information Sciences. Quantum technology is identified as a national strategic priority in China's new Five-Year Plan (2026–2030), aiming to make it "a new economic growth point."

Other Nations

Australia backed PsiQuantum with $620M AUD. Denmark (via the Novo Nordisk Foundation) is purchasing a Microsoft/Atom Computing quantum computer. Japan has received a QuEra fault-tolerance-ready system. Canada's QEYSSat program is advancing space-based QKD. The geopolitical competition continues to drive commitments across dozens of nations.

Key Milestones & Timeline

Date Milestone
Aug 2024NIST finalizes first three PQC standards (ML-KEM, ML-DSA, SLH-DSA)
Dec 2024Google Willow achieves first below-threshold quantum error correction
Feb 2025Microsoft unveils Majorana 1 topological qubit processor
Feb 2025AWS launches Ocelot cat-qubit chip
Mar 2025IonQ/Ansys demonstrate 12% quantum advantage in medical simulation
Mar 2025China's Zuchongzhi 3.2 achieves below-threshold QEC via all-microwave control
Mar 2025PsiQuantum secures $750M (including $620M AUD Australian government)
Apr 2025Fujitsu/RIKEN announce 256-qubit superconducting computer
Jul 2025EU unveils Quantum Europe Strategy
Sep 2025Quantinuum raises $600M at $10B valuation; sets QV record of 33.5M
Oct 2025Google demonstrates 13,000× speedup over classical supercomputers
Nov 2025IBM delivers Nighthawk (120 qubits) and Loon (fault-tolerance validation)
Nov 2025Quantinuum launches Helios (98-qubit trapped-ion, 99.921% fidelity)
Nov 2025DOE announces $625M for quantum research centers
Jan 2026National Quantum Initiative Reauthorization introduced in U.S. Senate
Feb 2026White House drafts executive order on quantum policy

What's Expected Next

Where the Field Is Heading

The quantum computing landscape in early 2026 is characterized by a striking diversity of approaches and an unprecedented pace of progress. Five distinct qubit technologies — superconducting, trapped ion, neutral atom, topological, and cat qubits — are all advancing simultaneously, each with credible roadmaps to fault tolerance. This "let a thousand flowers bloom" phase is healthy for the field, though consolidation will likely follow as some approaches prove more scalable or cost-effective than others.

The clearest near-term story is the convergence of error correction techniques. The theoretical and experimental breakthroughs of 2025 — below-threshold operation, qLDPC codes, 100× overhead reduction — have shifted the conversation from "can we correct quantum errors?" to "how efficiently can we do it?" This is fundamental progress.

Commercially, the field remains in a "pre-revenue" phase for most applications. The early quantum advantages demonstrated (drug discovery, chemistry simulation) are real but narrow. The hybrid quantum-classical model will dominate for the foreseeable future, with quantum processors serving as specialized accelerators for specific subroutines rather than general-purpose replacements for classical computers.

The investment picture is robust but not without risk. With $4.5+ billion raised in 2025 and massive government commitments, the industry has the capital to pursue its ambitious roadmaps. However, the gap between laboratory demonstrations and commercially useful applications remains significant, and the timeline to fault tolerance — while narrowing — is still measured in years, not months.

The bottom line: quantum computing is no longer a speculative technology. The physics works. The engineering is advancing rapidly. The question that remains is whether the practical, fault-tolerant machines will arrive by the end of this decade — as IBM, Google, and others are betting — or whether it will take longer. The breakthroughs of 2025 provide strong reason for optimism.

Sources