Quantum Computing Modalities: Majorana Qubits

(For the umbrella view of topological quantum computing, see Topological Quantum Computing. For other modalities, see Taxonomy of Quantum Computing: Modalities & Architectures.)

What It Is

A Majorana qubit encodes quantum information in Majorana zero modes (MZMs): quasiparticles that appear at the boundaries of topological superconductors and are, in a mathematical sense, their own antiparticles. Two spatially separated MZMs together form one fermionic mode, which can be occupied (|1⟩) or unoccupied (|0⟩). That occupation state is the qubit. Because the quantum information is stored non-locally across two points that may be micrometers apart, local perturbations at either site cannot determine or flip the encoded state. Only a process that connects both sites (braiding one MZM around another, or bringing them together to “fuse”) can change the qubit.

This non-local encoding is the source of the theoretical appeal. In every other qubit modality, the quantum state is a local property of a physical system: the spin of an electron (silicon spin), the energy level of a circuit (superconducting), the internal state of an atom (trapped ion, neutral atom). Local noise can flip any of these. A Majorana qubit’s information is not located at any single point, which makes it topologically protected against local noise. The protection is not perfect (non-local errors, thermal excitations above the topological gap, and quasiparticle poisoning can still cause errors), but the theoretical error rate could be orders of magnitude lower than any conventional qubit.

The catch, as I detail in the topological quantum computing umbrella article: no Majorana qubit has been demonstrated. The physics is theoretically sound. The experimental realization has been plagued by ambiguous signatures, contested claims, a high-profile retraction, and a persistent gap between what press releases announce and what peer-reviewed papers actually show.

How Majorana Zero Modes Form

The Kitaev Chain

The simplest model for understanding MZMs is Alexei Kitaev’s 1D chain (2001). Consider a one-dimensional wire of spinless fermions with p-wave superconducting pairing. At each site, you can decompose the electron creation and annihilation operators into two Majorana operators (γ_A and γ_B). In the topologically trivial phase, these Majorana operators pair up on each site: γ_A at site 1 pairs with γ_B at site 1, forming a regular fermion. But in the topological phase, the pairing shifts: γ_B at site 1 pairs with γ_A at site 2, γ_B at site 2 pairs with γ_A at site 3, and so on. This leaves γ_A at the first site and γ_B at the last site unpaired. These two dangling Majorana operators are the MZMs. They sit at zero energy (hence “zero modes”), they are spatially separated at opposite ends of the chain, and together they encode one fermionic degree of freedom that can be in one of two states.

The Kitaev chain is a toy model. Real materials do not have spinless p-wave superconductivity. The experimental challenge has been to engineer a physical system that effectively reproduces the Kitaev chain’s topological phase using available materials.

The Semiconductor-Superconductor Heterostructure

The breakthrough proposal came from Lutchyn et al. and Oreg et al. (2010), who showed that a semiconductor nanowire with three ingredients can mimic the Kitaev chain:

1. Strong spin-orbit coupling (the electron’s spin couples to its momentum, creating an effective spinless regime at certain momenta). Indium arsenide (InAs) and indium antimonide (InSb) nanowires provide this.
2. Proximity-induced superconductivity (a superconductor, typically aluminum, is deposited on or grown adjacent to the nanowire, inducing Cooper pairing in the semiconductor through the proximity effect).
3. An applied magnetic field (Zeeman splitting breaks time-reversal symmetry, driving the system into a topological phase when the field exceeds a critical value).

When all three conditions are met simultaneously, the theory predicts that MZMs appear at the ends of the proximitized wire segment. The experimental signature: a peak in tunneling conductance at exactly zero bias voltage, quantized at 2e²/h, that persists over a range of magnetic field and gate voltage. This zero-bias conductance peak (ZBCP) became the most-sought experimental signature in condensed matter physics for a decade.

The Topoconductor

Microsoft‘s current approach replaces the nanowire geometry with a planar InAs/Al heterostructure that the company calls a “topoconductor.” Instead of a thin wire with MZMs at its ends, the topoconductor is a 2D platform where gate electrodes define the topological region, and interferometric measurements probe the states at its boundaries. The topoconductor architecture is designed for scalability: multiple qubit units (tetrons, described below) can be patterned on the same chip using lithographic techniques, whereas nanowire devices require manual assembly of individual wires.

The Tetron: Four Majoranas, One Qubit

Microsoft’s qubit design, the tetron, uses four MZMs arranged in a specific geometry. Two pairs of MZMs form two fermionic modes. The joint parity of these two modes (whether the total fermion number is even or odd) is fixed by the superconducting ground state. Within that fixed-parity subspace, the occupation of the individual modes defines the qubit’s |0⟩ and |1⟩ states. Readout is performed by measuring the parity of specific loops in the tetron using quantum-dot charge sensors.

The tetron architecture is what Microsoft’s Majorana 1 chip implements. The chip has the geometry to host 8 tetron qubit units on a single die.

The Experimental History: Claims, Contests, and a Retraction

The history of Majorana experiments is a case study in the difficulty of distinguishing real topological states from mundane alternatives that produce similar signatures. Understanding this history is essential for assessing where the field stands today.

2012: The First Claim (Mourik et al., Delft)

Leo Kouwenhoven’s group at Delft observed a zero-bias conductance peak (ZBCP) in InSb nanowires proximitized with NbTiN superconductor. The peak appeared at finite magnetic field, consistent with the predicted topological phase transition. Published in Science, it was widely reported as the discovery of Majorana fermions in solid state.

The problem: a ZBCP can also be produced by non-topological mechanisms. Andreev bound states (ABS) in a disordered wire with smooth confinement potentials can mimic the Majorana signature. The ZBCP observed by Mourik et al. was not quantized at 2e²/h, and it appeared over a limited range of parameters. Subsequent theoretical and experimental work showed that trivial ZBCPs from disorder-induced ABS are common in these devices and difficult to distinguish from MZMs using tunneling conductance alone.

2014–2017: Additional Claims and Mounting Skepticism

Multiple groups reported ZBCPs in various material platforms: iron chains on superconducting lead (Nadj-Perge et al., Princeton, 2014), InAs nanowires (several groups), and vortex cores in Fe-based superconductors. Each observation generated excitement, followed by analysis showing that non-topological explanations could not be ruled out.

The Albrecht et al. (2016) experiment from the Copenhagen group (Microsoft-affiliated) reported quantized conductance in InAs/Al nanowire devices, which was initially taken as stronger evidence than a bare ZBCP. But the quantization was observed in only a fraction of devices and under specific conditions, and its interpretation remained debated.

2018: The Quantized Conductance Claim (and its 2021 Retraction)

A Microsoft-affiliated team at Delft published a paper in Nature (Zhang et al., 2018) claiming stable, quantized Majorana conductance in InSb/Al nanowires. The paper was high-profile and widely cited as near-definitive evidence for MZMs.

In 2021, the paper was retracted from Nature after Sergey Frolov (University of Pittsburgh) and others identified issues with the data analysis. Specific concerns included selective data presentation and undisclosed processing of the raw conductance data. The retraction was a watershed moment for the field: it raised the evidentiary bar for all subsequent claims and generated a credibility deficit that Microsoft’s program has been working to overcome since.

2022: Microsoft’s “Historic Milestone” (Without Peer Review)

In 2022, Microsoft announced that its Azure Quantum team had “engineered a topological qubit” and achieved a “historic milestone.” No peer-reviewed paper accompanied the announcement. The quantum physics community largely withheld judgment, and several prominent physicists publicly noted the absence of published data.

February 2025: Majorana 1

Microsoft unveiled Majorana 1, an 8-qubit-capacity test chip on the InAs/Al topoconductor platform. A companion paper was published in Nature with over 160 co-authors. The press release described it as the “first quantum processor powered by topological qubits” and claimed a path to one million qubits on a single chip.

The Nature paper told a different story. It described a device architecture that could host MZMs and presented interferometric measurements of low-energy states. The Nature editorial team attached an explicit disclaimer: the results “do not represent evidence for the presence of Majorana zero modes in the reported devices” and the paper was published for introducing “a device architecture that might enable fusion experiments using future Majorana zero modes.”

At APS March Meeting 2025, additional data was presented by Microsoft researchers. The community response was mixed. Sergey Frolov raised specific objections about the interpretation of the interferometric data. Patrick Lee (MIT) and Henry Legg questioned whether the observed signatures were uniquely topological.

My analysis at the time concluded that the gap between the press release and the peer-reviewed evidence was substantial, and that Majorana 1 should be understood as a test platform, not a functioning topological quantum processor.

July 2025: Parity Measurement Follow-Up

Microsoft published a follow-up preprint (“Distinct Lifetimes for X and Z Loop Measurements in a Majorana Tetron Device”) presenting the first complementary Pauli-X and Pauli-Z parity measurements on a tetron device. This was a more concrete experimental advance than the February Nature paper: it demonstrated the two measurement types needed to operate a Majorana qubit, with measured parity lifetimes and readout fidelities.

This result moved the needle. It showed that the tetron architecture can perform the basic measurement operations required for qubit control. Whether the states being measured are topological MZMs or trivial ABS remains debated, but the operational capability (measuring both X and Z parity on the same device) is a prerequisite for any subsequent braiding or gate demonstration.

July 2025: The 1/f Noise Concern

An Australian preprint (July 2025, arXiv, not yet peer-reviewed) analyzed 1/f charge noise decoherence in Majorana-based qubits and raised a fundamental concern: the semiconductor-superconductor interface that hosts MZMs also generates charge noise that could decohere the topological states faster than the topological gap protects them. If this analysis holds, the theoretical advantage of topological protection may be significantly reduced in practice, because the noise source is intrinsic to the material platform rather than an external perturbation that can be shielded away.

Why the Evidence Remains Contested

The core problem is that every experimental signature claimed as evidence for MZMs can also be produced by non-topological mechanisms:

Zero-bias conductance peaks can arise from trivial Andreev bound states in disordered nanowires. The disorder that is endemic to semiconductor-superconductor interfaces creates smooth potential variations that trap electrons at near-zero energy, mimicking the Majorana ZBCP.

Quantized conductance (2e²/h) was predicted as a smoking-gun signature, but the 2018 claim of this quantization was retracted. Non-topological near-zero-energy states can produce approximately quantized conductance under certain conditions.

Interferometric phase measurements (used in the Majorana 1 Nature paper) detect low-energy states but, as the Nature reviewers noted, cannot by themselves determine whether those states are topological.

Parity measurements (the July 2025 follow-up) demonstrate operational capability but do not prove that the states being measured are MZMs rather than trivial states.

The field needs a demonstration that is uniquely topological: an experiment whose outcome can only be explained by non-Abelian anyon physics, not by any combination of trivial effects. The most widely proposed such experiment is a braiding operation: physically exchanging two MZMs and measuring a state change that matches the predicted non-Abelian transformation. This has not been done.

Comparison to Other Modalities

The comparison is stark and can be stated briefly. Every other modality covered in this series has demonstrated working qubits, controllable gates, and (for the leading platforms) error-corrected logical qubits:

• Superconducting: 1,121 physical qubits, 99.97% gate fidelity, below-threshold QEC
• Trapped ion: 98 physical → 48 logical qubits, 99.921% two-qubit fidelity
• Neutral atom: 96 logical qubits from 448 atoms, room-temperature operation
• Silicon spin: >99% fidelity on 300 mm foundry wafers, first logical operations
• Photonic: 99.98% SPAM, fab-manufactured at GlobalFoundries

Majorana qubits have demonstrated: parity measurements on a tetron device consistent with (but not proven to be) topological states. No qubit, no gate, no braiding, no error correction.

The theoretical value proposition remains powerful: if Majorana qubits achieve ~10⁻⁶ hardware error rates as theorized, a few thousand topological qubits could match what other modalities need millions of physical qubits to accomplish. But “if” is carrying extraordinary weight. The other modalities are delivering results now. Majorana qubits are delivering promises and preliminary device architectures.

Advantages (Theoretical)

These advantages have not been experimentally demonstrated. I label them “theoretical” because none exists outside of mathematical models and theoretical physics papers.

Topological error protection. The defining advantage. Local noise cannot flip a qubit whose state is encoded non-locally across two spatially separated MZMs. The error rate scales exponentially with the ratio of MZM separation to the coherence length of the superconductor: wider separation means exponentially better protection. Theoretical estimates suggest hardware error rates of 10⁻⁶ or better, compared to ~10⁻³ for the best conventional physical qubits.

Reduced QEC overhead. If hardware error rates are 10⁻⁶, the number of physical qubits per logical qubit drops from ~1,000 (surface code on conventional qubits) to potentially 1–10. This would make fault-tolerant quantum computing practical with far fewer total resources.

Braiding-based gates. Gate operations via braiding depend only on the topology of the braid path, not on timing, speed, or path precision. This makes gates inherently resistant to calibration drift and control errors, problems that consume significant engineering effort in every other modality.

Disadvantages (Demonstrated)

No demonstrated qubit. The most fundamental problem. Over two decades of effort across multiple world-class laboratories have not produced a single controllable Majorana qubit. The latest device (Majorana 1) demonstrated parity measurements but has not shown qubit control, gate operations, or braiding.

Ambiguous experimental signatures. Every claimed signature of MZMs (ZBCPs, quantized conductance, interferometric phases) has plausible non-topological explanations. The field has not produced a uniquely topological experimental result.

A retracted high-profile paper. The 2021 retraction of the 2018 Nature paper on quantized Majorana conductance damaged the field’s credibility and raised the evidentiary bar for subsequent claims.

Limited universality from braiding. Even if Majorana braiding works, it only implements Clifford gates. Universal computation requires non-Clifford gates (T-gates), which must be produced through magic state distillation or non-topological operations, reintroducing some overhead.

1/f noise may undermine topological protection. The July 2025 Australian preprint suggests that charge noise intrinsic to the semiconductor-superconductor interface could decohere Majorana qubits faster than the topological gap protects them. If correct, this would significantly reduce the theoretical error-rate advantage.

Single-vendor concentration. Microsoft is the only major organization funding a Majorana-based quantum computing program at production scale. Academic groups contribute fundamental research, but the engineering pipeline depends almost entirely on one company’s commitment. Microsoft’s Atom Computing partnership for neutral-atom systems (delivering 50 logical qubits to Denmark by 2027) signals that Microsoft itself is hedging.

Extreme materials requirements. The topoconductor requires atomically clean InAs/Al interfaces, precise gate electrode geometries, millikelvin temperatures, and applied magnetic fields. The fabrication demands are at least as stringent as for superconducting transmon qubits, and the yield of working topological devices appears to be much lower.

Impact on Cybersecurity

Majorana qubits do not affect current CRQC timeline assessments. I do not include them in my CRQC Quantum Capability Framework because there is no demonstrated Majorana qubit from which to extrapolate a scaling trajectory.

If Majorana qubits were realized as theorized, a topological quantum computer running Shor’s algorithm on RSA-2048 could require far fewer physical qubits than any conventional approach. The Gidney 2025 estimate of ~1,400 logical qubits translates to millions of physical qubits for superconducting surface codes but potentially just thousands of topological qubits if hardware error rates are ~10⁻⁶. That would make a CRQC physically smaller, cheaper, and faster to build.

But this is speculation about a technology that has not demonstrated its basic building block. PQC migration timelines should be driven by demonstrated progress in superconducting, trapped-ion, and neutral-atom systems, not by theoretical projections for Majorana qubits. Regulators and clients are setting deadlines based on what has been built, not what has been theorized.

Key Academic Papers

Kitaev (2001/2003). “Fault-tolerant quantum computation by anyons.” Introduced the concept of encoding qubits in non-Abelian anyons and the Kitaev chain model. Published in Annals of Physics. The theoretical foundation for the entire field.

Fu & Kane (2008). Predicted that MZMs could form at the interface between a topological insulator and an s-wave superconductor. Published in Physical Review Letters. Opened the path to hybrid-system Majorana proposals.

Lutchyn et al. (2010); Oreg et al. (2010). Two independent proposals for MZMs in semiconductor nanowires with spin-orbit coupling, proximity superconductivity, and Zeeman field. Published in Physical Review Letters. These papers defined the experimental recipe that every subsequent Majorana experiment has followed.

Mourik et al. / Delft (2012). First ZBCP in an InSb nanowire, reported as a Majorana signature. Published in Science. The claim that launched a decade of experimental effort and debate.

Zhang et al. / Delft (2018, retracted 2021). Claimed quantized Majorana conductance in InSb/Al nanowires. Published in Nature, retracted after data analysis concerns were raised by Frolov et al. The retraction is a defining event in the field’s history.

Nayak, Simon, Stern, Freedman, Das Sarma (2008). “Non-Abelian anyons and topological quantum computation.” The comprehensive review that defined the theoretical framework. Published in Reviews of Modern Physics.

Microsoft Majorana 1 (February 2025, Nature). Device architecture for an 8-tetron chip on InAs/Al topoconductor. Nature reviewers noted the results do not determine whether observed states are topological. My analysis.

Microsoft parity measurements (July 2025). First X and Z loop parity measurements on a tetron device. Demonstrated operational capability needed for qubit control. My analysis.

Australian 1/f noise preprint (July 2025). Analysis suggesting that charge noise intrinsic to semiconductor-superconductor systems may decohere Majorana qubits faster than previously assumed.

Future Outlook

The Majorana qubit program is at a well-defined inflection point. Microsoft has the device architecture (topoconductor + tetron), the fabrication capability (InAs/Al heterostructures), and the measurement infrastructure (parity readout). What it does not have is proof that the states in these devices are topological, or a demonstration that they can be braided.

What would change the assessment. Three things, in order of increasing significance:

1. A measured topological gap in the tetron device that matches theoretical predictions and persists over a range of parameters, confirming that the system is in the topological phase.
2. A braiding operation between two MZMs in the same tetron, with the resulting state change measured and confirmed to match the predicted non-Abelian transformation. This would be the first definitive evidence for non-Abelian anyons in a solid-state system.
3. A two-qubit gate between two tetrons, with measured fidelity demonstrating that topological protection provides a measurable advantage over non-topological qubits in the same material system.

Until at least item 1 is achieved, Majorana qubits remain in the “promising theoretical concept with ambiguous experimental support” category.

Microsoft’s trajectory. Microsoft continues to fund the topological program through DARPA US2QC and internal R&D. The company’s stated goal is fault-tolerant quantum computing within this decade. The Azure Quantum near-term business is built on Atom Computing neutral atoms, with topological qubits as the long-term aspiration. Whether the long-term aspiration survives contact with engineering reality depends on the next 2-3 years of experimental results.

The field beyond Microsoft. Academic groups at Delft/QuTech, Copenhagen, Princeton, Caltech, and in Israel continue fundamental research. IBM has reported interferometry results in fractional quantum Hall systems that may be relevant to non-Abelian anyons. But these are fundamental physics programs, not engineering programs aimed at building computers. If Microsoft reduces its commitment, the pipeline from Majorana physics to Majorana computing narrows significantly.

My assessment. Majorana qubits are the most intellectually compelling idea in quantum computing. The physics is elegant. The theoretical advantages are enormous. And after 13 years of experimental effort since the first claimed observation, the community still cannot agree on whether MZMs have been conclusively observed, let alone controlled or braided. That is the fact that should anchor any assessment of this approach.