Alice & Bob

(This profile is one entry in my 2025 series on quantum hardware roadmaps and CRQC risk. For the cross‑vendor overview, filters, and links to all companies, see Quantum Hardware Companies and Roadmaps Comparison 2025.)

Introduction

Alice & Bob is a Paris-based quantum computing startup (founded in Feb 2020 by Dr. Théau Peronnin and Dr. Raphaël Lescanne) focused on building a universal, fault-tolerant quantum computer using a novel “cat qubit” architecture. The company’s approach leverages superconducting cat qubits, quantum bits that are inherently protected from certain errors, to drastically reduce the overhead for error correction. Backed by substantial funding and scientific pedigree, Alice & Bob has set out an ambitious roadmap aiming to deliver the world’s first “useful” fault-tolerant quantum computer by 2030, culminating in a system with 100 high-fidelity logical qubits (code-named Graphene) demonstrating real-world quantum advantage. In essence, Alice & Bob’s strategy is to “fight decoherence” at the hardware level via cat qubits, thereby making scalable quantum computing practical with far fewer qubits than conventional designs.

Milestones & Roadmap

Alice & Bob has outlined a clear five-phase hardware roadmap toward a fault-tolerant quantum computer by 2030:

Master the Cat Qubit (Boson series) – Achieved in 2024. This milestone produced a reliable and reproducible cat qubit (the Boson chip line) with record-breaking stability against bit-flip errors. The Boson 4 chip demonstrated a bit-flip lifetime exceeding 7 minutes – the longest of any superconducting qubit to date, effectively eliminating bit-flip errors in a single qubit.

Build a Logical Qubit (Helium series) – Underway (2025). Using the Helium chip lineup, Alice & Bob is constructing its first error-corrected logical qubit that operates below the error-correction threshold. This will encode one logical qubit into multiple cat qubits (the company reports ~16 physical qubits per logical qubit for Helium) to suppress the remaining errors. Achieving an under-threshold logical qubit is a critical step toward scalable quantum computing. With the goal of using advanced LDPC codes project to achieve 15:1 ratio for future systems.

Demonstrate Fault-Tolerant Quantum Computing (Lithium series) – Planned next. With the Lithium chip series, the company intends to scale up to multiple logical qubits and demonstrate the first error-corrected logical gate between them. In practice, this means showing that two logical qubits can interact (entangle and perform computations) while errors are being corrected on the fly – a hallmark of fault tolerance.

Unlock Universal Quantum Computing (Beryllium series) – Upcoming. The Beryllium chip family will introduce a universal set of logical gates, enabling execution of any quantum algorithm on the error-corrected platform. Notably, this phase involves generating “magic states” and incorporating Toffoli (or other non-Clifford) gates via on-chip magic state factories. Achieving this will make the architecture fully programmable and universal, capable of running complex algorithms (e.g. Shor’s or Grover’s) on logical qubits.

Deliver Useful Quantum Computing (Graphene series) –Target: 2030. The final milestone is the Graphene chip: a quantum processor with ≈100 high-fidelity logical qubits that can demonstrate quantum advantage in practical applications. At ~100 logical qubits (with an anticipated logical error rate ~10^-8), the system is expected to tackle early industrial use-cases and integrate with HPC infrastructures. This would mark the beginning of the “quantum era” for useful computations, fulfilling Alice & Bob’s goal of a commercially relevant fault-tolerant quantum computer by decade’s end.

Throughout these phases, Alice & Bob has given codenames to each chip series (Boson, Helium, Lithium, Beryllium, Graphene), mirroring the periodic table or materials, to denote progress. The roadmap is aggressive – essentially a five-year plan from 2025 to 2030 – but the company has thus far hit its early milestones on schedule.

Focus on Fault Tolerance

From the outset, Alice & Bob’s strategy has been “fault-tolerance first.” Its cat qubit design was chosen to attack the error problem at the hardware level. Cat qubits are a type of superconducting qubit offering an unrivaled trade-off in error rates: they drastically suppress bit-flip errors while only modestly increasing phase-flip errors. In a normal superconducting qubit (e.g. a transmon), environmental noise causes both bit-flips and phase-flips frequently. In a cat qubit, however, the information is encoded in a superposition of two robust coherent states of a microwave resonator (like Schrödinger’s cat being simultaneously “alive” and “dead”), which “actively stabilize” the qubit against bit-flips. By injecting photons and engineering the resonator’s nonlinearity, Alice & Bob’s qubits achieve an exponential decrease in bit-flip error rates, at the cost of only a linear increase in phase-flip error rate. This means bit-flip errors become vanishingly rare, and the qubit’s main vulnerability is phase errors (a much more manageable problem).

In practice, the focus on fault tolerance has yielded record-breaking results. By 2023, Alice & Bob demonstrated that their second-generation cat qubits were 10,000× more resistant to bit-flips than their first generation – achieving a bit-flip lifetime of about 10 seconds. In 2024, the Boson 4 chip pushed this further to over 7 minutes without a bit-flip (≈430 seconds). For context, dozens of bit-flips occur per second in a typical transmon qubit, whereas Boson 4’s cat qubit didn’t flip for minutes, effectively making bit-flip errors a non-issue during computation. This “six-nines” level of reliability (Alice & Bob’s target is a 10^-8 logical error rate, or about one error in a million operations) is well beyond the conventional fault-tolerance threshold. Importantly, such error rates are below the error-correction threshold, meaning a logical qubit can be formed that improves fidelity rather than worsens it. (To date, only Google has reported physical qubits below threshold using other methods.)

By eliminating bit-flip errors at the physical level, Alice & Bob only needs to correct phase-flip errors using quantum error correction. This is a far simpler task than correcting both error types. The company employs a 1D repetition code (a linear chain of cat qubits) to detect and correct phase flips. In fact, their analysis suggests that to reach a given logical error rate (e.g. 10^-8), the number of physical qubits required is orders of magnitude lower with cat qubits than with traditional qubits. A study cited by Alice & Bob estimates Google’s surface-code approach would need ~1,457 physical qubits to achieve 10^-8 error rate, whereas fewer than 100 physical cat qubits could achieve the same. This huge gain in hardware efficiency – roughly a 10×-15× reduction in qubit overhead – is a cornerstone of Alice & Bob’s fault-tolerant strategy.

The practical focus on fault tolerance is evident in their development path. After mastering the stable cat qubit, the next major step is building an error-corrected logical qubit on the Helium chip. Helium will combine multiple cat qubits (Alice & Bob says 16 physical qubits, vs. Google’s 105 in a comparable demo) to create a single logical qubit with automated error correction. The team has also developed an interim Hydrogen series chip as a testbed for the phase-flip error-correction code, where at least 3 cat qubits in series are used for simple repetition-code error correction. All of this is in service of demonstrating a fault-tolerant logical qubit – one that persistently corrects errors faster than they occur.

Moreover, Alice & Bob recognizes that fault tolerance isn’t just a hardware challenge but a full-stack one. They have partnered with software and quantum architecture experts (for example, working with Riverlane to integrate its advanced quantum error correction decoding stack) to ensure that error correction is handled optimally in real time. This integration of a fast classical decoder and control electronics is crucial for catching and correcting errors (like phase flips) on the fly, a requirement for true fault tolerance. In summary, every aspect of Alice & Bob’s approach – from qubit design to chip architecture to partnerships – is aligned with the goal of minimizing errors and achieving robust, fault-tolerant operation. Their cat qubits “reduce error correction from a 2D problem to a 1D problem,” drastically cutting complexity, and giving them a potential shortcut to scalable quantum computing that competitors relying on brute-force error correction might lack.

CRQC Implications

The prospect of a cryptographically relevant quantum computer (CRQC) looms in the background of Alice & Bob’s roadmap. A CRQC is typically defined as a quantum computer powerful enough to break current cryptographic schemes (like RSA or ECC) – essentially, able to run Shor’s algorithm on large keys within a reasonable time. Estimates vary, but solving a strong RSA encryption by brute-force Shor’s algorithm could require on the order of millions of physical qubits if using conventional error-correction methods. This is why many experts suggest that a large-scale, error-corrected quantum computer (capable of threatening cryptography) might be a decade or more away. Alice & Bob’s approach, however, aims to significantly reduce the resources and potentially the time needed to reach that regime.

According to the company, its cat qubit architecture could dramatically accelerate the path to a CRQC. By design, their qubits are so error-resistant that scaling to cryptographically relevant sizes would need far fewer qubits. “Where conventional approaches would require millions of qubits, we would need only thousands,” says Alice & Bob’s CEO, highlighting the promise that cat qubits could achieve large-scale fault-tolerance much more efficiently. In concrete terms, the Graphene machine planned for 2030 – with 100 logical qubits at 10^-8 error rates – is not itself at CRQC scale, but it’s a pivotal stepping stone. With 100 logical qubits, one could demonstrate quantum advantage in certain problems and possibly run medium-size instances of quantum algorithms. However, breaking RSA-2048 or similar would likely require a few thousand logical qubits (which, with Alice & Bob’s efficiency, might correspond to tens of thousands of physical cat qubits).

Crucially, Alice & Bob’s roadmap doesn’t stop at 100 logical qubits. They explicitly state an intention to “quickly scale to thousands of logical qubits” after Graphene, targeting industry-relevant use cases beyond the initial machine. If they achieve Graphene on schedule, one could envision an early-2030s device with, say, 500-1000 logical qubits, which starts encroaching on the lower end of CRQC capability. Their emphasis on a hardware-efficient, fault-tolerant architecture means that each additional qubit they add is a logical (usable) qubit rather than mostly overhead. This could shorten the timeline to a CRQC considerably, compared to approaches that might need to scale to millions of physical qubits for the same result.

That said, 100 logical qubits in 2030 will likely not be enough to crack modern cryptography outright. For example, one analysis cited that implementing Shor’s algorithm for cryptographically relevant numbers might demand on the order of 20 million physical qubits (using surface codes), which even with a 20:1 efficiency improvement would translate to perhaps ~1 million physical cat qubits. Alice & Bob’s thousand-logical-qubit scale machines (post-2030) would be a step toward that, but further scaling and engineering would be needed to fully realize a CRQC. Nonetheless, the implications are clear: if Alice & Bob’s technology works as hoped, the threshold for breaking encryption could be reached with far fewer qubits than previously thought, hastening the day when quantum computers pose a real threat to classical cryptography.

Indeed, Alice & Bob’s white paper emphasizes potential impacts on cybersecurity alongside scientific applications. A fault-tolerant quantum computer can, in principle, run algorithms to decrypt RSA or simulate AES-breaking Grover iterations at scales that insecure classical cryptography. Thus, as Alice & Bob and others progress, it reinforces the urgency for deploying post-quantum (quantum-resistant) cryptographic algorithms. In summary, while Alice & Bob’s 2030 machine will be a remarkable milestone but probably not a CRQC by itself, it will lay the groundwork for one. Should they successfully demonstrate 100 logical qubits with high fidelity, scaling that design up to the thousands needed for cryptographic tasks becomes an engineering project with a plausible path – potentially bringing a CRQC within reach sooner than many anticipated. Their roadmap, if achieved, would represent a significant acceleration toward the cryptographic breaking point, underlining how this startup’s progress is closely watched not just in computing but also in security circles.

Modality & Strengths/Trade-offs

Alice & Bob’s quantum hardware modality is a unique twist on the superconducting quantum circuit paradigm. Instead of using transmon qubits (the typical nonlinear superconducting circuits used by IBM, Google, etc.), Alice & Bob uses bosonic mode qubits, i.e., quantum states of microwave resonators often called “cat qubits.” In a cat qubit, information is encoded in two coherent states of a resonator – roughly speaking, the resonator holds a microwave photon field that can be in a superposition of two opposite-phase configurations (|α⟩ and |-α⟩). These two states form the qubit’s effective |0⟩ and |1⟩. Thanks to the physics of these states, a cat qubit gains stability: it is much less likely to randomly jump (flip) between |0⟩ and |1⟩ than a standard qubit would. Alice & Bob achieves this by designing the circuit with no stand-alone transmon: their chips use a nonlinear element (such as a Josephson circuit) in tandem with the resonator mode to create a so-called Kerr-cat qubit that does not suffer bit-flips for minutes. In fact, the Boson 4 device was described as a “transmon-free” design, emphasizing that the usual source of noise (the transmon) was eliminated to extend coherence.

Strengths: The cat qubit modality’s main strength is its error resilience. By hardware-encoding the qubit in a way that bit-flip errors are exponentially suppressed, Alice & Bob immediately removes one of the two major error channels. This yields extremely high coherence times for the qubit’s population (|0⟩/|1⟩) state – as demonstrated by the 7-minute bit-flip time record. The error that remains (phase flips, where |0⟩ turns into -|0⟩, for instance) can be corrected by redundancy, but because phase errors increase only linearly with the cat qubit’s parameters, the net error rate is much lower than in other qubit modalities. The trade-off between bit-flips and phase-flips is skewed very favorably in cat qubits: conventional superconducting qubits experience both types of errors at similar rates, whereas cat qubits turn the bit-flip rate into a negligible factor, at the cost of a modest phase-flip rate increase. This inherent bias drastically cuts down the overhead for quantum error correction. As noted, to reach a certain fidelity target, a cat-qubit-based quantum computer will need far fewer physical qubits than one based on, say, transmons with surface code. Alice & Bob expects that thousands of cat qubits could do what millions of regular qubits are projected to be required for. In other words, the hardware efficiency (logical qubits per physical qubits) is a key strength – they predict ~100 logical / 1500 physical at 10^-8 error rates (a 1:15 ratio), which is a dramatic improvement over the ratios estimated for other platforms.

Another strength of Alice & Bob’s modality is that it builds on well-understood superconducting circuit technology (microwave resonators, Josephson junctions, cryogenics, etc.), meaning it can leverage existing fabrication and control techniques. Yet, by using a novel qubit design, they evade some limitations of standard transmons (like relatively short T1 lifetimes). They also demonstrated the ability to integrate these qubits into a system: Boson 4 and Hydrogen/Helium chips indicate a roadmap of scaling from one qubit to a few and onward, showing that the approach is not just a theoretical idea but a working system. The cat qubit’s design also allows certain operations that align well with error correction; for example, measuring the joint parity of a pair of cat qubits can detect a bit-flip error without collapsing the logical state – this kind of operation is naturally supported in bosonic encodings and is useful for their repetition code error correction scheme.

Trade-offs and Challenges: The cat qubit approach does carry some trade-offs. One is that while bit-flip errors are suppressed, the phase-flip errors are still present and indeed will occur more frequently as the cat states are made larger (more photons). This means that an error-correcting code is still necessary – you must network multiple cat qubits to handle phase errors. Alice & Bob uses a 1D chain of cats (a repetition code) to correct phase flips, which itself requires extra qubits (e.g. three physical cat qubits might be the minimum to correct one phase error). Thus, the system is not free of overhead; it’s just far less overhead than 2D surface codes. Another consequence is that operations (gates) between cat qubits can be more involved. For instance, implementing a two-qubit gate might require an intermediary coupling element or careful pulse sequences that do not disturb the delicate cat states. The company will have to engineer high-fidelity logical gates – they plan to demonstrate the first in the Lithium stage – and this remains to be proven. There’s a known complexity in performing non-Clifford gates (like Toffoli) on encoded qubits; Alice & Bob’s solution is to incorporate magic state distillation on chip (the Beryllium stage), but magic state generation itself incurs significant overhead and error-handling. So, while cat qubits reduce the cost of error correction for storage of qubits, the cost for implementing certain operations might still be high (though presumably still cheaper than with other methods). This is a design trade-off: easy error correction of one type, but needing creative solutions for full universality.

From a hardware standpoint, cat qubits require maintaining high-quality resonators and potentially coupling multiple resonators. The hardware is still cryogenic – the entire system sits in a dilution refrigerator at ~10 millikelvin, just like other superconducting qubit setups. Scaling up to many qubits means scaling wiring, control electronics, and cryogenic infrastructure. In fact, resonator-based qubits can occupy a fair amount of physical space on a chip (a high-Q resonator may be several millimeters in length). Alice & Bob will have to innovate in packaging and integrating potentially thousands of resonators (or else find ways to make them smaller or 3D-integrated) to reach the Graphene scale. This challenge is common to all superconducting approaches but could be accentuated if each logical qubit spans multiple resonators, etc. The company’s Cat Factory project – aiming to make fault-tolerant architectures “10× cheaper and optimized” – likely is addressing such issues (how to arrange and fabricate large numbers of cat qubits efficiently).

In summary, Alice & Bob’s modality offers a clear advantage in error suppression, giving it a head start on the road to fault tolerance. The strengths are high fidelity (long coherence), lower qubit count requirements, and leveraging proven tech in a new way. The trade-offs include the need to tackle phase errors via coding (which, while much easier than full-blown surface coding, still requires work), the complexity of multi-qubit gate implementation, and the general engineering challenges of scaling a superconducting system. Nonetheless, the company’s distinctive focus – building a quantum computer “protected from bit-flips by design” – sets it apart. If successful, the cat qubit modality could fundamentally alter the quantum computing race by making fully error-corrected qubits attainable with far fewer resources than previously thought.

Track Record

Although a young company, Alice & Bob has compiled an impressive track record of achievements in both research and development. The scientific roots of the company are strong: the core concept of a bit-flip-protected qubit was validated early on. In fact, the co-founders’ work on cat qubits was published in Nature Physics in 2020, demonstrating the viability of exponentially suppressing bit-flips in a superconducting qubit. Building on that breakthrough, Alice & Bob was founded the same year, and it quickly set about turning theory into practice.

Over 2021-2023, the team developed a series of prototype chips (the Boson series), each improving on the last. By 2023, they announced that their cat qubits had achieved a bit-flip lifetime of 10 seconds, which was a four-orders-of-magnitude improvement over initial implementations. This 10-second coherence against bit-flips indicated that their qubits’ error rates were well below the threshold needed for error correction – a crucial proof-point that their hardware approach is valid. It essentially showed that the cat qubit could “lock in” quantum information for significantly longer than a normal qubit, without bit-flip errors occurring.

Then in May 2024, Alice & Bob launched the Boson 4 chip publicly on the cloud, garnering widespread attention. Boson 4 is a single-cat-qubit device that set a world record for stability: over 7 minutes without a bit-flip. (In comparison, typical superconducting qubits have T1 relaxation times on the order of 0.1 milliseconds to 0.1 seconds.) This accomplishment was not only reported in press releases but also made accessible via Google Cloud Marketplace, meaning external researchers and users could run experiments on this record-breaking cat qubit themselves. By making Boson 4 publicly available, Alice & Bob demonstrated confidence in their technology and a commitment to transparency and collaboration. The data from Boson 4 confirmed the predicted exponential suppression of bit-flips (each new iteration from Boson 1 to 4 showed orders of magnitude improvement, as plotted in their reports).

Alongside hardware milestones, Alice & Bob has been methodically building out its capabilities and resources. In 2022, it raised a Series A funding (€27M), and by early 2025 it closed a €100M Series B round, bringing total funding to about €130M to date. This makes Alice & Bob one of Europe’s best-funded quantum startups. The investors include both venture firms and public funds (e.g. Bpifrance, the French public investment bank), underlining confidence in the company’s vision. The funding has enabled rapid growth: the team doubled to ~110 employees (including 30+ PhDs) by late 2024, and the company expanded internationally, opening a Boston office in 2023 to connect with the U.S. quantum ecosystem. This U.S. presence helps in hiring talent and forging partnerships (for example, with U.S.-based quantum software firms and national labs).

In terms of partnerships and collaborations, Alice & Bob has been very active. One notable partnership is with Riverlane, a quantum computing software company, to integrate Riverlane’s error-correction stack with Alice & Bob’s hardware. This collaboration is critical for the real-time decoding of errors in a fault-tolerant quantum computer – essentially uniting Alice & Bob’s robust qubit hardware with state-of-the-art classical software for error syndrome processing. Another collaboration is the government-supported Cat Factory project: In March 2024, Alice & Bob and academic partners (ENS de Lyon and Mines Paris) won a €16.5M grant to develop a new architecture for fault-tolerant quantum computing by 2027. The Cat Factory initiative aims to make the scaling of cat qubits more cost-effective (targeting a 10× reduction in cost/complexity), which is a strategic endeavor as the company approaches the more complex phases of its roadmap. Additionally, Alice & Bob has engaged in joint efforts on applications – for instance, working with other French quantum players (like Quandela and EDF) on quantum computing’s energy efficiency and looking for “killer applications” that a medium-scale fault-tolerant machine (a few hundred qubits) could tackle. All these efforts build the ecosystem around Alice & Bob’s technology and prepare the market for when their hardware comes of age.

Finally, it’s worth noting that Alice & Bob’s progress has been recognized in the broader community. Their achievements (such as the 7-minute coherence record) have been covered in major industry publications and even resulted in academic publications. They have effectively established themselves as the leader in cat qubit technology – indeed, the company notes it is the only one exclusively pursuing cat qubits for quantum computing. Each milestone achieved (Boson 4, etc.) not only validates a piece of their roadmap but also builds credibility that the next milestone (Helium logical qubit, and so on) can be reached. In summary, Alice & Bob’s track record so far shows a steady and deliberate climb: from theoretical research to practical prototype, from prototypes to record-setting devices, and from devices to an integrated plan backed by strong funding and partnerships. This lends weight to their claim that they can deliver a useful quantum computer by 2030 – they have consistently met technical goals and secured the necessary support to move forward.

Challenges

While Alice & Bob’s vision is bold and its progress notable, there remain significant challenges and uncertainties on the road ahead. Achieving a fully fault-tolerant quantum computer is an enormous scientific and engineering feat, and the company must navigate multiple risk points:

Technical Scaling and Unproven Steps: Thus far, Alice & Bob has demonstrated single qubits with extraordinary coherence. The next steps – building a logical qubit (Helium), then a logical-qubit gate between two logical qubits (Lithium), and eventually a network of 100 logical qubits (Graphene) – each introduce new complexities. Creating the first logical qubit will test whether their error-correction code (a repetition code on cat qubits) truly works as expected in practice, catching phase-flip errors reliably without introducing too much overhead. This involves integrating multiple cat qubits and several ancilla/measurement circuits on one chip, something that hasn’t been done before with this technology. Then, demonstrating a fault-tolerant two-qubit gate (the Lithium milestone) is another leap – it requires coupling two logical qubits and performing an entangling operation while both remain error-corrected. Any uncontrolled interaction or excess noise could unravel the fragile error-correcting process. These are frontier problems: no one has yet shown a fully error-corrected logical gate in any platform, so Alice & Bob will be attempting a world-first. The Toffoli gate via magic states planned for Beryllium is yet another unproven element – magic state distillation consumes additional qubits and rounds of error correction; it’s a notoriously complex part of fault-tolerant quantum computing. Ensuring that this can be done on their hardware (and done repeatedly and quickly enough) is a big challenge. In essence, Alice & Bob must turn their qubits into an engineered system: it’s one thing to have long-lived qubits, but another to have them interact and compute reliably as a collective.

Engineering and Scale-Up: Hitting the 100 logical qubit target by 2030 means scaling the hardware roughly by two orders of magnitude in terms of qubit count (compared to near-term demonstrations). Graphene is envisioned to use ~1,500 physical qubits (to get 100 logical). Managing 1,500 microwave resonators with control lines, readout devices, etc., all while keeping coherence high, is a massive engineering undertaking. Issues such as cross-talk between qubits, frequency crowding (resonators need distinct frequencies or other isolation), and heat load in the cryostat could all become limiting factors. Alice & Bob is preemptively addressing some of this: they announced plans for a new 4,000 m² facility in Paris dedicated to quantum hardware production and integration. This large lab space (and the investment in equipment it implies) will be needed to fabricate and test complex chips like Graphene. They’re also engaged in the Cat Factory project to optimize architecture by 2027 – presumably developing techniques like perhaps 3D integration or better cryo wiring to handle the scale-up. Despite these measures, it remains challenging to deliver within five years a machine that outstrips anything built so far. By comparison, giants like IBM have outlined a path to thousands of qubits by 2026-2027 but even those are non-error-corrected physical qubits (and IBM has a large engineering team and infrastructure). Alice & Bob will need to marry innovation with robust engineering to not only build but also maintain a 100-logical-qubit processor. Reliability of every component (amplifiers, cryocoolers, control electronics) becomes crucial when running long algorithms – something the company will have to ensure through rigorous systems engineering.

Timeline and Competition: The 2030 timeline is ambitious. Alice & Bob essentially has to go from a few-qubit prototype to a full fault-tolerant machine in a half-decade. Meanwhile, the competition in the quantum computing race is intense. As noted by their team, PsiQuantum (a photonic quantum computing startup) has publicly aimed for ~200 logical qubits by 2027 using a silicon photonics approach. IBM has also declared goals to achieve large-scale error-corrected qubits in the second half of the 2020s (IBM’s roadmap includes testbed logical qubits and perhaps ~1,000+ physical qubits chips by 2025, and beyond) – one statement interpreted by Alice & Bob was IBM aiming for 200 logical qubits that can execute 100 million quantum gates (though with unspecified fidelity). In this landscape, Alice & Bob, as a startup, faces the challenge of keeping up momentum and not falling behind. The fact that they focus on quality (fidelity) over raw qubit count could be an advantage – e.g., 100 high-fidelity logical qubits might outperform a larger number of low-fidelity qubits. Still, if a competitor achieves a major fault-tolerance milestone earlier (say, the first logical qubit or a small logical computer by 2027), it could affect Alice & Bob’s funding or partnership opportunities. The unknown performance of competitors is something even their executives mention: “the devil is in the details” – one can promise logical qubits, but what matters is error rates and gate fidelity. Alice & Bob is betting that its approach yields superior error rates (10^-6 LER) and thus more useful qubits, albeit on a slightly later timeline. They will need to demonstrate each step clearly to maintain credibility in the face of big announcements from others.

Financial and Talent Needs: As a deep-tech venture, Alice & Bob will continuously require significant capital and skilled personnel. They have done well in funding so far (with €130M raised, plus government support), but building a fault-tolerant quantum computer might demand more resources than currently in hand. Their COO has noted that the “quantum industry needs more money… especially when competing with large American corporations that have almost unlimited cash“. This means Alice & Bob must carefully steward their funds and likely will seek further investments or partnerships down the line. On the talent front, they are expanding the team and have been successful in attracting top scientists and engineers (benefiting from Europe’s strong quantum education pipeline). But scaling from 110 people to the few hundred that might be needed to build and support a full quantum computing stack is non-trivial. They plan to reach ~200 employees in the next phase, which will require not just hiring but training people in this very specialized technology. Competition for quantum experts is global and intense; keeping their team “very selective” yet growing might be a delicate balance.

Unforeseen Technical Risks: Lastly, there are always unknown unknowns. The cat qubit approach, while backed by solid theory and early experiments, could encounter new problems when scaled. For instance, will maintaining dozens of photons in a resonator (for a very long time) lead to unexpected noise or interactions? Could there be materials or fabrication issues that limit coherence when many resonators are on one chip? Also, integrating magic state factories introduces probabilistic elements (since distillation circuits have a yield <100%), which could complicate the architecture’s performance. Alice & Bob will have to iterate quickly on designs if something doesn’t work as anticipated – flexibility and contingency plans are important in such cutting-edge development.

In summary, the road to 100 logical qubits is challenging on multiple fronts: scientific (demonstrating new quantum error correction feats), engineering (scaling the hardware and infrastructure), and competitive/business (maintaining leadership and resources in a fast-moving field). The company is mitigating these challenges through careful planning – e.g., defined milestones, new facilities, partnerships (for software, academia, and even finding applications) – and thus far has shown a realistic understanding of the work ahead. Achieving a universal fault-tolerant quantum computer by 2030 will require near-flawless execution of their roadmap. It’s a high-risk, high-reward endeavor: if Alice & Bob succeeds, it could vault ahead in the quantum race with a truly revolutionary machine; if some assumptions don’t hold, timelines may slip as with many complex technologies. Ultimately, their focus on fault tolerance via hardware efficiency gives them a unique fighting chance, but the next few years of results will be crucial to validate that their approach can scale as predicted.