In a significant quantum computing breakthrough, researchers from the University of Basel and IBM Research–Zurich have achieved a controlled interaction between two quantum bits inside a standard silicon transistor. The team’s new paper “Anisotropic exchange interaction of two hole-spin qubits” in Nature Physics reports that they realized high-speed, high-fidelity operations between hole-spin qubits implemented in a fin field-effect transistor (FinFET) – a workhorse device of modern computer chips. This is the first time a two-qubit logic gate has been demonstrated using holes (the absence of electrons) confined in an industry-standard transistor structure, without any trade-off between operation speed and accuracy. The accomplishment marks a major step toward integrating quantum computing with mainstream semiconductor technology, since it shows that quantum bits (qubits) can leverage the same devices and fabrication processes used for billions of classical transistors.
The significance of this result lies in its promise of scalability and compatibility. Quantum computers today remain limited to relatively small numbers of qubits, in part because they rely on exotic hardware that doesn’t easily scale. By contrast, FinFET transistors are ubiquitous in CMOS chips, so demonstrating qubits in a FinFET suggests a path to manufacturing quantum processors on a massive scale. The Basel/IBM team’s qubits achieved fast and reliable gate operations within a device essentially identical to those in commercial chips, underscoring the potential for combining quantum and classical computing architectures on the same silicon platform. In other words, this breakthrough hints that future quantum processors might be built with the very same technology that powers today’s smartphones and CPUs, vastly accelerating the marriage of quantum computing with the semiconductor industry.
Understanding Hole-Spin Qubits and the New Research
What is a hole-spin qubit?
In a semiconductor, a “hole” is the absence of an electron – essentially a positively charged carrier that can move and has an intrinsic spin (up or down) just like an electron. A hole-spin qubit encodes quantum information in the spin state of a single hole, with “spin-up” and “spin-down” serving as the qubit’s two states (analogs of 0 and 1). By trapping an individual hole in a tiny region (a quantum dot) under a transistor gate, researchers can isolate a controllable qubit. Hole-spin qubits are especially promising for scalable quantum computing because of a key advantage: they can be controlled entirely by electrical signals rather than by magnetic fields. In silicon nanostructures, holes exhibit strong spin–orbit coupling, meaning the orientation of a hole’s spin is sensitive to its motion and electric fields. This allows engineers to flip or rotate the spin using fast voltage pulses on a gate electrode, without needing on-chip microwave antennas or magnets. Electrical control is not only convenient; it’s compatible with standard chip technology, making hole-spin qubits easier to integrate into devices like transistors.
Notably, hole-spin qubits can operate at higher temperatures than many other qubit types. In prior work (2022), the Basel group showed that a single hole spin could be trapped and coherently controlled in a silicon FinFET even at temperatures above 4 kelvin. This “hot” operation (by quantum standards) is far above the millikelvin regimes required by superconducting qubits, easing cryogenic overhead. At 4 K, cooling is much simpler and more power is available, opening the door to integrating classical control circuits alongside the qubits in the same cryogenic environment. The latest research builds on that foundation of electrically controlled, relatively high-temperature qubits, and takes the crucial next step: getting two qubits to interact in a controlled way inside one FinFET device.
What did the new experiment achieve?
The researchers fabricated a double quantum dot within a single FinFET, creating two nearby hole-spin qubits, and succeeded in coupling them to perform a two-qubit logic operation. In essence, they demonstrated a controlled spin-flip gate: the spin of one qubit flips only if the other qubit is in a particular spin state. This is analogous to a controlled-NOT (CNOT) gate, a fundamental building block of quantum computation. Achieving such a gate is non-trivial because it requires the two spins to interact strongly yet remain individually addressable and coherent. The team exploited the natural exchange interaction between the two holes – an electrostatic interaction that effectively correlates their spins when they are made indistinguishable and in close proximity. Importantly, they found that the exchange coupling in their device is highly anisotropic, meaning its strength depends on the relative spin orientation, a direct consequence of the spin–orbit physics in silicon FinFETs. By electrically tuning this coupling via the gate voltages, they could turn the interaction on and off and implement the controlled spin-flip with precision.
Crucially, this two-qubit gate was achieved without compromising on speed or fidelity. Fast gate operations are desired to outrun decoherence, but speeding up interactions can often introduce errors. Here, thanks to the anisotropic exchange mechanism, the researchers were able to enact a swift two-qubit rotation while maintaining high fidelity (low error rates) of the operation. As Dr. Andreas Kuhlmann of the team explained, the intrinsic anisotropy “makes two-qubit gates possible without the usual trade-off between speed and fidelity” that plagues many quantum technologies. In other words, the hole-spin qubits flipped as fast as needed for practical computing, yet behaved reliably enough to be viable for larger quantum circuits. This result was observed in a device identical in design to a commercial 15-nm FinFET transistor, underscoring that quantum functionality was achieved without any exotic materials or structures. The demonstration of two high-quality qubits inside a single transistor is a landmark for semiconductor-based quantum dots, showing that one can perform real quantum logic operations using the building blocks of conventional chips.
Comparison with Other Qubit Technologies
How do hole-spin qubits in silicon stack up against other leading qubit platforms? Each approach to quantum computing comes with its own pros and cons, and the new results accentuate the unique strengths of semiconductor spin qubits vis-à-vis competitors:
Superconducting Qubits: These qubits (used by IBM, Google, and others) are based on circuits of superconducting materials and manipulated with microwave signals. They offer fast gate speeds on the order of tens of nanoseconds and have benefited from integrated chip fabrication techniques. However, superconducting qubits require ultra-cold dilution refrigerators (~10–20 millikelvin) for stability, and their coherence times (typically microseconds to milliseconds) are shorter than some alternatives. Today’s superconducting quantum processors have reached dozens to a few hundred qubits, but scaling further faces challenges in wiring and cross-talk at such extreme cold temperatures.
Trapped Ion Qubits: Trapped-ion systems (pursued by IonQ, Quantinuum and academic labs) use individual ions levitated in electromagnetic traps and manipulated with laser beams. Their chief advantage is excellent qubit quality: ions of certain elements can maintain superposition for seconds or longer, giving them very long coherence times and high-fidelity operations (single- and two-qubit gate fidelities over 99%). The trade-off is speed and size – gate operations take on the order of milliseconds, and the apparatus (vacuum chambers, lasers) is relatively bulky, which makes scaling to millions of qubits difficult in practice. Connectivity in ion traps can be high (many ions in one trap can be fully entangled), but shuttling ions or using photonic links is needed to scale beyond a single trap, which adds complexity.
Photonic Qubits: Photonic quantum computing encodes information in particles of light (photons), often using polarization or path as the qubit. Photons can travel at room temperature and over long distances, making them ideal carriers of quantum information for networking and communication. Moreover, photonic qubits do not decohere in the same way matter-based qubits do, as they rarely interact with their environment. The big challenge is performing two-qubit logic with photons: it typically requires delicate interference setups or effective nonlinear interactions, which are probabilistic or demand special materials. As a result, while photonic systems are great for quantum communication and have achieved small-scale quantum logic, they face significant hurdles for building a general-purpose quantum computer with many logic gates.
Semiconductor Spin Qubits (Electrons or Holes): This category includes the new silicon hole-spin qubits and the closely related electron-spin qubits in quantum dots. Their standout feature is extreme miniaturization – the qubit is a single charge in a nanoscale transistor, so devices can be densely packed. They can leverage the vast CMOS semiconductor manufacturing infrastructure, promising compatibility with existing chip fabrication and potentially the integration of classical control electronics on the same chip. Additionally, spin qubits can tolerate oK to 1–4 K temperatures (significantly higher than superconducting qubits’ mK range), simplifying cooling requirements slightly. Long coherence times are possible especially for electron spins in purified silicon (T2 of seconds has been shown in isolated cases), though the strong spin–orbit coupling in hole spins means they must carefully mitigate charge noise to approach such stability. So far, semiconductor spin qubits are at an earlier stage of development: only small numbers of qubits (handfuls) have been entangled in labs, and controlling large arrays of spins and their interconnections is an active research challenge. However, the Basel/IBM result shows that their performance can rival other platforms in speed and fidelity on the two-qubit scale, and the promise of monolithic integration and high density gives spin qubits a compelling long-term advantage.
In summary, superconducting qubits currently lead in maturity (with tens to hundreds of qubits demonstrated) and offer fast operations, but require extreme cooling and face scaling bottlenecks; trapped ions excel in precision and connectivity but are slow and hard to miniaturize; photonic qubits work at room temperature and are natural for communication, yet are difficult to use for computation; and semiconductor hole-spin qubits, as showcased in this breakthrough, offer a path to nano-scale quantum processors that piggyback on the semiconductor industry’s technology, though they are still catching up in realized qubit count and overall complexity. It’s possible that these paradigms will coexist, each suited to different applications – for example, photonic qubits for networking, ion qubits for specialized high-precision needs, and silicon spin qubits for large-scale integrated quantum chips.
Industry Adoption and Scalability Potential
This advance comes at an interesting intersection of industry efforts. Tech giants and startups alike are racing to build useful quantum computers, but they have taken divergent technological routes. The success of hole-spin qubits in a FinFET suggests a new convergence point where quantum computing meets the established semiconductor industry. How does this development fit into the broader strategies of companies like IBM, Intel, and Google?
IBM: IBM has been a leader in quantum computing, primarily with its superconducting qubit processors. It has already built devices with 433 qubits (IBM Osprey) and is aiming for thousands by 2025 on its roadmap. That said, IBM is also exploring quantum devices that leverage silicon technology through its research divisions. The collaboration with University of Basel indicates IBM’s interest in hedging its bets with silicon spin qubits that could offer better scalability. IBM’s vision for quantum computing involves modular “quantum centric” supercomputers in the long term, which will likely require millions of physical qubits. The fact that two high-quality qubits can reside in one silicon transistor is a tantalizing prospect: IBM could potentially incorporate such silicon-based qubit arrays into future systems or pair them with its superconducting technology, marrying fast solid-state qubits with the integration density of CMOS. In practical terms, IBM and other firms might continue their current platform (superconducting in IBM’s case) to reach intermediate milestones, while in parallel refining silicon-based qubits for the generation beyond, when packing qubits by the million on a chip becomes the priority.
Intel: Unlike IBM and Google, Intel’s quantum strategy has focused squarely on silicon spin qubits from the start, leveraging the company’s unparalleled expertise in semiconductor manufacturing. Intel has been designing quantum dot spin qubits on 300 mm silicon wafers using standard CMOS processes, essentially treating qubits as just another transistor type. The rationale is clear: if qubits can be made with the same processes as classical processors, scaling to large numbers and high yield should be much more straightforward. Recent milestones from Intel support this approach – for instance, its researchers developed a cryogenic wafer probe to characterize thousands of quantum dot devices at once, and achieved single-qubit fidelities of 99.9% on devices made entirely with industrial fab techniques. Those are record numbers for any qubit made in a mass-production environment. The hole-spin qubit breakthrough reinforces Intel’s path: demonstrating two-qubit gates in a FinFET means the fundamental logic operations can happen in a standard transistor geometry. It aligns with Intel’s vision of eventually having large-scale quantum chips fabricated with high uniformity, where each qubit is as reliable as a transistor on a logic chip. In the bigger picture, Intel – and partners like TSMC or Samsung if they enter the fray – could one day produce quantum processor chips in their fabs, potentially integrating thousands or millions of spin qubits along with classical co-processors on the same die. The compatibility with FinFET architecture is crucial here: it means quantum circuits don’t require a whole new manufacturing paradigm, only some adaptation of existing ones. This could drastically accelerate adoption, since it piggybacks on the decades of refinement in semiconductor mass production.
Google and others: Google’s quantum effort has centered on superconducting qubits (e.g. the Sycamore processor) and on demonstrating quantum error correction. It hasn’t announced a pivot toward silicon qubits, but the company is certainly monitoring all breakthroughs. As the field progresses, even companies heavily invested in one platform may consider hybrid approaches. Google, for instance, might continue with superconducting qubits for the near term – given their current lead in that area – but could down the road explore integrating spin-qubit-based memory or scalable modules if those outperform in density. Other players like IBM (as noted), Rigetti (superconducting startup), IonQ (trapped-ion systems), and academic consortia each have their chosen technologies, but the advent of a CMOS-compatible qubit is something of a game-changer. It invites the semiconductor industry at large into the quantum computing race. Companies that build classical chips could now have a clearer entry path to build quantum chips. We may soon see foundries collaborating on quantum dot arrays or tech giants forming partnerships with chip manufacturers to produce quantum processors. In short, this result boosts the scalability potential of qubits: if quantum bits can be made using the same techniques as classical bits, scaling up becomes a matter of engineering and refinement, not fundamental reinvention. As the study authors noted, qubits based on hole spins “leverage the tried-and-tested fabrication of silicon chips” and have proven to be fast and robust, giving this approach a strong chance in the race toward large-scale quantum computers.
Moreover, integration with semiconductor technology could enable on-chip control logic and error correction circuits physically close to the qubits. Because the hole-spin qubits can operate at around 4 K, one can imagine cryogenic CMOS control electronics sitting at the same temperature tier, reading and programming qubits without needing thousands of long wire connections out to room temperature. This addresses the so-called “wiring bottleneck” that current quantum systems face, where each qubit often requires dedicated cabling that doesn’t scale. Industry adoption will hinge on such practical engineering solutions. The fact that the demonstrated qubits are in a FinFET means engineers are already familiar with the device structure, and it can be tiled in two-dimensional arrays on a chip. All these factors contribute to a scalable roadmap: in principle, millions of qubits on a single chip (necessary for fault-tolerant quantum computing) might be realized by tiling arrays of FinFET qubits, much as we tile billions of transistors today. While significant work remains to reach that point, the path forward has become much clearer with this breakthrough.
Challenges and Roadblocks
Despite the excitement, it’s important to temper expectations with the challenges that hole-spin qubits – and quantum processors generally – still face on the road to practical adoption. Here are some of the key hurdles and considerations moving forward:
Coherence and Noise: Quantum states are notoriously fragile. Hole-spin qubits benefit from fast operation, but the same spin–orbit interaction that lets us control them electrically also makes them sensitive to charge noise in the semiconductor environment. Fluctuations in electric fields or charges in nearby materials can disturb the hole’s spin, limiting how long the qubit can retain information (its coherence time). While electron-spin qubits in purified silicon have achieved coherence times of seconds in isolated lab setups, hole-spin qubits in realistic devices will need materials and engineering improvements to approach those levels. The Basel/IBM team’s ongoing work is to make the qubits more coherent (i.e. extend their useful lifetime) while also reducing gate times further. Techniques like finding “sweet spots” where the qubit is first-order insensitive to electric noise (through device design or bias tuning) are being explored to mitigate decoherence. Progress is being made, but maintaining quantum coherence in a complex chip with millions of holes will be an enormous challenge.
Error Rates and Scaling Up Logic: A practical quantum computer needs not just a few high-fidelity qubits, but thousands or more, all interacting in a controllable way with error rates low enough for error correction. The two-qubit gate demonstrated is fast and high-fidelity on its own, but as more qubits are added, errors can accumulate and interactions can become harder to control precisely. Error rates need to be pushed down toward the threshold required by quantum error-correcting codes (often quoted on the order of ~0.1% or lower per gate for surface codes). The best reported single-qubit gate fidelities for silicon spin qubits are around 99.9%, and two-qubit gates around ~98-99% in experimental settings, which is encouraging. However, integrating many qubits could introduce new cross-talk and noise channels. As one scales up, calibration of numerous qubits and couplers becomes labor-intensive. New techniques in automation, calibration, and perhaps machine-learning-driven tuning will be needed to handle large arrays of quantum dots. Error correction itself will require a large overhead in qubit count – potentially thousands of physical qubits for one logical qubit. Ensuring that a silicon-based qubit architecture can support that (and do so better or more cheaply than other architectures) is still an open question.
Fabrication and Uniformity: While using standard CMOS fabs is a huge advantage, quantum devices impose more stringent demands on variability and disorder. Each qubit – each tiny quantum dot and transistor – must have well-controlled properties. Even atomic-scale imperfections can affect qubit behavior. The FinFET design helps in reproducibility, but as qubit counts rise, uniform fabrication and yield become critical. Intel’s recent work shows that it’s possible to achieve high yield and uniformity for spin qubits across a whole 300 mm wafer, which bodes well. Still, any large-scale quantum chip will need thorough testing of every qubit and the ability to discard or correct for fabrication defects. The fabrication challenge also includes integrating classical and quantum devices – for instance, adding control electronics or interconnects near the qubits without disrupting the qubit performance. This will require clever co-design of classical and quantum circuits, new cryogenic electronic components, and perhaps 3D integration (stacking) to separate layers of qubits and control wiring. All of this must be done while keeping the process compatible with commercial fabrication lines.
Thermal Management: Operating at ~4 K is much warmer than millikelvin, but it’s still cryogenic. If millions of qubits and associated control circuits are on a chip, they will generate heat that needs to be removed to keep the device at 4 K. Each classical control transistor dissipating even a tiny amount can add up across a large chip. Thus, engineering the system for heat removal and perhaps using multiple cooling stages is important. The advantage is that commercial cryocoolers at 4 K can provide a lot more cooling power (watts) than dilution refrigerators at mK (which provide milliwatts or less). This makes the concept of on-chip control feasible, but it’s still a demanding thermal design problem to ensure stability at scale.
Architectural and Interconnect Challenges: In a processor with many qubits, not every qubit can directly interact with every other; there will be some layout constraints (typically nearest-neighbor coupling in a grid). Managing qubit connectivity and performing operations between distant qubits may require moving quantum information around (quantum state transfer between quantum dots, or swap networks). Techniques like shuttling spins through quantum dot arrays or using microwave resonators as buses are being researched. Each approach adds complexity and potential error sources. Developing a robust architecture – perhaps tileable quantum modules that can be linked – is a task ahead for designers of silicon quantum chips. This is analogous to how classical CPUs moved from a few cores to many-core designs and had to introduce network-on-chip architectures; quantum chips might need analogous solutions at the quantum level.
Possibility of Hybrid Approaches: It’s conceivable that the first fully practical quantum computers will combine different qubit technologies to harness the best of each. For example, one could use fast, densely integrated silicon spin qubits for computation, but connect clusters of them via photonic links for long-distance communication, or use more robust qubits (like trapped ions or superconducting qubits) as memory or ancilla qubits for error correction in combination with spin qubit arrays. Research is ongoing into hybrid systems such as coupling spins to photons (to convert stationary qubit states into “flying” qubits for networking). While the current breakthrough stands on its own in silicon, practical quantum computing might involve a heterogeneous approach – much as classical computing uses CPUs, GPUs, and specialized accelerators together. In the long run, each qubit type might play a role: superconducting circuits or ions in specialized high-fidelity roles, photonics for interconnections, and silicon spin qubits forming the dense computational core. Building such hybrid quantum architectures raises its own set of challenges (interface compatibility, complexity of multiple subsystems), but could be a way to circumvent the limitations of any single qubit technology.
In summary, the road from a two-qubit demonstration to a full-fledged quantum computer is still long and winding for hole-spin qubits. The encouraging news is that none of the challenges above appear to be fundamental show-stoppers – they are matters of engineering, refinement, and inventive design. As Dr. Kuhlmann noted, the latest research shows the principle is sound and “underscores that this approach has a strong chance in the race to develop a large-scale quantum computer.” Now, it becomes a multi-disciplinary effort: combining physics, materials science, and chip engineering to turn this promising approach into a competitive quantum computing platform.
Impact on Cryptography and Cybersecurity
Advances in quantum computing inevitably raise the question: Does this bring closer the day when quantum computers can crack our encryption algorithms? In the context of cryptography, the ultimate threat is a quantum computer that can run Shor’s algorithm to factor large numbers or compute discrete logarithms, thereby breaking RSA and ECC encryption. However, running Shor’s algorithm for real-world key sizes (such as 2048-bit RSA) would require a huge quantum computer – estimates typically say on the order of millions of physical qubits with error correction to succeed. Today’s devices are nowhere near that scale. The Basel/IBM result doesn’t immediately change that calculus, but it is aimed squarely at the scalability problem. By pointing to a path where qubits can be manufactured and integrated in great numbers (leveraging the semiconductor industry), this breakthrough could indeed accelerate the timeline to a cryptographically relevant quantum computer. If industrial fabrication of quantum chips takes off in the next few years and qubit counts start growing at rates similar to transistor counts in the early integrated circuits era, the hypothetical quantum threat to encryption might materialize sooner than previously anticipated.
It’s important to note that we are still talking years of development – no one is breaking RSA with a FinFET qubit pair anytime soon. But the strategic direction is clear, and that’s why governments and companies are already moving toward post-quantum cryptography (PQC). The understanding in the cybersecurity community is that it’s a matter of “when, not if” quantum attacks on today’s public-key cryptosystems become feasible. In anticipation of that “Q-day,” standards bodies like NIST have been soliciting and standardizing quantum-resistant encryption algorithms. In July 2022, for example, NIST announced the first batch of PQC algorithms (such as CRYSTALS-Kyber for key exchange) to be implemented in the coming years. The advancement of silicon-based qubits reinforces the urgency of those efforts – it suggests that a quantum computer capable of threatening classical encryption could emerge on the horizon as soon as the engineering challenges are ironed out. Organizations are therefore advised to pursue crypto agility: being ready to swap out vulnerable cryptographic algorithms for PQC alternatives well before large-scale quantum computers come online. The timeline for quantum computing remains uncertain, but developments like this underscore that the field is progressing steadily. From a cybersecurity perspective, the safe approach is to assume that breakthrough innovations will continue and to prepare now for future quantum capabilities, rather than be caught off-guard.
Future Outlook
The demonstration of hole-spin qubits in a silicon FinFET is a bellwether for where quantum computing is heading. Experts see it as an indicator that quantum hardware can evolve from laboratory curiosities to engineered systems leveraging the best of classical semiconductor know-how. What does this breakthrough mean for the future of quantum computing? In a nutshell, it suggests that quantum bits might eventually be as ubiquitous and scalable as the transistors on a microchip, which could revolutionize the field by enabling complex quantum processors with millions of qubits.
In the near term, the Basel and IBM researchers will be looking to build on this result. We can expect to see attempts to integrate more than two qubits in a single device — for instance, a 3- or 4-qubit FinFET quantum dot array, which would allow demonstration of small quantum algorithms or error-correction primitives. They also plan to continue improving qubit performance: lengthening coherence times via materials purification and optimized device geometry, and further reducing gate error rates. Another likely step is the integration of on-chip electronics. Since their qubits operate at 4 K, it’s feasible to put cryogenic classical logic (such as multiplexers, amplifiers, or even a control microprocessor) on the same chip or substrate. A prototype of a quantum System-on-Chip with both qubits and control circuitry would be a game-changing milestone, showing how to address the wiring and control complexity at scale.
From an industry standpoint, this breakthrough will spur investments and research in quantum silicon technology. We might see more collaborations between quantum computing firms and semiconductor foundries. Intel’s aggressive roadmap on spin qubits will likely be accelerated, and other companies could launch their own silicon-based quantum initiatives to not be left behind. Startups focusing on semiconductor qubits or hybrid semiconductor-superconducting approaches may get more attention and funding. In a broader sense, the quantum computing landscape could shift: superconducting and ion-trap systems have a head start and will continue to advance in the next few years, but if and when silicon qubit chips begin to demonstrate, say, tens of qubits with good control, the balance might tilt as the long-term scalability advantage becomes evident. As one report on the Basel experiment concluded, these silicon qubits are “highly scalable and have proven to be fast and robust in experiments,” giving them a strong footing in the race toward a large-scale machine.
It’s also conceivable that the future of quantum computing will not be a single-platform monopoly but rather a multifaceted ecosystem. Just as classical computing uses CPUs, GPUs, FPGAs, and more recently TPUs or AI accelerators in tandem, quantum computing might use different qubit types for different layers of the system. For example, a quantum data center might use superconducting qubits as dedicated high-speed coprocessors for certain tasks, while using CMOS-integrated spin qubits as the main memory and processing grid for large-scale operations, and photonic links to connect modules together. The Basel/IBM result increases the likelihood that silicon spin qubits will earn a place in that ecosystem – possibly as the platform that links everything together with the semiconductor industry’s fabrication capability.
Looking further out, the ultimate promise of this line of work is a fault-tolerant quantum computer that can solve classically intractable problems (in chemistry, material science, optimization, and beyond). To reach fault tolerance, error-correcting codes will need to be implemented, which might require hundreds of physical qubits per logical qubit. The only way that’s possible is if the physical qubits are extremely abundant and sufficiently reliable – exactly what the silicon FinFET approach aims to deliver. If one day we have chips with, say, 1 million spin qubits with each 1,000 forming a logical qubit, we could run very deep quantum circuits reliably. Achieving that will take perhaps a decade or more of dedicated effort, but the path has been illuminated by results like this. As Prof. Dominik Zumbühl from Basel’s team noted in earlier work, they chose FinFETs because “why not build a quantum computer with a platform that has successfully mastered” the challenge of scaling to billions of devices? That philosophy now seems to be paying off.
In conclusion, the successful marriage of hole-spin qubits with a silicon transistor is a profound validation of the vision of quantum computing that piggybacks on classical microelectronics. It provides a tangible blueprint for how to merge quantum bits into the fabric of classical chips. The next few years will be critical to watch: will we see rapid improvements and scaling in these silicon qubit systems, akin to a “Moore’s Law” for qubits? If so, quantum computing could transition from specialized lab setups to something that looks much more like today’s computing hardware, only imbued with quantum mechanical power. This breakthrough has essentially brought quantum computing one step closer to the world of mainstream technology – and in doing so, it has made the future of quantum computing both more exciting and more foreseeable. Each new development will bring challenges, but also carry the field further along the clear path now forming towards practical, large-scale quantum machines.
