The race to build useful quantum computers may be becoming a race to decide what a qubit should actually be made of.
IBM has spent years becoming one of the most recognizable names in quantum computing. Its machines use superconducting qubits, tiny electrical circuits cooled to temperatures extremely close to absolute zero. IBM has built increasingly large processors, developed quantum error-correction systems, and laid out an ambitious roadmap toward fault-tolerant quantum computing.
So its latest move is particularly interesting.
In July 2026, IBM announced plans to acquire HRL Laboratories, a research laboratory jointly owned by Boeing and General Motors. HRL works across several advanced technologies, but one area makes the acquisition especially important for quantum computing: silicon spin qubits.
A qubit is not one specific object. It is any controllable quantum system that can store quantum information. That means companies can build quantum computers using very different pieces of physics.
IBM and Google are best known for superconducting circuits. Quantinuum uses trapped ions. Companies such as QuEra work with neutral atoms. Photonic quantum computers manipulate particles of light. Microsoft has invested heavily in the possibility of topological qubits.
HRL is pursuing another approach: storing quantum information in the spin of electrons inside silicon devices.
That difference matters because the quantum industry still does not know which type of qubit will ultimately scale best.
Silicon spin qubits begin with something the technology world already understands extremely well: semiconductors.
Inside specially designed silicon structures, researchers can confine individual electrons in extremely small regions called quantum dots. An electron has a quantum property known as spin, and different spin states can be used to represent quantum information.
In a simplified picture, one spin state can represent (|0⟩), another can represent (|1⟩), and quantum mechanics allows combinations of the two.
The exciting part is not only that electrons can behave as qubits. It is that these qubits can potentially be fabricated using techniques related to the same semiconductor technology that has spent decades producing increasingly sophisticated computer chips.
Modern factories can manufacture chips containing billions of transistors with extraordinary precision. Quantum computing is nowhere near that level of manufacturing maturity.
Silicon spin qubits raise an obvious question: could some of the technology that enabled the classical computing revolution eventually help scale quantum computers as well?
HRL has already demonstrated substantial progress toward answering that question.
Its researchers have developed silicon devices containing dozens of quantum dots and demonstrated operations between encoded spin qubits. Instead of simply controlling isolated quantum states, researchers are beginning to build systems in which several physical components cooperate to represent and manipulate more reliable quantum information.
That connects directly with one of the biggest problems facing the entire quantum industry.
Building a few excellent qubits is difficult.
Building thousands, millions, or eventually even more physical components that behave consistently enough for fault-tolerant quantum computation is an entirely different engineering problem.
At that scale, quantum computing starts looking as much like semiconductor engineering and advanced manufacturing as theoretical physics.
IBM's acquisition does not mean the company is abandoning superconducting quantum computing.
Its current roadmap still centers on superconducting processors and large-scale quantum error correction. Instead, HRL gives IBM access to expertise in a very different hardware architecture.
That may become increasingly important because the future quantum computer does not necessarily need to contain only one type of quantum device.
Classical computers already work this way. A modern computer combines CPUs, GPUs, memory, networking hardware, storage, and specialized accelerators. Each component performs a different job.
Future quantum systems could eventually develop in a similar direction. One technology might be particularly effective for processing quantum information. Another might provide long-lived quantum memory. Different devices could handle communication, error correction, or connections between processors.
The final architecture of a useful quantum computer may therefore look much more complicated than a single giant chip filled with identical qubits.
This is also why today's competition between quantum companies is so unusual.
During the early history of classical computing, many technologies competed before semiconductor transistors became dominant. Quantum computing may currently be experiencing its own version of that period.
Superconducting qubits offer extremely fast operations and already power some of today's largest quantum processors. Trapped ions can maintain quantum states for long periods and achieve extremely accurate operations. Neutral atoms can naturally form large programmable arrays. Photons are attractive for communication. Spin qubits could benefit from their extremely small size and connection to semiconductor manufacturing.
Topological approaches attempt something even more ambitious: encoding information in quantum states that could be naturally protected from certain errors.
Each architecture solves some problems while introducing others.
There is still no quantum equivalent of the transistor: a technology that has clearly emerged as the standard building block for an entire industry.
That makes IBM's decision to acquire HRL more interesting than a normal corporate acquisition.
IBM is one of the companies most strongly associated with a particular quantum architecture, yet it is expanding its expertise into another major approach. Google has also begun exploring neutral-atom systems alongside its superconducting program. Across the industry, the boundaries between different approaches are becoming less rigid.
The biggest quantum competition may therefore not simply be about which company reaches the highest qubit count.
It may be about discovering which combination of physics, materials, manufacturing, error correction, and computer architecture can actually turn quantum mechanics into a scalable technology.
For students entering quantum computing today, that makes the field unusually open.
Some of the most important questions have not been answered yet.
What should a qubit be made from? How should millions of them be manufactured? Should future quantum computers combine several kinds of qubits? And will one architecture eventually dominate in the same way silicon transistors transformed classical computing?
IBM's move toward silicon spin qubits suggests that even the largest companies in quantum computing are still searching for those answers.
The race to build the quantum computer may ultimately begin with a much more fundamental question: What should we build it out of?