IBM's plans for a fault-tolerant quantum computer are moving from promises toward a tougher test: independent technical scrutiny.
The company has advanced to Stage C, the final phase of the Defense Advanced Research Projects Agency's (DARPA) Quantum Benchmarking Initiative. The program aims to determine whether competing quantum computing approaches can produce systems whose computational value exceeds their cost.
For IBM, the advancement brings its quantum roadmap closer to external evaluation as the company targets 2029 for its first fault-tolerant quantum computer. But reaching the final stage does not mean DARPA has confirmed that IBM's technology will work as planned.
Making progress toward useful quantum tech
DARPA launched the Quantum Benchmarking Initiative (QBI) to assess whether companies have credible paths toward utility-scale quantum computing.
Nearly 20 companies entered Stage A in 2025, when participants were asked to demonstrate the technical plausibility of their proposed approaches. In November 2025, DARPA selected 11 companies to enter Stage B of the QBI challenge. At this stage, the US agency was testing whether the participants had a credible development plan.
Stage C moves the evaluation closer to the technology itself. DARPA will scrutinize hardware and systems as they develop, assessing whether their performance matches the companies' technical models and roadmaps.
Micah Stoutimore, QBI Managing Director, said: “Stage C is where we begin moving from plans to proof. Our team will be closely scrutinizing the hardware and systems as they develop to determine whether the technology performs the way the models and roadmaps say it should.”
IBM is among the companies advancing to this stage. The company is developing IBM Quantum Starling, a fault-tolerant quantum computer it expects to deliver in 2029.
However, IBM has not explicitly identified Starling as the system it will use for DARPA's evaluation. The company's 2029 target also remains a projected milestone rather than a verified achievement.
What eWeek found: DARPA's quantum test moves from promises to proof
IBM's advancement to Stage C marks an important step, but it does not mean the company has solved the technical challenges of fault-tolerant quantum computing. Instead, it brings IBM's proposed approach into a more demanding phase of independent evaluation.
The reason why this distinction matters is because quantum computing companies are pursuing different technologies to overcome the same problem: building systems that can perform reliable calculations despite errors caused by noise and other disturbances.
IBM is developing modular superconducting processors, while IonQ uses trapped ions. Other Stage C participants are pursuing approaches based on neutral atoms and silicon spin qubits, according to DARPA.
Rather than favoring one architecture, DARPA's Quantum Benchmarking Initiative evaluates whether these different approaches offer credible paths to useful quantum computing. The agency's goal is to determine whether an industrially useful system, whose computational value exceeds its cost, can be built by 2033.
This makes Stage C particularly important. DARPA is moving beyond evaluating development plans to examining whether proposed systems can actually be built and operated as intended. The agency has also emphasized that the initiative is not designed to select a single winner.
The value of this process lies in the evidence it could produce. Independent evaluations may help distinguish approaches that can overcome their engineering challenges from those that still face significant barriers to scaling.
For enterprises, the results could eventually provide a more reliable basis for assessing quantum computing vendors and their long-term potential. But Stage C is not a guarantee that commercially useful, fault-tolerant systems are imminent.
The more meaningful milestone will be whether IBM and its competitors can demonstrate that their technologies work as intended and offer a credible path to useful computation at a justifiable cost.
Other news: IBM and Red Hat have fixed more than 400 previously unknown vulnerabilities in widely used Java libraries through their Lightwell initiative, which combines AI-assisted patching with automated testing and human validation.