AIThis post was created with the assistance of artificial intelligence (AI).

TL;DR

Buying for a business?Offer from Amazon

Get business pricing on monitors, keyboards and dev gear

  • Business-only prices and quantity discounts
  • Tax-exempt purchasing
  • Multiple users, one account, clear invoices
As an affiliate, we earn on qualifying purchases.

IBM has been selected for Stage C of DARPA’s Quantum Benchmarking Initiative, moving from planning reviews to assessment by the program’s independent verification team. IBM says it expects to deliver its Quantum Starling fault-tolerant system in 2029; DARPA’s broader goal is to assess whether a useful, cost-effective fault-tolerant quantum computer can be built by 2033.

IBM has been selected for Stage C of the U.S. Defense Advanced Research Projects Agency’s Quantum Benchmarking Initiative, moving the company’s fault-tolerant quantum computing plans into independent technical evaluation. The phase shifts the review beyond proposals and research plans toward testing hardware and assessing whether the company’s approach could support a fault-tolerant system.

IBM said the DARPA initiative’s independent verification and validation team will assess its quantum hardware, technology, systems engineering and architectures. IBM described Stage C as a testing phase. The announcement does not specify which components will be tested, how long the evaluation will take, or when DARPA will publish results.

The company says it expects to deliver IBM Quantum Starling in 2029, describing it as a fault-tolerant quantum computer. IBM says the system will combine work on error correction, processor design and systems engineering. These dates and technical plans are IBM’s projections; selection for Stage C is not itself a demonstration that Starling has achieved fault tolerance or that it will meet the stated schedule.

IBM said it had advanced through the initiative’s first two stages, which involved an initial technical concept and a more detailed research and development plan, including risks and mitigation strategies. The company also points to prior work on hardware components and error-correction decoding, but the announcement does not provide independent test results or metrics for those efforts.

At a glance
announcementWhen: Announced October 7, 2026
The developmentIBM announced its selection for Stage C of DARPA’s Quantum Benchmarking Initiative, where independent experts will evaluate company hardware and technical approaches.

Independent Testing Begins for IBM

Stage C places IBM’s proposal in a different kind of review: independent evaluation of hardware and engineering, rather than assessment limited to the company’s plans. That could give DARPA and other observers a more evidence-based view of the technical challenges and the plausibility of IBM’s route toward fault tolerance.

The initiative’s central question is not simply whether a quantum processor can run calculations. DARPA says it is assessing whether an industrially useful fault-tolerant computer can be built by 2033, with computational value that exceeds the system’s cost. For researchers, potential users and policymakers, the outcome could inform expectations about the time and resources needed to move from today’s quantum systems to more capable machines.

Selection does not establish that IBM has met that goal. It marks a milestone in a staged review, and the program’s eventual findings will matter more than the advancement alone. Results could help distinguish demonstrated capabilities from projected performance and identify the remaining technical and economic hurdles.

Amazon

quantum computing hardware

As an affiliate, we earn on qualifying purchases.

As an affiliate, we earn on qualifying purchases.

How DARPA’s Review Is Structured

DARPA launched the Quantum Benchmarking Initiative in 2024 to examine whether a fault-tolerant quantum computer with practical computational value could be built by 2033. In the program’s framing, the system must provide computational value greater than its cost. The initiative is intended to benchmark approaches, rather than certify a commercial product or guarantee a particular company’s timeline.

As described by IBM, Stage A asked participants for an initial technical concept with a plausible path to a cost-effective fault-tolerant computer. Stage B required a comprehensive research and development plan, with risks and proposed ways to reduce them. Stage C adds independent verification and hardware testing, taking the assessment from plans toward technical evidence.

IBM’s public roadmap includes Quantum Starling, which it said last year it aims to deliver in 2029. The company says it has continued work on components and error-correction methods. Those developments provide context for its selection, but DARPA’s assessment is separate from IBM’s own account of its progress.

“Advancing to the third stage of DARPA’s Quantum Benchmarking Initiative validates the strength of IBM’s path to building a fault-tolerant quantum computer.”

— Jay Gambetta, director of IBM Research and an IBM Fellow

Amazon

fault-tolerant quantum computer

As an affiliate, we earn on qualifying purchases.

As an affiliate, we earn on qualifying purchases.

Test Scope and Results Await

The announcement does not detail the test procedures, evaluation schedule or publication plans for Stage C. It also does not disclose specific hardware measurements, thresholds, or what evidence DARPA would require to judge a system fault-tolerant or industrially useful.

It remains unclear whether DARPA will release findings for individual participants, how Stage C results will affect later program decisions, or whether IBM’s 2029 target will change as testing proceeds. IBM’s selection confirms its advancement in the initiative, but it does not confirm that Starling is fault-tolerant, that its projected performance has been independently verified, or that the 2033 program goal will be met.

Amazon

quantum error correction systems

As an affiliate, we earn on qualifying purchases.

As an affiliate, we earn on qualifying purchases.

DARPA’s Hardware Evaluation

The next stated step is for DARPA’s independent verification and validation team to evaluate participating companies’ hardware. IBM says it will work with the team as the review moves into testing. The agency has not announced a timetable for completing Stage C or publishing outcomes in the information provided.

Readers should watch for details on what hardware is tested, what criteria are used, and whether DARPA reports comparative findings. Separately, IBM’s 2029 Quantum Starling target remains a company roadmap milestone; subsequent technical disclosures and independent results will show how its development is progressing.

Amazon

IBM Quantum Starling

As an affiliate, we earn on qualifying purchases.

As an affiliate, we earn on qualifying purchases.

Key Questions

What does IBM’s selection for Stage C mean?

It means IBM has advanced to the phase where DARPA’s independent verification and validation team is expected to evaluate hardware and technical approaches. It does not mean the system has passed that evaluation.

What is DARPA trying to determine?

The Quantum Benchmarking Initiative aims to assess whether an industrially useful fault-tolerant quantum computer can be built by 2033, with computational value greater than its cost.

When does IBM expect to deliver Quantum Starling?

IBM says it expects to deliver its Quantum Starling system in 2029. That is the company’s stated target, not a date independently confirmed by DARPA.

Has DARPA confirmed that IBM’s system is fault-tolerant?

No such confirmation is included in the announcement. Stage C begins an independent evaluation; the results and testing criteria have not been disclosed in the source material.

Source: rss

FALL

Fall Picks

As an affiliate, we earn on qualifying purchases.

You May Also Like

Anomaly Detection Techniques in Data Mining

Understanding anomaly detection techniques in data mining unlocks powerful insights, but exploring their nuances is essential for effective implementation.

A Skill Is a Folder, Not a Prompt: What Anthropic Learned Running Hundreds of Them

Anthropic reveals that effective AI Skills are folders containing instructions, scripts, and assets, transforming ad-hoc prompts into durable organizational assets.

Propensity Score Matching: What It Fixes and What It Can’t

Not all biases are eliminated by propensity score matching, so understanding its limitations is crucial for accurate causal inference—discover why next.

Gibbs Sampling and Metropolis-Hastings Explained

Gibbs Sampling and Metropolis-Hastings are powerful algorithms for sampling complex distributions, and understanding their differences can transform your approach to Bayesian inference.