Quantum Computing Inc. Announces Dirac-3S, Scaling to Nearly 10,000 Variables

Hardware, manufacturing and supply-chain changes are designed to support increased production volumes and broader commercial deployment.

Rhea-AI Summary

Quantum Computing (QUBT) announced Dirac-3S, its next-generation quantum optimization machine, with the first customer shipment expected in November 2026.

The system demonstrated configurations supporting up to 9,980 variables with one Expansion Module, compared with the current system's 1,000-variable capacity. Its modular design accommodates additional Expansion Modules. QCi said its benchmark testing found optimal solutions across all tested synthetic problems with known global optima ranging from 2,000 to 9,980 variables. The company also advanced its hardware architecture, manufacturing processes and supply chain to support increased production volumes and broader commercial deployment.

Positive

• Minor point. Forward-looking: it has not happened yet and may not happen.Manufacturing and supply-chain upgrades support increased production volumes and broader commercial deployment.

Negative

• None.

Key Figures

Demonstrated problem size: 9,980 variables Capacity increase: 10x First customer shipment: November 2026 +1 more

Demonstrated problem size

9,980 variables

Dirac-3S optimization problems

Capacity increase

10x

Compared with QCi's current 1,000-variable system

First customer shipment

November 2026

Expected shipment of Dirac-3S

Synthetic benchmark range

2,000 to 9,980 variables

Dirac-3S achieved the optimal solution across all tested instances

Historical Context

2 past events · Latest: Aug 10

2 events

1. Aug 10

   platform delivery

   24h Move

   +0.2%

   QCi reported delivering a Dirac-3 system to a global consulting firm.

2. Sep 14

   university agreement

   24h Move

   -0.3%

   HBKU agreement included cloud access to Dirac-3 and installation of a platform in Qatar.

24h Move is the share-price change in the day after each event; other market factors may also have contributed.

Key Terms

np-hard, simplex formulations, maximum-clique, integer optimization

4 terms

np-hard technical

"how NP-hard problems can be mapped to simplex formulations"

A problem is NP-hard if solving it efficiently (in polynomial time) would let you solve every problem in NP efficiently; formally, every NP problem can be transformed into it by a polynomial-time reduction. NP-hard problems are at least as difficult as the hardest problems in NP, but they need not themselves be decision problems or even belong to NP, and they may lack a known polynomial-time solution.

simplex formulations technical

"mapped to simplex formulations native to Dirac-3S"

Pharmaceutical products or preparations whose active ingredients and dosage forms are intended specifically to prevent, treat, or reduce symptoms of infections caused by herpes simplex viruses (commonly HSV‑1 and HSV‑2). The phrase covers different routes and forms—topical creams, oral tablets, injectables, or vaccines—and refers to the intended viral target and therapeutic goal rather than any single drug or formulation. It does not, by itself, specify the active compound, route of administration, regulatory status, or whether the product is therapeutic or prophylactic.

maximum-clique technical

"industry-standard DIMACS maximum-clique benchmarks"

A maximum-clique in a graph is a largest possible set of nodes such that every pair of nodes in that set is directly connected to each other; equivalently, it is a complete subgraph of greatest size. There can be more than one maximum-clique of the same size in a graph, and finding a maximum-clique is a computationally hard problem (NP‑hard), which is different from a maximal clique (a clique that cannot be enlarged but is not necessarily the largest).

integer optimization technical

"including continuous, integer and higher-order optimization"

An optimization problem in which one or more decision variables are required to take integer values (often 0/1 for yes/no choices). It combines objective functions and constraints like other optimization problems but adds integrality requirements that make the problem discrete and typically much harder to solve; standard solution methods include branch-and-bound, cutting planes, and integer-specific variants of linear or nonlinear programming. A common practical approach is to solve a relaxed (continuous) version to get bounds and then enforce integrality to find feasible, optimal or near‑optimal integer solutions; the discrete nature leads to combinatorial complexity as problem size grows.

AI-generated analysis. How Rhea-AI works. Not financial advice.

See more from StockTitan in Google Search and AI answers. Adds StockTitan as a preferred source · opens Google

Add on Google Not now

• Scales quantum optimization to up to nearly 10,000 variables, a 10x increase over QCi's current 1,000-variable system, while delivering faster performance, production-ready reliability and scalable manufacturing
• First customer shipment of Dirac-3S upgrade expected in Q4 2026

HOBOKEN, N.J., Oct. 1, 2026 /PRNewswire/ -- Quantum Computing Inc. ("QCi" or the "Company") (Nasdaq: QUBT), a vertically integrated quantum systems company pioneering photonics and semiconductor manufacturing, today announced Dirac-3S, the next-generation of its Dirac-3 quantum optimization machine, marking a significant step toward commercial deployment of QCi's quantum optimization technology.

Dirac-3S advances the platform across three areas – Speed, Scale and Solutions – with faster performance, improved solution quality, increased computing capacity and a modular architecture designed to support customers as their requirements grow.  The first customer shipment of Dirac-3S is expected in November 2026.

Speed. Scale. Solutions.

Speed: Dirac-3S is engineered to efficiently solve increasingly complex optimization problems, with practical computation times demonstrated on problems involving up to 9,980 variables while consistently finding the optimal solution. This enables customers to tackle larger problems and explore more possibilities in less time.

Scale: Dirac-3S features a modular architecture designed to scale with customer requirements. The compact 5U base system can be extended with Expansion Modules to support larger and more complex optimization problems, with QCi demonstrating configurations supporting up to 9,980 variables with one Expansion Module. The architecture is designed to accommodate additional Expansion Modules, enabling customers to expand their system as their optimization needs grow. QCi has also advanced the Dirac-3S hardware architecture, manufacturing processes and supply chain to support increased production volumes, consistent product quality and broader commercial deployment. These improvements include enhanced error correction, component control and operational reliability.

Solutions: Dirac-3S is designed to address a broad range of real-world optimization problems on a single platform, through multiple optimization approaches including continuous, integer and higher-order optimization. QCi's latest error-correction techniques are designed to improve solution quality across applications in financial services, logistics and supply chains, manufacturing, energy, telecommunications, scientific research, and aerospace and defense.

The platform can be deployed on-premises or in the cloud and integrated into AI/ML pipelines, enabling organizations to incorporate quantum optimization into existing computational workflows.

"Dirac-3S takes our quantum optimization platform from demonstrating what is possible to delivering a system built for practical commercial use," said Yong Meng Sua, Chief Technology Officer of QCi. "We have increased performance and computing capacity, expanded the size and complexity of problems the system can address through a modular architecture, and advanced the hardware and manufacturing architecture needed to support broader deployment. This gives customers a flexible path to adopt quantum optimization and scale their computing capacity as their needs grow."

Building on the Dirac-3 Foundation

Dirac-3S builds on QCi's first-generation Dirac-3 quantum optimization machine, which has been used by academic researchers exploring quantum optimization algorithms and applications.

"Our experience with the first-generation Dirac-3 quantum optimization machine validated its potential as a powerful tool for solving meaningful optimization problems. QCi's continued innovation with the Dirac-3S, including its expanded computational capabilities, enables us to pursue significantly larger and more complex optimization challenges. We look forward to exploring new research directions and accelerating the development of practical, high-impact applications," said Professor Paul Griffin, Associate Professor, Singapore Management University.

QCi has published a new Technical Review evaluating Dirac-3S across synthetic, graph-based and industry-standard optimization benchmarks, including comparisons with Projected Gradient Descent and commercial classical solver Hexaly. On synthetic problems with known global optima ranging from 2,000 to 9,980 variables, Dirac-3S was the only evaluated solver to achieve the optimal solution across all tested instances, while solution times increased only modestly as problem size approached 9,980 variables. On industry-standard DIMACS maximum-clique benchmarks, Dirac-3S more frequently identified the highest-quality solutions than the other evaluated methods.

QCi has also published a survey paper demonstrating how NP-hard problems can be mapped to simplex formulations native to Dirac-3S, expanding its potential applications across continuous, discrete and combinatorial optimization. The Technical Review, along with the detailed benchmarking methodology and results is available in QCi's Dirac-3S Technical Reference at: https://quantumcomputinginc.com/learn/module/dirac-3s-technical-reference.

To learn more about the Dirac-3S, visit QCi's Dirac-3S webpage.

Quantum Computing Inc. (Nasdaq: QUBT) is a vertically integrated quantum systems company pioneering photonics and semiconductor manufacturing, and delivering accessible, scalable, and cost-effective quantum machines, photonics products, and advanced packaging. The Company provides foundry services for photonic chips and semiconductor manufacturing and offers a vertically integrated portfolio spanning photonics and electronic components, subsystems, and full-stack systems.

Designed to operate at room-temperature with low-power requirements, QCi's technologies enable practical deployment across high-growth markets, including high-performance computing, artificial intelligence, cybersecurity, aerospace and defense, and advanced sensing and imaging.

Headquartered in Hoboken, New Jersey, QCi also has operations in Arizona, California, Illinois, Indiana, Massachusetts, North Carolina and Virginia. By combining advanced materials, device engineering, and scalable manufacturing, QCi delivers integrated quantum, photonics, and semiconductor technologies, accelerating commercialization and real-world adoption.

Company Contacts:

Investor Relations
David Moskowtiz, Head of Investor Relations
investors@quantumcomputinginc.com

Media & Communications
Karen Jeffers, Director, Marketing and Communications
media@quantumcomputinginc.com

Agency Contact:

IMS Investor Relations
John Nesbett/Zach Nevas
qci@imsinvestorrelations.com

Forward-Looking Statements

This press release contains forward-looking statements as defined within Section 27A of the Securities Act of 1933, as amended, and Section 21E of the Securities Exchange Act of 1934, as amended. These forward-looking statements and forecasts, generally identified by terms such as "may," "will," "expect," "believe," "anticipate," "estimate," "enhance," "intends," "goal," "objective," "seek," "attempt," "aim to," or variations of these or similar words, involve risks and uncertainties because they relate to events and depend on circumstances that will occur in the future. Those statements include statements regarding the intent, belief, or current expectations of QCi and members of its management as well as the assumptions on which such statements are based.  Any such forward-looking statements are not guarantees of future performance and involve risks and uncertainties, including the timing of the first delivery of a Dirac-3S, the ability of the  Dirac-3S to solve larger and more complex problems than the first generation of Dirac-3, deliver higher quality solutions and faster performance, the functionality of the Dirac-3S modular design, and delivering improvements in the performance, system maturity and manufacturability, and that actual results may differ materially from those contemplated by such forward-looking statements. Except as required by federal securities law, QCi undertakes no obligation to update or revise forward- looking statements to reflect changed conditions.

View original content to download multimedia:https://www.prnewswire.com/news-releases/quantum-computing-inc-announces-dirac-3s-scaling-to-nearly-10-000-variables-302895913.html

SOURCE Quantum Computing Inc.

FAQ

AI-generated questions and answers. How Rhea-AI works. Not financial advice.

When does Quantum Computing expect to ship Dirac-3S to its first customer?

Quantum Computing expects the first customer shipment of Dirac-3S in November 2026.

How many variables can Quantum Computing's Dirac-3S support?

Dirac-3S demonstrated configurations supporting up to 9,980 variables with one Expansion Module. The current system supports 1,000 variables, and the new architecture is designed to accommodate additional Expansion Modules.