From theory to impact: Real-world results in quantum machine learning

Quantum Machine Learning has now crossed from theory into practice, delivering measurable accuracy gains today through hybrid models on commercial quantum hardware.

Download PDF

When Classical Performance Plateaus

Classical machine learning excels at pattern recognition, but struggles as relationships become more nonlinear and feature interactions grow more intricate. In high stakes environments, even small performance gains can translate into outsized economic or operational impact—yet those gains are increasingly difficult to achieve with classical methods alone. Quantum machine learning addresses this limitation by introducing quantum dynamics that are infeasible to simulate classically, enabling richer feature representations that enhance downstream learning

As enterprise machine learning confronts increasingly complex, high‑dimensional data, purely classical models begin to reach performance limits. Gains plateau even as architectures grow more sophisticated.

This report highlights joint research from KPMG, IBM, and Kipu Quantum demonstrating how quantum‑enhanced machine learning can augment classical models now, without waiting for fault‑tolerant quantum computers. By combining proven classical feature extraction with quantum feature mapping on real hardware, hybrid workflows capture higher‑order relationships that classical systems alone cannot—driving materially improved predictions, classifications, and decisions.

Ready to move from theory to measurable results?

Access the full PDF to review the study and results demonstrating how quantum enhanced machine learning is delivering measurable performance gains on real hardware today.

Download now

What we've achieved here is a testament to the power of convergence. By combining KPMG's business-centric approach with Kipu's algorithmic ingenuity and IBM's state-of-the-art quantum systems, we've collectively pushed beyond the theoretical and set a new bar for what's possible in the era of quantum-enhanced machine learning.

Dr. Aaron Kemp

US Quantum Leader, KPMG LLP

Joint research shows that quantum enhanced feature extraction delivers consistent, reproducible accuracy improvements on real world datasets when executed on IBM quantum processors. Across multiple quantum devices and configurations, hybrid quantum–classical models outperform best in class classical baselines—demonstrating that quantum techniques are already delivering measurable value.

In select hardware scenarios, quantum only models exceeded both classical and hybrid performance, highlighting the standalone potential of quantum feature representations under specific conditions. In domains such as risk modeling, fraud detection, supply chain optimization, satellite analytics, and medical diagnostics, a 2–3% accuracy gain is not incremental—it is transformational, translating directly into material financial impact, reduced risk, and competitive differentiation.

What this validation demonstrates:

Validated on commercially available IBM quantum systems, not simulations or lab prototypes
Reproducible performance across different devices, qubit scales, and experimental runs

Quantum-enhanced gains persist across hardware scales

“Quantum-enhanced satellite image classification”, February 2026.

This approach enhances classical machine learning rather than replacing it. Classical models establish a strong baseline, while quantum feature mapping extends their reach into higher‑dimensional structure that classical systems struggle to capture.

The workflow delivers reproducible performance improvements today, while also enabling organizations to build hands‑on experience and readiness for future advances in quantum computing.

Classical feature extraction:

Proven deep learning models (e.g., ResNet architectures) extract compact, high value representations from complex, multi modal data—establishing a strong classical foundation.

Quantum feature mapping:

These classical features are encoded into quantum circuits using a Hamiltonian based method known as Digitized Quantum Feature Mapping (DQFM), where quantum dynamics reveal higher order relationships beyond classical reach.

Established models, quantum enhanced Inputs:

Quantum derived features are fed back into conventional classifiers, improving separability and overall accuracy while keeping the final learning and decision process unchanged.

Designed for high-complexity cross-industry data environments

While the featured research focuses on multi‑sensor satellite image classification, the underlying QML pattern is broadly applicable wherever classical models struggle with complex feature interactions:

Risk and fraud detection
Credit and financial risk analysis
Supply‑chain and logistics optimization
Remote sensing and geospatial intelligence
Medical imaging and diagnostics
Scientific research and materials discovery

Key takeaway:
Quantum machine learning is not a niche capability—it represents a generalizable enhancement layer for the most demanding machine‑learning problems.

Ready to move from theory to measurable results?

Access the full PDF to review the study and results demonstrating how quantum‑enhanced machine learning is delivering measurable performance gains on real hardware today.

Download now

About the research collaboration

This research was conducted through a collaboration between KPMG Quantum Research, IBM, and Kipu Quantum, combining enterprise problem framing, quantum algorithm design, and execution on leading quantum hardware platforms.

The work focuses on translating cutting‑edge quantum science into practical, enterprise‑ready machine‑learning applications—grounded in measurable outcomes rather than theoretical benchmarks.

All findings were developed and authored by human researchers using established classical and quantum machine learning methodologies.

Download the research paper

Proven performance gains, grounded in real hardware

Access the full study, “Quantum‑enhanced satellite image classification,” to explore the methodology, results, and analysis demonstrating how quantum‑enhanced machine learning delivers measurable accuracy improvements on real hardware today.

Learn more

	Demonstrated, reproducible performance The research shows a consistent 2–3% absolute accuracy gain over best in class classical models, validated across multiple hardware platforms and repeated experimental runs.
	Executed on commercial quantum hardware All results were achieved on commercially available IBM quantum systems—not simulations—confirming near term devices can deliver practical value.
	Hybrid by design Quantum systems enhance feature extraction while preserving existing classical ML pipelines, improving performance without wholesale replacement of proven models.
	Focused on high value use cases The approach targets domains where small accuracy gains drive disproportionate business impact, such as risk analysis, fraud detection, medical diagnostics, and supply chain optimization.
	Translating results into capability By working with real hardware and hybrid workflows, organizations can build quantum capability incrementally—gaining operational insight while delivering value today.

Looking ahead:

Why act now: A practical path to quantum readiness

Quantum enhanced machine learning is no longer hypothetical. The research shows that near term quantum devices can already deliver additive value when integrated thoughtfully into classical ML pipelines.

Organizations that begin experimenting now can:

Build internal expertise ahead of competitors

Capture incremental performance gains sooner

Inform long term quantum strategies with real operating data

KPMG works with organizations to translate quantum machine learning research into practical experimentation, grounded in real‑world results rather than theory or simulation. Drawing directly from collaborative research with IBM and Kipu Quantum, we focus on deploying hybrid quantum–classical approaches where classical methods begin to plateau.

We partner with organizations to:

Identify high‑complexity machine‑learning problems where incremental accuracy gains can drive material business value
Design and pilot hybrid quantum–classical workflows validated on real commercially available quantum systems
Evaluate performance, reproducibility, and scalability to inform next‑step investment decisions

This work is not about replacing classical computing, but enhancing it—using quantum techniques to uncover higher‑dimensional patterns that improve model performance today while building organizational readiness for future advances.

The journey to quantum value is iterative—but the first, validated step has already been taken.

About IBM and KPMG

Empowering clients with the power of AI

As organizations advance their transformation agendas and embrace technologies to remain competitive, it becomes imperative for them to unlock value. A strategic collaboration can help improve efficiency, develop insights from the available ecosystems, and lower costs. However, the search for one that understands the intricacies of the diverse technologies at play can be challenging.

Through a client-first approach, KPMG LLP and IBM bring innovative offerings to market across Hybrid Cloud, Data, AI, and other emerging technologies, including quantum computing and post-quantum cryptography, that help clients accelerate digital transformation. The regulatory, risk, and finance experience of KPMG, combined with the robust consultancy and technology competencies of IBM, can help clients address challenges with knowledge-driven and actionable solutions.

Insight

Webcast Replay Webcast Upcoming Listen Now From The Web

Quantum Dawn

Are you prepared to turn quantum disruption into competitive advantage? Quantum computing is rewriting the rules of security and organizations must act now to protect data before today's encryption becomes obsolete.

Insight

Webcast Replay Webcast Upcoming Listen Now From The Web

Pushing Boundaries: How to lead with Edge AI computing A collaboration between IBM and KPMG ushers in a new era of rapid, secure, sustainable AI solutions

Is your enterprise ready to move AI from the data center to the point of action? Discover how agentic, edge native architectures deliver sub second decisions, stronger privacy, and measurable value—without compromising trust. With IBM and KPMG, AI runs at the edge for real time insight, resilience, and secure, scalable modernization.

Insight

Webcast Replay Webcast Upcoming Listen Now From The Web

Computing infrastructure in a new era

Is your computing infrastructure ready to fuel business growth in the AI era? Explore how emerging tech and flexible, AI-powered solutions can transform digital operations into strategic growth drivers for organizations.

Insight

Webcast Replay Webcast Upcoming Listen Now From The Web

KPMG 2025 Futures Report

Seizing opportunities in an era of disruption

Insight

Webcast Replay Webcast Upcoming Listen Now From The Web

AI Insights

Discover leading research, analysis, and perspectives to help business leaders confidently navigate AI trends and priorities.

Dr. Aaron Kemp

US Quantum Leader, KPMG LLP

Read bio

Richard Entrup

Managing Director and Head of Emerging Solutions, Enterprise Innovation, KPMG US

Read bio

Dr. Aaron Kemp

US Quantum Leader, KPMG LLP

Read bio

Richard Entrup

Managing Director and Head of Emerging Solutions, Enterprise Innovation, KPMG US

Read bio

Relevant Results

From theory to impact: Real-world results in quantum machine learning

Higher‑dimensional insights, measurable gains

From theory to impact: Real-world results in quantum machine learning.

Quantum performance, proven in practice

Not simulation. Not theory. Measured results.

How hybrid quantum workflows deliver value

Extending classical models with quantum feature mapping

Where quantum machine learning creates value:

From theory to impact: Real-world results in quantum machine learning.

Download the research paper

Why this quantum approach is different

Built for measurable impact, not experimentation

Looking ahead:

How KPMG can help

From validated research to practical application: Building for the quantum future

Our latest Insights

Get in touch

Contact KPMG

Headline