Novo Holdings Says Quantum Investment Must Move Beyond Hardware

The majority of funding for quantum computing has been directed towards hardware, like processors, cryogenics, control electronics, and qubit components have absorbed the bulk of private capital. A report from Novo Holdings, summarized by The Quantum Insider on September 11, 2026, suggests that this concentration has left another critical area underfunded: the software and applications necessary to make quantum machines valuable to customers.

Novo Holdings’ Claims

Specifically, Novo Holdings, the investment company managing the Novo Nordisk Foundation’s assets, is a significant backer of life sciences. Indeed, the company states it has committed DKK 1.4 billion (approximately €188 million) to a quantum start-up ecosystem, anchored by its Seed Investments team, per its quantum investments page. A separate report from Quantum Computing Report headlines €100 million (about $145.5 million) as invested.

The report also highlights a key finding: around 70% of the approximately $13.9 billion invested in private quantum-computing companies between 2014 and 2025 went to hardware and components. This total excludes public-market deals and government programs, meaning the actual figure is higher. 

The report deviates from the usual practice of ranking hardware, software, and applications by which layer is most critical. Instead, it focuses on “control points”, or scarce capabilities that customers become dependent on and that a company can own. Technical difficulty alone does not qualify. Novo uses the semiconductor supply chain as an analogy: ASML in lithography, TSMC in fabrication, Arm in licensed architecture, and Nvidia in chips, software, and developers.

Applied to quantum, the report identifies domain-specific algorithms and vertical application platforms as the strongest positions for new entrants. 

These require less capital than hardware, are closer to the customer, and remain largely unclaimed because no company has yet deeply embedded itself. Life sciences serves as a prime example, as pharmaceutical firms already pay for computational chemistry and could benefit from improved electronic-structure calculations for specific, critical parts of molecules. 

Novo is also cautious here, not claiming biology is inherently a quantum problem and warning that a published algorithm does not create a competitive advantage. Currently, there is no quantum advantage in drug discovery, and the scientific case for advantage in ground-state chemistry remains debated, according to the report.

Application-Layer Investment in Practice

Fujitsu provides a concrete example of investment in the application layer. The company open-sourced OpenQARP under Apache 2.0, a package featuring over 100 components, algorithm building blocks, simulator support, and CUDA-Q integration. 

Fujitsu claims this tooling reduced one implementation from approximately 130 lines of code to fewer than 40. This claim is unverified, and a shorter codebase does not inherently demonstrate quantum advantage. Its stated plan to extend OpenQARP to its STAR architecture is an announced intention, not a delivered capability.

The Need for Systems Supporting Hardware

Supporting applications does not negate the need for foundational infrastructure. Altera and Riverlane, in a LinkedIn post from Altera, described a validated QEC Interface design example on Agilex 7 FPGAs for low-latency error-correction processing. Error correction, compilers, and runtimes may prove essential but could still fail to capture lasting value if hardware manufacturers or cloud platforms absorb them as features, a risk directly mentioned in the Novo report.

The scale of capital provides additional context. IonQ agreed to acquire SkyWater Technology to build a vertically integrated stack, per its investor release. This is a company-reported item reflecting one firm’s integration strategy, not an indication of application-segment revenue.

Policy and Private Capital

Public funding also addresses this gap. The EU Quantum Act prioritizes coordination, supply-chain resilience, industrial capacity, and skills, with connections to semiconductors, photonics, and cloud. The timeline for this act has shifted. 

A Greek Quantum Gate post from around May 2026 cited a Q2 2026 proposal window, and TQI’s September 4, 2026 reporting states the Commission expects an announcement by late 2026, with potential delays into 2027. No primary Commission text is yet available.

Evaluating Application Investment

The true test for application investment isn’t qubit counts or roadmap slides. Instead, it requires independent benchmarks, access to hardware or simulators, reproducible software, a clearly defined customer problem, specified error and resource requirements, total cost analysis, and evidence that the workflow outperforms classical alternatives. 

Novo’s own conclusion aligns with this: capital should be used to build reusable software, validation records, and proprietary data that accumulate value over time, as these take years to establish and the window to secure an application position narrows as workflows become embedded.

Invest in the machines, and the layer that transforms them into actionable decisions. Though the sources cited illustrate the existing gap, they do not yet definitively prove the eventual payoff.