Cyprus company develops software layer for Europe’s quantum ambitions

Superpositions Studio is building a quantum computing software platform from Limassol, aiming to help businesses determine whether the rapidly developing technology can deliver practical value without requiring them to employ specialist quantum scientists.

The company spoke exclusively to the Cyprus Mail about its origins in Cyprus, its work across finance and energy, and its ambition to develop Limassol into a centre for quantum software research.

Superpositions Studio was created to address what it described as a fundamental problem in the quantum computing industry, namely that companies are increasingly interested in the technology but often struggle to determine whether it can solve a particular business problem.

“Exactly that ‘do not yet know’ is the problem,” the company said.

More than $40 billion has been invested in quantum computers worldwide and more than €11 billion in Europe, according to the company, while the machines themselves are already operating.

Yet businesses considering quantum computing have traditionally faced a choice between hiring specialist staff, who are scarce and can cost more than $150,000 a year, or commissioning consultancy work costing around $50,000 and taking months to produce a report.

Superpositions Studio aims to turn that expert process into software.

Its platform allows a customer to describe a business problem in ordinary language before translating it into mathematical and quantum formulations, identifying suitable quantum and classical algorithms, running experiments and comparing the results.

The company said the platform can provide an answer in less than an hour on whether quantum computing is useful for a particular problem, where it may have an advantage and when that advantage could emerge, without the customer needing a quantum specialist.

Cyprus, and Limassol in particular, was chosen as the base for the company because the founders saw the island as a suitable environment for developing an international deep-tech business.

As an EU member state, Cyprus provides access to the European research and innovation ecosystem while also attracting international engineering, scientific and entrepreneurial talent.

The local innovation landscape, including support from organisations such as the Research and Innovation Foundation, also offers opportunities for deep-tech companies to develop and collaborate.

For Superpositions Studio, transparent benchmarking is central to its approach because the company believes quantum computing is currently caught between excessive optimism and scepticism.

“Because at this stage of the technology, honesty is the scarcest resource,” the company said.

The company argued that the reality is problem-specific, with quantum and hybrid approaches already producing measurable benefits in some industrial applications while classical computing remains more effective for others.

It therefore designed its platform so that every quantum experiment is accompanied by a classical baseline.

“The platform must run the best classical baseline alongside every quantum experiment, and the report says plainly which one won,” the company said.

In some cases, the most valuable conclusion for a customer may therefore be that its particular problem is not yet suitable for quantum computing.

“Sometimes the most valuable output is ‘your problem is not yet a fit for quantum’ and that answer saves a company from an expensive mistake,” the company said.

It added that such transparency could help establish the trust needed for customers to return when the technology becomes more suitable for their needs.

There is also a practical reason for the emphasis on reproducible results, with enterprise procurement processes and increasingly regulators requiring evidence that can be independently examined rather than relying on enthusiasm about the technology.

Every result produced through the platform comes with a reproducible and citable report, together with the underlying code.

The company’s approach starts with the business problem rather than quantum terminology.

A company looking to improve wind energy forecasting or detect fraudulent payments, for example, can describe its task to an agent and attach the relevant data.

The platform then uses a multi-agent artificial intelligence engine consisting of dozens of specialised AI agents with knowledge of quantum computing and industrial applications.

The system analyses the problem, translates it into an appropriate mathematical formulation and selects methods for solving it.

It then writes the required code, conducts experiments using high-performance simulators and, where appropriate, runs them on quantum processors.

The results are assessed against classical methods using measures including accuracy, speed and cost.

The customer receives a report identifying the most effective approach for the particular task and explaining how the result could change as quantum hardware improves.

The company also provides downloadable Python code, allowing a customer’s technical team to rerun, modify and integrate the solution.

For standard tasks, the complete process takes less than an hour.

Enterprises handling sensitive information can also operate the platform as a closed instance, meaning that their data does not leave their own environment.

Superpositions Studio has so far tested its platform across more than 20 industry use cases and 10 quantum algorithms.

Finance has emerged as one of its most active areas, which the company attributed to the combination of demanding computational problems, significant financial budgets and regulatory requirements.

In payment fraud detection, a hybrid quantum neural network achieved what the company described as good precision while being executed on real IBM quantum hardware.

In anti-money-laundering screening, the company said its hybrid model, working with heavily imbalanced data, outperformed classical approaches including random forests and gradient boosting.

For risk estimation, quantum amplitude estimation produced value-at-risk with less than 0.6 per cent error using 13 qubits.

Energy has provided another promising area, particularly through work on wind power forecasting.

In one example, a hybrid quantum neural network was trained and evaluated using a real industrial dataset containing more than 26,000 observations and was executed on IBM’s latest-generation quantum processors.

The model produced a prediction error of 7.1 per cent, compared with 7.3 per cent for a powerful classical neural network, while using approximately one-third fewer model parameters.

The company said this demonstrated that hybrid quantum components could potentially represent certain patterns more efficiently than classical layers.

“Fewer parameters mean less data needed, faster training, lower cost, less overfitting,” the company said.

It also pointed to the direct economic value of more accurate renewable energy forecasting, particularly as European electricity grids become increasingly reliant on renewable generation.

The company said the wider pattern across its tests was what encouraged it most, with hybrid quantum models repeatedly matching or outperforming classical models while using fewer parameters and less training data.

Superpositions Studio’s platform is deliberately hardware-agnostic, allowing experiments to be conducted across systems from IBM, IonQ, IQM and Rigetti.

The company considers this important because several quantum hardware technologies are developing simultaneously, with different strengths and without a clear indication of which approach will ultimately prevail.

“For a company assessing quantum today, hardware-agnosticism is the only responsible way in,” the company said.

It argued that committing to a single hardware provider would effectively force a business to make a technology bet at a time when the outcome of the wider hardware race remains uncertain.

The same customer code can instead run across 10 backends, with Superpositions Studio verifying that results match across them.

The company also sees a European dimension to its hardware strategy, with its integration roadmap including European quantum processors.

This would allow European companies to assess European hardware through the same independent platform rather than relying on a single provider.

“If the race leader changes, our customers don’t lose their investment, because they gain a new option in a dropdown menu,” the company said.

The company’s Early Access programme also shaped the platform before its move to general availability.

Superpositions Studio now has more than 150 users across 20 countries, giving it a broader understanding of what customers actually want from the technology.

One of the most important lessons was that the final report was as important to users as the experiment itself.

The company had initially expected users to spend most of their time building models, testing algorithms and examining different configurations.

Instead, many users were primarily interested in what they could take away from the process and use afterwards.

The programme also reinforced the value of interacting with the platform through ordinary language.

Many of the users who gained the most from Studio were not quantum specialists but experts in their own industries.

That experience strengthened the company’s belief that the platform should begin with a customer’s actual business problem rather than with quantum terminology.

Education has also become an important part of the company’s strategy.

Superpositions Studio is a technology partner in the European Business Institute of Luxembourg’s Q-Ready initiative, providing the hands-on platform used as the foundation of the programme.

Through the initiative, students can conduct quantum experiments and see direct comparisons between quantum and classical approaches.

The company said this was particularly important for younger people because understanding quantum computing cannot be achieved simply through presentations and theoretical material.

“I’d argue education isn’t adjacent to our mission, because it is our mission in another form,” the company said.

It pointed to the skills shortage as one of the biggest barriers to wider adoption of quantum technology, saying two-thirds of companies identify the skills gap as a major obstacle.

The company also sees education as a way of developing professionals who can properly assess quantum technologies rather than simply becoming users of individual hardware platforms.

Universities and educational institutions are therefore one of three pillars of its business model, alongside enterprises and equipment manufacturers.

The company said universities remain an important source of new algorithms and talent, while partnerships such as Q-Ready and its work with universities across Europe are intended to help develop a workforce capable of supporting Europe’s growing investment in quantum technology.

For Superpositions Studio, the long-term ambition extends beyond individual experiments and use cases.

“Our long-term ambition is simple to state, we want Superpositions to be the default software layer on quantum computers, helping businesses turn quantum hardware into measurable value in much the same way that operating systems made hardware useful and accessible to people,” the company said.

For its Limassol operation, that ambition means developing the Cyprus hub into a recognised centre for quantum software research, supported by a growing engineering and research team and stronger academic partnerships.

The company sees an opportunity for Cyprus precisely because the island does not need to compete with the large European countries that are investing in physical quantum infrastructure.

Europe has invested more than €11 billion in quantum hardware, while major machines are expected to be housed in large national laboratories.

Software, however, does not require the specialised physical infrastructure needed to build a quantum computer.

“Cyprus doesn’t need to build the quantum computer to matter in quantum computing,” the company said.

“It can build the layer that tells the rest of Europe what those computers are good for, that’s exactly what we’re doing from Limassol.”