Making quantum computing a reality

Dr. Axmann, could you tell us a little about the DLR Quantum Computing Initiative?

Dr. Robert Axmann: The initiative was originally launched by the Federal Ministry for Economic Affairs and Climate Action (BMWK). It commissioned DLR to promote the development, construction and use of quantum computers together with industrial partners. We started this in 2021 and have since established two centers in Hamburg and Ulm. We invite tenders for competitive procedures for the construction of quantum computers and the development of software. Together with the contractors, we pursue various technological approaches for quantum computers with different specifications and the development of different technology platforms. At the same time, we consider software and applications, in some cases in DLR-internal projects and in other cases in projects together with industrial partners. Once we have the computers, we will also use them in operations within DLR.

What is the mission of your start-up planqc, Dr. Glätzle?

Dr. Alexander Glätzle: Our mission is to build quantum computers: full-stack computers – from the hardware and the qubits to the quantum algorithms that we develop in co-design. The major goal is to solve one of the currently unsolvable trillion-dollar problems in materials science, pharmaceutical development or climate research that cannot yet be calculated even with the best available supercomputers. To achieve this, we are using neutral atom technology, which is based on decades of cutting-edge research at the Max Planck Institute of Quantum Optics in Garching – and we also receive proactive support from the photonics industry in Germany. Our mission also includes proving that technology transfer from science to industry can succeed. The aim is to create globally competitive quantum computers “Made in Germany”.

What role do start-ups play in the project to provide DLR institutes and your external partners with access to quantum computers?

Axmann: We invite tenders on a competitive basis – and we have been excited to discover that most of the bids come from start-ups. We rarely receive bids from traditional large-scale industry; so far, ideas and offers from this area have not come out on top. Bids are more likely to come from spin-offs from research institutes that can draw on their academic background, that have a very high scientific and technical quality and that are focused on their ideas. I actually see start-ups as the driving force behind the construction of quantum computers and the corresponding supply chains. Whether they are developing ion trap or neutral atom quantum computers – one thing all these companies have in common is their expertise, their outstanding teams and their access to the latest research results and scientific advice.

How do you see the role of start-ups in this young market and in the DLR network, Mr. Glätzle?

Glätzle: Compared to large companies or academic institutes, start-ups are agile, fast and highly focused – like speedboats compared to supertankers. This is probably also the reason why they are so strongly represented in DLR tenders. The essential building blocks for quantum computers have already been demonstrated in an academic context. We know that the approaches work on a laboratory scale and, as a young company, we are starting to translate them into market-ready computers for industrial applications. To achieve this, we need to increase technological maturity, miniaturize the systems and design them for low maintenance. As start-ups, we can concentrate fully on these tasks. We also have the advantage that the Max Planck Institute of Quantum Optics is behind us and is constantly providing new, important findings in its basic research, which are quickly incorporated into our development.

What is the situation with funding?

Glätzle: Start-ups begin with a big dream and little cash. We sought venture capital and acquired three very renowned European investors – UVC, Speedinvest and APEX Amadeus. The orders from DLR came at just the right time to further finance the development of our technology platform, generate revenue and continue developing our internal structures and processes. Not only did we win the contract to build Europe’s first digital neutral atom quantum computer, we were also successful in several DLR tenders for algorithm projects. We also gained maturity quickly during the implementation with the customer and its partners. It is important to establish sustainable structures and supply chains: we need first-class lasers, optics and vacuum technology which are, in principle, available in Germany. However, we have to work closely with suppliers to ensure that their components meet the necessary requirements.

Axmann: We shouldn’t expect miracles with quantum computing. Development takes longer than one legislative period, ties up a lot of capital and constantly requires new bright minds. This makes it difficult for start-ups. The BMWK has therefore commissioned us to put large contracts out to tender that offer the teams a long-term perspective. We pay when they reach agreed milestones.

The inventor of the laser, Theodore Harold Maiman, initially spoke of a “solution in search of a problem”. Does this also apply to today’s quantum computers?

Axmann: The theoretical foundations have been laid; it is now a matter of building market-ready solutions on this basis through good engineering. The focus here is on issues such as laser stability and the quality of the optics. Unlike in the early days of lasers, we now focus on applications for quantum computing from the outset. Complex climate modeling, highly flexible production planning, simulation of complicated reactions in the chemical and pharmaceutical industries, improved algorithms for search engines as well as Shor’s algorithm and the protection of cryptographic systems – there are many fields of application in which quantum computers promise major advances compared to conventional computers thanks to their performance. What’s more, it has been mathematically proven that quantum computers are in some cases exponentially superior to conventional computers.

Glätzle: Quantum computers are superior wherever quanta occur; for example, in simulations of molecules with their natural quantum effects. For instance, a single caffeine molecule has more than 10⁴⁰ possible different quantum states. To store this information, every atom on earth would have to be used as a classical bit. Therefore, problems of quantum chemistry and quantum pharmaceuticals cannot be solved accurately with conventional computers. The question of how exactly two molecules join together remains unanswered, as do numerous challenges in personalized medicine and materials research. Quantum computers could create real added value here – as soon as the hardware is ready. If I may make one comment in passing: I don't think Maiman could have imagined that lasers would be installed in supermarket checkouts or CD players today. Even today, we cannot begin to imagine which applications for quantum computers are yet to emerge. Because we inevitably think in terms of today’s possibilities.

Do you see quantum computers in large data centers with access via the cloud? Or are decentralized quantum computers in satellites, airplanes, automobiles or factories conceivable in the future?

Axmann: It’s too early to say because it’s unclear how the technologies will develop. The complex vacuum technology and cooling are still factors that would make decentralized use in aircraft, ships or satellites difficult or impossible. Therefore, I see the first systems as a supplementary solution for very high computing requirements in the cloud. But that may change because there are approaches at all levels to miniaturize the systems and make them less complicated. This would be interesting for space travel, for example, in order to immediately evaluate the huge volumes of data generated by satellites.

Glätzle: I also see quantum computers as powerful, accelerating co-processors in data centers. The current focus is therefore on investigating synergies with conventional mainframes. We are involved in corresponding projects funded by the Federal Ministry of Education and Research and are currently building a quantum computer for the Leibniz Supercomputing Center, which operates the SuperMUC NG supercomputer. In the near future, our quantum computers will run in data centers – and we ourselves will be able to offer customers from science and industry computing time via cloud access. Satellites are an interesting prospect for quantum computers to process data directly where it is generated. Who could have imagined a few decades ago that today every GPS satellite would have an integrated atomic clock? But it will be a long time before I can say the same about quantum computers.

How mature is the hardware? Is construction of the DLR quantum computer infrastructure progressing well?

Axmann: We successfully accepted the first two quantum computers at the beginning of July. It is by no means a walk in the park for the companies that develop the hardware. The challenges surrounding the number of qubits, gate fidelities, gates and stability as well as the challenges for the software are high. Especially as it involves the transition from experimental laboratory systems to CE-compliant, low-maintenance products. I am convinced that many companies will make it this far – but not all of them. With the increasing specifications in the next expansion stages, some approaches will reach their limits. But that is precisely the point: our initiative is investigating which quantum computing technology platforms are viable and scalable. The race is open, and it is in full swing worldwide.

What is planqc's technological approach to building the computer – and what role do lasers and optical technologies play in this?

Glätzle: IBM and Google rely on superconducting circuits for their quantum computers. We use a different technology, which we believe has many advantages. Above all, it allows us to better scale the number and quality of qubits, which is one of the key prerequisites for fault-tolerant quantum computers. Specifically, we use individual neutral atoms as qubits in which we can store and manipulate quantum information. The atoms are created by nature and are all completely identical. It is therefore a very coherent system – and the coherence times of the qubits, in which calculations are possible, are in the range of seconds. This is an eternity in the quantum world. The central component of our neutral atom quantum computer is an ultra-high vacuum chamber in which the atoms are trapped and enter a state of immobility at temperatures around absolute zero. The atoms – or qubits – are cooled, trapped and manipulated using lasers. In order to manipulate the individual atoms for calculations and change their logical state, we need low-noise lasers with precise frequency and wavelength. This is the only way to perform computing operations with high fidelity. In addition, we need high-end optics to be able to focus the atoms, which are arranged a few micrometers apart, individually by laser. Otherwise, there would be undesirable interactions with neighboring atoms. Optical switches are also needed so that we can control the atoms one nanosecond at a time. Our system works without cryogenic cooling at room temperature and is therefore more scalable and energy-efficient, but also less complex than systems based on superconducting circuits, for example.

At present there is a lot of hype associated with quantum computing and artificial intelligence. Are there real similarities and are visions of quantum machine learning realistic?

Axmann: Yes, we are pursuing this topic in various projects. These include Mat-QML, a sub-project of our materials research project QuantiCoM; the aim is to use quantum machine learning to predict the material properties of alloys, such as tensile strength or electrical and thermal conductivity. Another topic is quantum reinforcement learning, which we are working on in our QCI QLearning project. There are also approaches in climate modeling; for example, adapting variables to take account of new research findings. We will be excited to see how our handling of data will change in the future as a result of such “high-performance data processing engines”.

Glätzle: I find the combination of these two megatrends very exciting. planqc is also working extensively on projects to explore the potential of this link for climate models and the development of materials and active ingredients.