– Achieving quantum supremacy, overcoming the limitations of existing supercomputers
– Calculation that took 100 hours, now in 40 minutes… From road maintenance to new drug and new material development
– “Securing market-leading technology at the inflection point where the quantum computer market begins in earnest”
Qunova Computing (hereinafter referred to as Qunova), a leader in quantum computing innovation, has achieved the world's first quantum advantage with the 'HI-VQE (Handover Iteration Variational Quantum Eigensolver)' algorithm. This has enabled AI to overcome the limitations of only learning previously created materials and create and predict new materials.
Qnova, which started as a faculty startup at KAIST in 2021, is a quantum computing software company led by CEO Lee Jun-gu (concurrently a professor at the Department of Electrical and Electronic Engineering at KAIST). Qnova plans to release all algorithms, including HI-VQE for solving chemical problems, HI-VQA for optimization solutions, and HI-VQLS for thermodynamics, within the year, and Milky Way, a quantum algorithm-based new drug and new material design platform for the general public, and Pulsar, a molecular analysis software, by 2027.
Qnova is also conducting research with RIKEN in Japan, which owns the world's second-largest supercomputer, Fugaku. Through joint research with RIKEN, Qnova is proving that its quantum algorithm is faster and more accurate than the world's best supercomputer.
We met with CEO Lee Jun-goo at his office in Gangnam-gu, Seoul, and talked about quantum computing, quantum supremacy, the algorithms and software that Qnova is developing, and its business plans.
Quantum Computers and Quantum Supremacy
The computers we use today process information using bits of 0 and 1. However, quantum computers use a different unit called 'qubit'. The core of qubits lies in the quantum mechanical phenomena called 'superposition' and 'entanglement'. Superposition allows 0 and 1 to exist at the same time, and entanglement allows multiple qubits to be connected to each other and share information. If a conventional computer finds a way to solve a maze one by one, a quantum computer has the ability to explore all paths simultaneously. Therefore, it can calculate exponentially faster in certain problems. For example, the way a conventional computer finds the exit in a maze is like exploring paths one by one. It goes through path A, and if it's a dead end, it comes back and tries path B sequentially. On the other hand, a quantum computer can explore all paths simultaneously by utilizing the quantum mechanical phenomena called superposition and entanglement.
Each additional qubit doubles the computational power. With 40 qubits, petascale computing is possible, and with 60 qubits, problems that are impossible with current supercomputers can be solved. This is why 60 qubits are important. Quantum supremacy refers to a state where a quantum computer can calculate a specific problem faster and more accurately than a conventional supercomputer. In other words, problems that were impossible with a conventional supercomputer can be solved with a quantum computer. In theory, quantum supremacy starts with 40 qubits, but for practical industrial use, 60 qubits or more are needed.
World's first quantum supremacy achieved… Calculation results publicly demonstrated
Qnova has recently attracted attention by commercializing the world's first algorithm that achieved quantum supremacy. Currently, VQE (Variational Quantum Eigensolver, a hybrid quantum algorithm that combines quantum computers and classical computers) is a core algorithm that accounts for 40-50% of the quantum computing market. Qnova's HI-VQE is 100 times more efficient than VQE, and HI-VQE is 68 qubits in scale.
In March, Qnova registered the HI-VQE algorithm on IBM's Qiskit Functions platform (a core platform created by IBM to build a quantum computing ecosystem, which can be called the 'app store' of the quantum computing field).
The representative explained, “There are seven software registered in IBM’s Qiskit Functions, and Qnova is the only one that has achieved quantum supremacy. There are about 80 quantum computing software companies around the world, but most of them create products by combining software floating around the Internet. Without original technology, there is no true competitiveness.”
“The important thing is not the number of qubits, but achieving a use case that can compute faster and more accurately than current supercomputers, or more accurately. Making something that was previously impossible possible is the essence of quantum supremacy.”
At the '2024 Quantum Korea' event, Qnova presented the world's first real-time demonstration. It showed the molecular structure analysis of a compound called lithium sulfide (Li2S) in real time. This was the world's first public demonstration of the process of obtaining meaningful calculation results in real time using a quantum computer. Until now, quantum computer hardware companies have been able to show that their equipment works, but there has been no demonstration of obtaining meaningful calculation results in real time.
Founded after 30 years of quantum computing research
Qnova, the world’s only quantum supremacy, did not happen overnight. CEO Lee Jun-gu’s quantum computing journey dates back 30 years. CEO Lee, who began quantum computing research at the NEC Research Institute in the United States in 1996, can be said to be a living witness to this field.
“We started full-scale research in 2018 when we first established the university quantum computing ITRC center in Korea. However, research alone had its limitations. We wanted to create technology that could be applied to actual industries, and to do that, we had to start a business.”
Professor Francesco Petruccione of South Africa, who has been conducting joint research with this representative, participated as a co-founder. Professor Petruccione is a world authority in the field of quantum machine learning and is the author of the first quantum machine learning textbook.
Currently, half of the 26 employees are foreigners, 21 are R&D workers, and 14 of them have doctorates. Recently, they have been spurring their entry into the global market by hiring Dr. Jaewan Kim from Amazon as CPO (Chief Product Officer). CPO Jaewan Kim is a software engineering expert with experience in achieving $1 billion in sales at Amazon.
“There were difficulties in securing domestic manpower, but the more important reason is that most of our customers are overseas. We organized our manpower like this for globalization.”
Graduating from the University of Toronto’s Creative Destruction Lab Quantum Stream program in 2022 has also been a huge help in building a global network. Through this program, which started with 50 teams and only 10 graduated, I was able to build a network with global investors such as Bessemer Partners.
The results of 30 years of research into quantum computing are now becoming a reality.
New drugs and materials can be developed even without data
The core of HI-VQE is the 'hybrid approach'. It is a method that alternately utilizes the strengths of quantum computers and conventional computers. The innovation of this methodology is also recognized in the industry. In the past, it was thought that meaningful calculations were impossible with imperfect NISQ (Noisy Intermediate-Scale Quantum, the current imperfect but practical quantum computer) quantum computers, but Qnova's hybrid algorithm broke that conventional wisdom.
The representative said, “Just as CPUs did not disappear when GPUs were introduced, quantum computers do not replace existing computers, but rather complement each other. When the strengths of each are extracted and combined, amazing performance is achieved. We were the first to introduce this methodology, and we have also secured the original patent. With this technology, we have achieved quantum supremacy for the first time in the world.”
The area where the HI-VQE algorithm will bring the greatest innovation is the development of new drugs and new materials. Currently, AI-based new drug development is in the spotlight, but its limitations are clear. AI must learn from existing data, and it does not work well in new areas where there is no data.
“Quantum computers are responsible for new areas of calculation that AI cannot do. Since they interpret problems based solely on chemical principles, they can obtain accurate answers even without data.”
The principle of hemoglobin's oxygen transport was also discovered by HI-VQE. How hemoglobin transports oxygen in our bodies is the core of life phenomena, but it has been difficult to calculate the exact mechanism until now. At the center of hemoglobin is a compound called Fe-porphyrin. To accurately calculate the energy change that occurs when this substance combines with oxygen, it is necessary to analyze how approximately 10 million electron orbitals interact with each other. Qnova accurately interpreted this on a 44-qubit scale.
It is also possible to solve complex problems directly related to our daily lives.
Optimization problems are also easy to solve
For example, let's say you're repairing tens of thousands of roads in a certain administrative district. Deciding when and in what order to repair each road is a very complex problem. If you set the variable indicating whether to repair each road to 0 or 1, then choosing which roads to repair this week becomes a decision variable. If the total number of roads is a few hundred or a few thousand, this is not a problem, but if it exceeds 100,000, the problem is different. It is impossible to establish an optimal repair plan that considers the entire road network at once using existing methods.
Qnova's HI-VQA algorithm can optimize from 100,000 to 1 million variables. This means that nationwide road network optimization is also possible. "Problems with about 10,000 variables can be solved on current supercomputers, but problems with about 100,000 variables become unsolvable even for supercomputers. If optimization calculations are possible for complex problems such as road maintenance, we can drastically reduce costs while maximizing convenience for citizens," Lee emphasized.
Thermodynamic-based aircraft design is also possible
The design problem of aircraft wings shows another level of complexity. In order to increase the fuel efficiency and improve the safety of aircraft, the airflow around the wings must be accurately predicted. This involves complex physical phenomena such as vortex analysis of the airframe and water condensation or condensation on the wing surface. To solve these problems, the space must be divided into very small grids and the physical state at each point must be calculated. “Current supercomputers cannot calculate when the number of grid nodes exceeds 10,000, but in actual aircraft design, much more precise calculations are required,” explained Lee Jun-gu, CEO. Qnova’s HI-VQLS can solve problems on the scale of 1 billion nodes, enabling precise designs that were previously unimaginable.
HI-VQE, 98 claims original patent application
One of Qnova’s core competitive advantages is its strong patent portfolio. It has filed and registered a patent in the US with 98 claims related to HI-VQE alone. The CEO likened this to Qualcomm’s CDMA patents.
“Just as CDMA solved the mobile network access problem and triggered a rapid growth in the mobile communications market, our technology will be the world’s first practical use case that will accelerate the formation of the quantum computing market.”
In fact, Qualcomm earned $6 billion from CDMA patent licensing in 2016 alone. Qnova can also expect significant patent revenue once the quantum computing market takes off.
The current quantum computing hardware market has exceeded 1 trillion won this year and is expected to grow to 2 trillion won next year. McKinsey forecasts the quantum computing market to reach 93 billion dollars (approximately 130 trillion won) in 2030.
In preparation for the quantum computer market expected to emerge in 1-2 years
The growth potential of the quantum computing market is limitless. The market size of computational chemistry, numerical analysis, new drug, new material discovery, and AI, which Qnova is targeting, is expected to be 12.8 trillion won in 2030. Of this, the market that quantum computing can occupy (SAM) is 3.6 trillion won, and the market that Qnova can directly attack (SOM) is 538 billion won.
“We predict that the inflection point where the quantum computer market begins in earnest will occur in 1-2 years. We have already completed preparations and secured the technological capabilities to lead the market.”
The representative emphasized that quantum computing is not simply an extension of existing computing, but a completely new paradigm. “The evolution from CPUs to GPUs and now to QPUs (Quantum Processing Units) is a natural flow. We are at the forefront of this change.”
CEO Lee Jun-goo expressed his ambition to “act as a trigger for the quantum computing market.” In fact, Qnova’s quantum supremacy, achieved for the first time in the world, is a historic achievement that has made theoretical possibility a reality.
Qnova’s journey is only just beginning. It is moving towards specific goals of completing the integrated platform by 2027 and achieving sales of 20 billion won by 2030. Furthermore, it is also preparing next-generation technologies such as quantum AI and ultra-low-power GPT.
Whether Korea can demonstrate global leadership in quantum computing, following semiconductors and batteries, is the reason why Qnova’s steps are drawing attention. We look forward to the day when the “iPhone of the quantum computing market” that CEO Lee Jun-gu dreams of will be born in Korea.
You must be logged in to post a comment.