In this article, Gaurav Agrawal and T.S. Ramakrishnan of the SPE Research and Development Technical Section (RDTS) spoke with Madhava Syamlal of QubitSolve, a startup developing quantum solutions for computational fluid dynamics, about the status of quantum computing as part of the advanced computing roadmap.
This series highlights innovative ideas and analysis shaping the future of energy, with a focus on emerging technologies and their roadmaps, potential, and impact. With these conversations, we hope to inspire dialogue and accelerate progress across new energy frontiers.
Madhava Syamlal, CEO and founder of QubitSolve, was previously a senior Fellow for Computational Sciences and Engineering at the US Department of Energy’s National Energy Technology Lab and is credited for creating MFIX, a widely used multiphase CFD software. He holds a BTech in chemical engineering from the Indian Institute of Technology, Varanasi, and MS and PhD degrees from the Illinois Institute of Technology, Chicago.
RDTS: Tell us, what is quantum computing?
Syamlal: Quantum computing is based on quantum physics which has given us lasers and transistors, the foundation for common devices such as cell phones, computers, and televisions we all use. In 1985, David Deutsch at Oxford University described a quantum computer. Peter Shor, now at the Massachusetts Institute of Technology, published an efficient algorithm for factoring composite integers on a quantum computer in 1994. Shor’s algorithm shattered the belief that factoring large composite numbers is nearly impossible, challenging modern cryptography. This breakthrough sparked intense global interest in quantum computing.
RDTS: How is quantum computing different from traditional computing?
Syamlal: In traditional computing, information such as text or video is represented by a series of binary digits or bits, each of which is a 0 or 1. A computer processes the information by rapidly changing a programmable switch between the only two possible states, open (0) or closed (1).
Qubit is the quantum computing equivalent of a bit. Unlike a classical bit, which can be only 0 or 1, a qubit can exist in a state known as superposition, in both the 0 and 1 states simultaneously. Qubits are created by manipulating and measuring quantum objects such as electrons, photons, or superconducting circuits.
An N-qubit system can exist in a superposition of 2N states, represented by (2N-1) complex numbers called amplitudes. Storing these numbers would overwhelm the memory capacity of a supercomputer, even for N as low as 60. Thus, quantum computing offers significant computational power that relies on only a small number of qubits.
However, quantum computer functioning is more complex than classical computers. Unlike the definitive binary states of 0 and 1, the measurement of quantum states results in a random state. The probability of obtaining a particular state is determined by the square of the magnitude of its amplitude, which can be manipulated through quantum interference.
The most intriguing quantum phenomenon is quantum entanglement, wherein the measurement of a qubit instantly changes the state of all other qubits that are entangled with it.
RDTS: What are the differences in basic hardware and software?
Syamlal: Unlike the matured hardware representing a bit in classical computers and storage devices, multiple technologies are being developed to represent a qubit. Google and IBM are pursuing superconducting circuits, while IonQ and Quantinuum focus on trapped ions. Atom computing and Infleqtion are exploring neutral atoms. PsiQuantum and Xanadu are leveraging photons. Also, there is no equivalent of a classical hard disk drive for the permanent storage of qubits.
Quantum programs are written in classical programming languages like Python that invoke quantum libraries like Qiskit from IBM or Cirq from Google, or specialized languages like Q# from Microsoft.
A significant difference is in how hardware errors are handled. In classical computers, bit errors are easily detected and corrected by storing the state of the machine periodically and by encoding data with redundant bits. These are impossible in a quantum computer because copying an unknown quantum state is physically not possible. Instead, quantum error correction (QEC) protocols are used, which are still under development.
To achieve fault-tolerant quantum computers, robust logical qubits must be prepared from many noisy physical qubits. In 2024, Google achieved a key QEC milestone by demonstrating logical qubits whose error decreased with the number of physical qubits.
RDTS: Are quantum computers available for commercial sale or accessible over the cloud?
Syamlal: Yes, we are seeing them in limited quantities. In 2023, IBM installed a 27-qubit quantum computer at the Cleveland Clinic. In 2025, Pasqal plans to install a 200-qubit quantum computer at a Saudi Aramco facility in Saudi Arabia. Users can buy quantum computing time from Amazon Braket, Microsoft Azure, and IBM Q.
In the US, researchers can access quantum computers through Department of Energy labs, such as Oak Ridge National Laboratory, Lawrence Berkeley National Laboratory, and National Science Foundation centers at universities.
Simulators suitable for running small quantum circuits, say fewer than 40 qubits, are available for algorithm development. IBM’s Qiskit includes a simulator that can be downloaded and used at no cost. BlueQubit provides access to graphic processing units (GPUs) simulations of up to 36 qubits.
RDTS: Is there a strong open-source community for researchers to access quantum computing?
Syamlal: There are several open-source software options available in quantum computing. An extensive list of full-stack libraries, quantum simulators, and quantum algorithms can be found on the Quantum Open-Source Foundation website.
One of the most widely used libraries, called Qiskit, is open source. Open Quantum Design (OQD) recently launched the world’s first open-source, full-stack trapped-ion quantum computer, an initiative supported by Xanadu, the University of Waterloo, the Unitary Foundation, and Haiqu. This initiative allows researchers and developers worldwide to contribute to and benefit from a shared quantum ecosystem.
RDTS: Is quantum computing ideally suited for certain kinds of applications?
Syamlal: Any algorithm that can be run on a classical computer can also be run on a quantum computer but usually not efficiently. Quantum computing best suits “small data, big compute problems.” Small datasets are preferred because loading and reading data from quantum computers is slow. Furthermore, quantum gate operations are slow. A quantum advantage is gained only by reducing the number of steps required in the algorithm, enabled by exploiting quantum superposition, interference, and entanglement.
Industries focus on three main application areas: simulation, optimization, and machine learning. Quantum computers can tackle classically intractable chemistry problems, which could enable the design of new drugs or materials, building them from the atomic level. Quantum algorithms like Grover’s search can provide a quadratic speedup for unstructured searches, helping to identify optimal solutions in finance, logistics, and energy. Using the capability of quantum computers to represent complex data structures that are difficult to simulate classically may lead to innovative machine-learning algorithms.
RDTS: What potential applications do you foresee in the oil and gas industry?
Syamlal: The oil and gas industry has always faced challenging applications requiring advanced computing solutions. The industry’s need to process large amounts of seismic data aided supercomputer development and adoption. Likewise, quantum computing has the potential to significantly impact oil and gas by enabling highly complex simulations of subsurface geology, molecular modeling of reservoir fluids, and optimization of extraction processes. Such applications allow for more accurate predictions and improved decision-making in exploring, developing, and producing oil and gas resources as summarized in Table 1.
RDTS: Reservoir engineering and petrophysical inversion/log interpretation hinges on banded linear solvers. Do you see the potential for quantum computing here?
Syamlal: One of the quantum algorithms with provable speedup is the Harrow–Hassidim–Lloyd (HHL) algorithm, designed to solve a system of linear equations. This algorithm offers an exponential speedup compared to the fastest classical methods. However, it has several limitations that can hinder its performance.
First, the matrix involved must be sparse and well-conditioned. Second, the right-hand side vector must be easily loadable onto the quantum computer. Third, the output of the HHL algorithm is a quantum state representing the solution vector, making it inefficient to recover the entire solution.
This limitation may not be significant if the desired output is merely the locations of huge values within the solution or the value of an inner product of the solution with another vector.
Furthermore, the HHL algorithm necessitates a fault-tolerant quantum computer, which is not yet available. Meanwhile, alternatives like the variational quantum linear solver (VQLS) are being developed to run on the quantum computers currently accessible. However, VQLS is a heuristic algorithm whose speedup can only be empirically verified and not proven. VQLS may be beneficial for some types of problems.
RDTS: What can we do to enhance awareness of quantum computing in the upstream industry?
Syamlal: The time is ideal for the oil and gas industry to explore potential quantum computing applications and influence quantum computing’s development. SPE could play an important role in this through workshops, increasing engagement with the wider quantum computing community through collaborations with organizations such as the Quantum Economic Development Consortium (QED-C).
SPE could also organize computing challenges like the one by Airbus and the BMW Group toward enhancing aircraft and cars’ design, manufacturing, and operation.
The mobility industry is applying quantum computing to understand varied applications such as the energy absorption of materials, synergistic solutions with classical computing to minimize aircraft noise and maximize aerodynamic efficiency, minimizing CO2 emissions, creating realistic nighttime conditions using limited daytime data, and improving the reliability of autonomous systems.
RDTS: Is quantum computing having much of an impact on the artificial intelligence (AI) industry?
Syamlal: Quantum machine learning (QML) is an active area of research. Many institutions are involved, and papers are published annually. The field has dedicated conferences, such as (1) AI+Quantum at the Aspen Center for Physics, and (2) Quantum Artificial Intelligence and Optimization. Quantinuum recently announced a Generative Quantum AI Framework designed to utilize quantum-generated data to solve problems currently intractable using classical methods.
There is growing optimism that significant challenges in AI can be addressed through the synergy of AI and quantum technologies. However, these are early times. We must showcase practical demonstrations before we can assess the effectiveness of QML.
RDTS: Are there applications where traditional computing is unlikely to be replaced by quantum computing?
Syamlal: Both will coexist.
For many problems, classical computation is faster than quantum computation. For example, when implementing Shor’s algorithm, quantum computing is only utilized for period finding, a challenging task for classical computers; all other steps in the algorithm are performed using classical computing. Furthermore, the compilation of quantum programs is done using classical computers.
In the future, quantum processing units (QPUs) will likely function as co-processors, like modern supercomputers use GPUs alongside central processing units (CPUs). For example, IBM, Microsoft, and Atom Computing have announced plans to develop quantum supercomputers in that manner.
RDTS: What resources are available for further learning in this field?
Syamlal: A growing number of resources and reference materials is becoming available from corporations, online courses from MIT, seminal books by Nielsen and Chuang, and N. David Mermin, and various blogs.
Developing a quantum algorithm involves more than just converting a classical algorithm into a different programming language, as one would do when adapting CPU code for GPUs. Instead, it requires a complete algorithm redesign using quantum gates and measurement operations, which could potentially lead to groundbreaking applications.
Gaurav Agrawal, PhD, SPE, is the vice chair of the SPE R&D Technical Section, He has organized events in energy transition, artificial intelligence, digital technologies, and other areas. Previously he was a senior vice president of R&D at Newpark Resources in Houston and vice president of the Saudi Arabia Dhahran Technology Center at Baker Hughes. He has over 85 granted US patents.
T.S. Ramakrishnan, PhD, SPE, retired from SLB as a senior science advisor. He is a member of the National Academy of Engineering and was a recipient of the SPE Formation Evaluation award. He has published in several fields such as multicomponent multiphase flow in porous media, array induction logging, carbonate petrophysics, intelligent completions, carbon storage, plug and abandonment, magnetic resonance imaging, formation testing, and capillary dynamics. He is also an author of approximately 80 US patents.
Read the first article in this series, a Q&A with Amy Bason, deputy vice president of strategy and policy at the Oil and Gas Climate Initiative, “Navigating the Path to Net Zero in the Transportation Sector.”