Cutting-edge artificial intelligence (AI) models are incredibly powerful, but training and running them is computationally intensive. Today's largest AI models require clusters of powerful data center graphics processing units (GPUs). While AI is capable of solving many types of problems and answering many types of questions, the technology can still get things wrong and go off the rails.
Some of the limitations of current AI technology could potentially be overcome with quantum computing. A quantum computer uses the probabilistic properties of quantum bits, or qubits, to exponentially speed up certain types of calculations. While no company has managed to build a commercially viable quantum computer to date, many are working toward that goal.
Image source: Getty Images.
A quantum pioneer
International Business Machines (IBM 1.05%) is one of them. IBM has been doing fundamental research on quantum computing for decades and has a roadmap that leads to a large fault-tolerate quantum system by 2029. The company has already generated nearly $1 billion worth of quantum computing-related business and offers access to its fleet of quantum computers to organizations eager to test out the emerging technology.
IBM demonstrated the first computer based on quantum principles in 1998, assembling atoms of hydrogen and chlorine to successfully sort an unordered list of items. Nearly 30 years later, quantum computing is still not useful in the real world, but immense progress has been made. IBM's latest quantum computers top 100 qubits, and the plan is to exceed 1,000 qubits in 2027 and 2,000 qubits in 2033.
IBM has also been actively researching ways to apply quantum computing technology to machine learning and artificial intelligence applications. Back in 2021, the company's researchers developed a quantum algorithm that can deliver an exponential speedup for a certain class of machine learning classification problems.
Once large-scale, fault-tolerate quantum computers are a reality, the technology could revolutionize various fields, including AI. Training an AI model is a time-consuming, computationally intensive process that involves massive amounts of data. A quantum computer could potentially accelerate the process and lower the costs of AI training. Quantum computing could also speed up AI inference and lead to better results.
Enterprise AI and quantum computing in one package
While other tech giants are working on quantum computing and AI technology, IBM's long track record in quantum computing and its successful enterprise AI business will make it a potent force in the quantum AI age.
IBM has booked $7.5 billion worth of business related to generative AI, and much of that comes from the company's consulting arm. IBM's pairing of AI consulting, which covers implementation and other services, and AI software, primarily its watsonx platform, has proven to be a winner among its enterprise clients. A similar story is likely to play out once quantum computing becomes commercially viable.
IBM may not seem like the most cutting-edge tech company, but it's well-positioned in both AI and quantum computing. As quantum computing technology matures and is eventually applied to AI, IBM is almost certain to be in the thick of it.
