IBM advances its quantum roadmap as competition heats up


IBM reached a quantum-computing milestone in March with the first U.S. deployment of an on-site, private-sector, IBM-managed quantum computer. The IBM Quantum System One, installed at the Cleveland Clinic, is the world’s first quantum computer to be specifically dedicated to healthcare research, with the goal of helping the Cleveland Clinic accelerate biomedical discoveries, according to IBM.

The announcement didn’t surprise Scott Buchholz, global quantum computing lead at enterprise advisory firm Deloitte. “IBM is a leader in the race to build useful, scalable quantum computers,” he says. “Their research teams have been working to build the software, hardware, and supplier ecosystem necessary to support the long-term development of these important technologies.”

Buchholz, like most followers of the nascent quantum computer market, is both pleased and surprised by the field’s rapid progress. “As it turns out, manipulating individual atoms to do useful work at temperatures colder than outer space is an incredibly complex undertaking, and yet we’re collectively making progress,” he says. “At the same time, we’re comparing today’s quantum computers with the result of more than six decades of the world’s smartest minds working on today’s classical computers, so the bar for utility is very, very high.”

Cloud access to quantum computers

Since 2016, when IBM put the first quantum computer online for anyone to use, the company has established an ecosystem of over 450,000 users with access to more than 20 quantum computers via the cloud, says Oliver Dial, CTO of IBM Quantum.

Four years later, IBM released its Quantum Development Roadmap, which outlined how the company’s engineers and developers would advance the quantum field across both hardware and software. “The team continues to deliver on the goals set out in our roadmaps,” Dial says.

For many observers, IBM’s progress appears to be occurring in quantum leaps. Last November, IBM Quantum deployed a 433-qubit processor. (In terms of storing information, a qubit plays a similar role as a bit, but it behaves much differently due to the quantum properties on which it’s based.)

“Later this year, we will be unveiling the Heron processor, our first modular processor,” Dial says. “The modular approach will allow for rapid scaling, with a goal of more than 4,000 qubits by 2025,” he predicts.

Organizations participating within the IBM Quantum Network, spanning a wide range of industries, are working on case studies in multiple areas, including healthcare, automotive, chemistry, finance, machine learning, and more, Dial says. So far, IBM has installed IBM Quantum System One systems in Germany, Japan, and the U.S.

100-qubit quantum system on the horizon

Within the next two years, IBM anticipates rolling out a quantum system capable of supporting 100 qubits and 100 gates. “That should be enough to begin some demonstrations of limited advantage over existing classical machines,” Buchholz says. As the technology advances, he expects IBM to offer upgraded models with the ability to address increasingly complex problems with larger data sets. “Ultimately, we anticipate quantum computers opening new opportunities that are difficult to envision today, in the same way it was difficult to foresee today’s Internet in the early 1990s.”

Yet there’s also a downside to quantum computing—specifically the ability to complete the factorization of large numbers using Shor’s algorithm, which could be used to break current, widely-used public-key cryptography schemes. “When quantum computing becomes available at scale, current encryption technologies will become vulnerable,” says Doug Saylors, a partner with technology research and advisory firm ISG. He warns that this vulnerability will open the potential for large scale breaches across platforms that use TLS encryption, such as websites handling online financial transactions, and systems storing personal healthcare data.

IBM is addressing this issue by banking on its ability to make significant advancements in quantum computing technology, services, and security through its IBM Quantum Safe program. Over 100 research and academic institutions and 100-plus major global enterprises are participating in IBM Quantum Safe, contributing use cases and working with the company to understand the importance of quantum in their respective areas, Saylors says.

Saylors notes that IBM has built a large team of internal and external resources focused on “crypto-agility,” the cybersecurity side of quantum. “As a participant in the NIST cryptography program, focused on building algorithms that provide some form of resiliency against a quantum computing attack, IBM was recognized as a leader in this space.”

Dial states that IBM’s mission is to make the world quantum-safe. “Last year, NIST released the first four algorithms to protect data from quantum hackers, and IBM worked on three of them,” he says. “We are working with enterprises and government agencies worldwide to implement new quantum-proof encryption of data.”

Dozens race to deliver functional quantum computers

The race to join IBM as a producer of fully functional quantum computers is heating up rapidly. “There are more than a half-dozen different approaches to building quantum computers, and dozens of companies are involved, from startups to household names,” Buchholz says.

The largest competitors and potential competitors include Capgemini, Amazon Web Services, Google, and Atos. “Atos is a smaller player from a budget perspective but is working in quantum through France’s Bull Systems in a nationalistic model,” Saylors says. “Capgemini is working closely with IBM from a hardware perspective while leveraging their life sciences and application and development expertise to focus on services around quantum,” he adds.

Saylors notes that there were more than 110 mergers and acquisitions in the quantum computing space between January 2018 and July 2022. “Some of these were focused on specific industries, others focused on materials, and others on the type of processor used to create the quantum computer,” he explains. “As an example, IBM uses a traditional superconducting approach while others are using trapped ion technology.” (Trapped ion qubits encode quantum information in the electronic energy levels of ions suspended in a vacuum. Ion traps are generally regarded as being more stable than superconducter qubits.)

Potential quantum-computing customers

Customers spanning multiple fields are exploring quantum computing’s potential, from governments to financial services to pharmaceuticals to chemicals to manufacturing to logistics. “Because of its broad applicability, as the technology becomes increasingly capable, organizations of all sorts will start using quantum computers to solve problems,” Buchholz says. “In the same way GPUs started as a way to accelerate computer games and have become fundamental to machine learning, expect that quantum computers will start in known spaces and become more useful and ubiquitous over time.”

Yet before quantum computing can become a mainstream technology, it must first overcome some significant operational hurdles, particularly unintended interactions between qubits and the environment— commonly referred to as noise. A qubit’s ability to maintain a superposition state can fall apart due to noise, significantly degrading performance.

Dial declares that IBM’s mission is to bring useful quantum computing to the world. “The goal [is] delivering applications with a quantum advantage: where a computational task of business or scientific relevance can be performed more efficiently, cost-effectively, or accurately using a quantum computer than with classical computations alone.” He adds that developments are leading to an inflection point where quantum computers will be able to meaningfully outperform classical computers on problems of real value. “In other words, useful quantum computing.”

Quantum computers are generally considered to be better at solving highly complex problems across small data sets. “Things like molecular modeling can be completed in a fraction of the time it takes with standard and high-performance computing (HPC) platforms today,” Saylors says. Quantum will allow acceleration of complex research at a scale not seen before. “In fact, when used in combination with standard and HPC, it could reduce time to market by years for products in specific industries,” he notes.

To deliver useful quantum computing it will be necessary, at least for the time being, to combine quantum technology with classical high-performance supercomputing in the same system. “We’re calling this ‘quantum-centric supercomputing’,” Dial says. “The idea is that anyone trying to solve a complex problem can utilize quantum technology without knowing the details of quantum physics, just like they can utilize classical resources today without knowing the specs of the CPU or GPU they’re using.”

Early adopters find the quantum advantage

Organizations in the IBM Quantum Network are already conducting research and case studies that apply to their particular industries. In one of these studies, Dial notes that IBM and semiconductor lithography and materials firm JSR used a quantum computer to simulate a key chemical process in the fabrication of semiconductors. “This research is an early example of implications for the global semiconductor industry with quantum computing’s application to solve a hard chemical engineering problem,” he explains.

Besides quantum hardware, Dial says that IBM developers are also working on intelligent software to efficiently distribute workloads across quantum and classical systems. The developers are building on the work IBM has already done in the creation of the Qiskit Runtime, the firm’s containerized quantum computing service and programming model.

Saylors believes the most promising use cases will be in life sciences and aerospace and defense, addressing the complex problems these industries face. “However, we’re already seeing commercial product development and the ability to advance cybersecurity protections in financial services and energy,” he adds.

What’s next from IBM?

IBM’s game plan includes further advancement of quantum computers through both qubit capacity increases as well as range improvements, which boosts accuracy, Saylors says. “Longer term, expect IBM to become a leader in this space and establish a dominance similar to that of the IBM mainframe platforms when they first became available,” he predicts.

Dial says that IBM has a target to deliver a 4,000-plus qubit system in 2025. “But we don’t do that with a single giant processor,” he states. “It’s a combination of different chips wired together through classical technology as well as short-range and long-range quantum coupling.”

IBM’s Quantum System Two, which the company announced last December, is scheduled for shipment before the end of this year. The system supports multiple processors, as well as the ability for systems to be interconnected. “This will give our clients the flexibility to come up with the combination of processors and systems that works best for them,” Dial says. “This modular architecture will usher in a new era of scaling and provide a path to systems with 100,000 qubits and beyond.”

Look ahead to possible quantum applications

Buchholz believes it’s important to remember that quantum computers represent an entirely new technology—not simply advanced versions of conventional systems. “It can require one to two years of retraining to be able to use a quantum computer effectively, so organizations need to plan in advance,” he says.

IT leaders should begin considering how quantum computing might fit into their future plans, Buchholz advises. Given the technology’s potential and high profile, it would be shortsighted not to evaluate possible applications in order to keep a step ahead of competitors and internal colleagues. “Whether you know it or not, there are people in your organization exploring quantum computers on their own today.”

Copyright © 2023 IDG Communications, Inc.



Source link