By Damien Petty. August 4, 2020.
We recently announced our first quantum computing investment with Rigetti Computing. Why quantum computing, why now and why Rigetti?
For those unfamiliar with the sector, quantum computers harness quantum-mechanical physics phenomena like entanglement and superposition to unlock significant advances in computing capacity.
On a fundamental level, classical computers operate with digital bits of 1 or 0 and logic gates that control the relationship between bits. For solving many important problems—such as modelling complex molecules for drug discovery, producing more precise and accurate weather predictions, or optimizing networks—the required classical computing resources scale exponentially, and thus tend to reach upper bounds on size or are out of reach due to the expense and limitations of today’s supercomputers.
Quantum computers operate with qubits, which—unlike classical bits—can be 1, 0, or in a superposition of 1 and 0 (i.e. a linear combination of both) at the same time. Further, information can be encoded in precise correlations between the states of individual qubits through a property called quantum entanglement. These properties of superposition and entanglement can be exploited by quantum algorithms to allow quantum computers to match up to certain exponential problems with a number of qubits that scales only linearly.
If you wanted to solve a set of linear equations with four unknowns, you could easily do it by hand. Expand the scale of the problem to 20 unknowns, and you’ll find it far easier to use a classical computer to solve the equations. How about a system of linear equations with 16 million unknowns? Today it would require a supercomputer like IBM Summit—a $300 million+ machine—about 5.5 hours and 56,000 kWh of electricity to find the solution. On a superconducting circuit-based quantum computer—a machine that will likely cost less than $10 million dollars—it might take around 20 minutes, and less than 10kWh of electricity to find the solution.
This dramatic reduction in ownership and operating costs will open many opportunities and creates a path to solving many previously inaccessible problems. It is a quintessentially disruptive technology.
There are multiple ways—modalities—to construct a quantum computer. They include superconducting circuit/Josephson junction systems, trapped ion systems, linear optical/photonics systems, and spin qubit chips, to name the most common. Each modality has its own development challenges, stages of maturity, and theoretical benefits.
The quantum computing industry is still in its infancy. Many companies are racing to build useable systems or developing programing tools and applications to harness them. The market is projected to grow from $500 million in 2019 to over $60 billion by 2030, according to some analysts. Because of market size and nature, we do not believe this is a winner-take-all market, nor one given to a natural duopoly. Many companies will carve out their own space and be successful. However, because of where the industry is at in its development, hardware developers will end up heavily influencing the tools that sit atop their machines and owning the foundational IP. Therefore, we chose to initially target hardware-focused, vertically integrated companies that promise to get to market early with commercially viable services and that have focused on providing access through their own or partner platforms.
After surveying the market, Rigetti clearly stood out. Compared to other modalities, superconducting quantum computers are the most mature, having benefitted from the largest research expenditure and sustained engineering development to date. Within this modality, Rigetti is far and away the leading independent company, having developed much of the foundational technology and brought multiple pioneering advances to the burgeoning market over the past few years. Google and IBM, both also focused on superconducting qubits, have generated eye-catching headlines tussling over quantum supremacy. And other companies have touted their advances within less mature modalities. With this backdrop, Rigetti has been heads-down, busy developing the necessary technology components to build a 100+ qubit commercial system. In parallel, their team has been building an impressive revenue pipeline, including significant contracts with DARPA and with the United States Air Force, and forging a suite of business partnerships, including with Amazon’s Braket platform.
Developing a commercially viable 100+ qubit system requires not only building larger quantum computer chips, but—critically—also requires reducing the system error rates so that the qubits are truly useable. Rigetti’s development plan leverages its proven 32-qubit Aspen-architecture chip technology. This technology is already commercially accessible. With its proprietary IP, Rigetti has a clear path to drastically reducing 1Q and 2Q error rates, while scaling its systems to over 100 usable qubits over the next two years. It is a full-stack developer with its own chip-fabrication facility, which affords it fast development cycles with few external supplier dependencies. We believe Rigetti will be one of the first to market with a commercially viable 100+ qubit system.
Rigetti has developed a hybrid quantum-classical architecture for the cloud. Users are able to access Rigetti’s production systems and quantum simulation systems through its own platform or via Amazon’s Braket platform and use Rigetti’s QUIL instruction language and SDK. Rigetti is also working with government and private-sector customers to build dedicated quantum computers. Rigetti’s commercial offerings are helping to embed it in the heart of the community of developers, researchers and academics, which will pay dividends down the line as it’s larger, more powerful systems are rolled out.