The quantum computing age appears to be fast approaching. A computational chemist has launched a new company to help chemists and companies harness quantum computing’s potential to more accurately simulate chemicals and their interactions.
Computers have long ago proven their worth in chemistry, but there’s not enough classical computing power in the world to simulate molecules exactly. Computational methods like density functional theory make very good approximations, but quantum computers have the potential to model how compounds really behave. Unlike a classical computer, in which bits can represent either a one or a zero, a quantum computer’s qubits exist in a superposition of both one and zero simultaneously. That duality more closely resembles how electrons behave. Right now, the number of qubits holds researchers back. The best quantum computers have a few dozen qubits, but scientists estimate tens or hundreds of thousands of qubits are necessary to do useful simulations.
Alán Aspuru-Guzik of Harvard University has developed a number of algorithms for simulating chemical behavior using quantum computing. In 2005, he demonstrated an algorithm that could calculate a molecule’s ground state energy using simulated qubits.
Now, Aspuru-Guzik has founded Zapata Computing to help chemists and companies take advantage of what quantum chemistry has to offer. He named the company after Emiliano Zapata, a leader of the Mexican Revolution. Aspuru-Guzik, who moves to the University of Toronto this summer, thinks quantum computers will be better at chemical simulations than classical computers in 10 years. Others think it could take much longer. Predicting when that happens is one of Zapata Computing’s missions. But Markus Reiher, a theoretical chemist at ETH Zurich, says it is not too early to start using quantum computers. Because computational hardware is still imperfect, researchers need software that can exploit its current capabilities until better quantum computers are developed, he says.
Zapata Computing will advise customers on how to solve chemistry and materials problems using its algorithms. CEO Christopher Savoie says Zapata will be hardware agnostic, directing users to devices in development at IBM, Google, IonQ, and other companies. Differences in companies’ hardware, such as different arrangements of qubits, might be useful for addressing different problems, explains Savoie. In some cases all the qubits will be connected to each other, while in others it might be that a few qubits are running calculations and others are being used to evaluate their performance. That will change which algorithms and what hardware Zapata would recommend.
Zapata Computing announced it has licensed all of Aspuru-Guzik’s relevant intellectual property from Harvard. It is reporting $5.4 million in seed funding from Massachusetts Institute of Technology’s The Engine venture firm, Pillar VC, and others. Four of Aspuru-Guzik’s postdocs from Harvard will join him as founders of the company. Savoie says the company will start with about 10 employees and expects to continue hiring.
Chemistry may be just the beginning for Zapata Computing. Aspuru-Guzik says he intends it to be the top quantum software company in the US. Savoie says chemistry is the problem they can address right now, but he thinks the algorithms and other tools they will develop could be more broadly applicable in fields such as machine learning and finance.
CORRECTION: This story was updated May 17, 2018, to correct Alán Aspuru-Guzik’s affiliation.