Zapata Computing

STEX25 Startup:
May 3, 2019 - May 2, 2020

Accelerating the quantum revolution

By: Daniel de Wolff

Depending on whom you ask, the debate around whether or not quantum supremacy has been achieved may continue for some time. Or at least until the arrival of a quantum computer with more qubits. What is not up for discussion, according to Christopher Savoie, CEO of Zapata Computing, is that quantum technologies are going to change the world. “You may have heard of Google’s quantum supremacy experiment,” says Savoie. “Well, we’re the folks who write the software for those machines.” Meet Zapata Computing, the MIT-connected startup designing software for the quantum revolution.

"What differentiates Zapata from our competitors, is that we created the initial algorithms that run on these NISQ computers,” says Savoie. For the uninitiated, NISQ stands for Noisy Intermediate-Scale Quantum. Coined by Jack Preskill, the term refers to the relatively small size, in terms of qubits (intermediate scale), and our lack of perfect control (noisy) over the qubits. Our near-term, or NISQ-era, devices lack error correction, unlike their classical counterparts. So, while we have quantum algorithms to prove we can do things that classical computers aren’t capable of, we don’t yet have quantum computers big enough to provide fault-tolerant implementations of these algorithms.

“There are maybe 100 people in the world who know how to program these new computers. We can count 20 of them on our team,” says Savoie. This includes one of his co-founders, CSO of Zapata and world leader in quantum computing algorithms, Alán Aspuru-Guzik. Guzik authored the first algorithm for near-term quantum computing and the first quantum algorithm for chemistry. “We have a terrific advantage just by having more technical specialists under one roof than just about anyone else,” Savoie says.

There are maybe 100 people in the world who know how to program these new computers. We can count 20 of them on our team.

While helping fortune 100 companies to solve their toughest problems with quantum and quantum-inspired devices, the team at Zapata realized they needed a better platform on which to design and run their programs. “There wasn’t an existing workflow system,” says Savoie. “There wasn’t an existing database storage system, cloud system, or high-performance computing system. And there wasn't anything that orchestrated all the moving pieces into a platform that could manage the work for our customers, so we went out and built it.”

Zapata has taken a deep knowledge of quantum computing to create an interoperable platform for quantum-enabled workflows that is designed to be relatively easy for a domain expert to use. The platform is called Orquestra, and it organizes all of the components needed to build and deploy quantum workflows. Whether you’re an industrial mathematician or a data analyst, Orquestra lets you do your work without having to learn a new language or needing to understand the physics of how each machine works.

Because it's important to recognize that while all quantum computers use qubits—the fundamental units of information needed for quantum computation—they aren’t all built on the same technologies. Some quantum computers rely on bits of superconducting metal within microwaves, while others use trapped ions for their qubits.

“If a chemist says, ‘I want to do the ground state energy calculation for water and write a workflow for that,’ the platform can leverage mostly classical resources on AWS or Google Cloud or Azure that are spun up to do that, then run the most computationally complex portion on Honeywell’s ion machine or IBM’s superconducting machine,” explains Savoie. “You can stay at a fairly high level or go very deep into the hardware to create truly optimal performing solutions.”

Orquestra is the complete quantum toolkit. Users can work in any quantum language, use Zapata’s proprietary algorithms and other open-source libraries, and leverage optimized quantum and classical resources to tackle the toughest real-world problems. What’s more, it allows its users to interact with all of the hardware platforms out there for quantum computers. It can interact with any quantum machine in a language that domain experts can understand. 

If and when quantum computing becomes mainstream, it’s not unreasonable to expect accurate models of quantum systems. This is good news for chemical companies, materials companies, and pharmaceutical companies. Imagine having the ability to model a molecule ahead of time, before going into the lab, and knowing exactly how it will behave. “We’re talking about new chemicals for photo emissions or solar panels that are 100 percent efficient and pharmaceuticals that work 100 percent of the time for the target,” says Savoie.

The advent of quantum computing also means that problems surrounding optimization will witness a sea change. This means orders of magnitude improvement around the optimization of financial portfolios or supply chains and logistics. Today, there's no fancy algorithm helping Amazon to determine how many pallets will fit into a truck for delivery. Classical computers help. But for the most part, it's a guy with a clipboard determining if one more pallet can fit on a truck (or boat or plane).

With quantum computers and Zapata’s platform, this could all be optimized, saving tens of percentages of a logistics budget, not to mention the fuel saved. “It’s kind of hard to fit into your mind just how much influence it would have if we could use a quantum computer to pre-place where we’re going to send packages ahead of time. We’re talking millions of tons of carbon being released into the atmosphere unnecessarily. It’s that kind of scale,” says Savoie.  

Machine learning will also benefit from the advent of quantum technology. Consider our use of machine learning for autonomous vehicles. The computer needs to understand the image it is seeing: is it a ball bouncing or is it a child crossing the street? Mistakes in this arena have serious ramifications. Richer models will be more accurate, which is extremely valuable.

Originally spun out of Harvard University, Zapata’s seed round of funding was co-led by The Engine, MIT’s venture to invest in tough tech. But The Engine didn’t just provide funding. According to Savoie, Zapata developed and found its legs there. At one point, Savoie and his team took up the entire third floor of The Engine’s Cambridge headquarters.

What we’ve been able to do with our platform is make quantum computers more accessible to domain experts.

“We were very fortunate to have been in the Engine. Now that we’ve moved out of their space, we still benefit tremendously from being a part of the MIT community. We work closely with the Startup Exchange and MIT ILP, which allows us to engage, in a vetted manner, with people who are serious on the corporate side of things as partners who are interested in quantum computing.”

When Zapata was founded, quantum supremacy hadn’t yet been announced. There were only a few qubits, now we’re up to 50. Savoie and others predict that we'll reach 100 qubits soon. So it should come as no surprise that Zapata has plenty of industrial customers knocking on its door for solutions. “What we’ve been able to do with our platform is make quantum computers more accessible to domain experts,” he says. “We can put these tools in the hands of capable industrial scientists to be able to create applications for quantum computers, and that will scale. Things are very much looking onwards and upwards.”