Our first investment in Zapata Computing — a quantum computing software spinout from Harvard — was done already at the end of 2017. We are very proud to have followed Zapata’s team from incubation until today as the first pre-seed investors in the company. As a first-time fund investing in rather tough areas to grasp, we get many questions around this investment and have therefore decided to elaborate to our best ability on how we identified and why we decided to invest in Zapata.
Short Intro to Quantum Computing:
There are increasing doubts around Moore’s law continued application as the transistor density of conventional computers is approaching the atomic length scale. At the same time, advances in nanoscience now makes it possible to engineer, control and manipulate inherently quantum mechanical systems.
While classical computers compute by performing (Boolean) logical operations on binary bits — 1s and 0s — quantum computers operate according to the laws of quantum mechanics. In a quantum computer, the elementary store of information is a qubit. Unlike a classical bit, a qubit is not binary but can also be in a state “in between” 1 and 0 by a principle called superposition. But the real benefit comes from the superposition of a system of several qubits. The quantum correlations of such a system (or entanglement — what Einstein called “Spukhafte Fernwirkung”) represents a store or system of information that has no analog in a classical computer. This gives quantum computers an exponential advantage over classical computers as the information capacity scales exponentially with the number of qubits. In fact, writing down the full description of a system of 300 qubits would require more numbers than the number of atoms in the universe!
For long quantum computers remained a theoretical possibility, but research advances have opened a new world of practical possibilities. Tech forethinkers quickly catched on and over the past 5 years Google, Rigetti, IBM have demonstrated an ability to scale up their hardware going from 5–50 qubits.
Where we came in:
Prof. Alán Aspuru-Guzik has long been a pioneer in the development of quantum algorithms. Anders approached Alán back in 2016 as he identified his group’s research as world leading and over the years developed a close relationship to him and his group. When Alán aired his ideas about spinning out a company around his quantum algorithms, Anders was thrilled.
Not just was the research world-leading, Alán was also able to gather an impressive team around him with Christopher Savoie in the lead who has a strong commercial and tech track record having developed the core parts of the natural language technology that later went into Siri, and founded an OLED company. The major moat of the company is indeed its talent. Considering that our society lacks ML/AI talent, think about the lack of quantum developers!
While other companies in the field were focusing on the hardware of quantum computers (or software for those specific hardware solutions), Zapata is positioning itself as THE software company for a wide range of hardware providers, already working with IBM and Google and others. It’s algorithms can be described as a hardware-agnostic “operating system” for clients to interact with quantum computers to e.g. simulate chemistry on a molecular level (quantum chemistry & materials) or for machine learning or optimization applications.
One important hurdle to overcome for quantum computers is their susceptibility to errors and noise. Therefore, successful applications for near to intermediate term quantum computers would either have to be especially designed to run on NISQ (Noisy Intermediate-Scale Quantum) hardware or employ clever algorithms for error correction. Zapata has made important contributions in both categories exemplified by the Variational Quantum Eigensolver (quantum chemistry) and qVector for error correction.
Future business models will likely be an offer of “algorithms-as-a-service” to commercial clients or software licenses to hardware companies. While fully fledged commercialization is not to be expected in the near-term, both commercial players as well as nation states have started to invest heavily in the field (China allocated $10bn to a 5-year quantum science initiative, EU more than $1bn to its quantum research project and the US National Quantum Initiative Act recently proposed a budget of $1.3bn).
When we invested, Zapata was an early university spin-off in a promising tech area with a leading team — this is exactly where Propagator thinks it can take a strong position. We were lucky to catch the right timing — both from a technological point of view and of VC interest, getting top US VCs onboard in the following seed-round. Bridging leading tech developments right before early market adoption is where we feel comfortable.
Where we’re headed:
The most current (extremely) exciting news is the latest publication by the Zapata team showcasing their last algorithm for quantum factoring on near-term quantum computers, possibly starting a much sooner development toward quantum encryption than previously expected. While Shor’s algorithms for quantum factoring has long been one of the most famous potential applications for quantum computers, it’s applicability has still been a distant goal as it would require on the order of 100 000 error corrected qubits to work for relevant problems in encryption.
The ingenuity of Zapata’s new factoring algorithm is in mapping the problem onto a model that can be efficiently solved using a clever trick called the variational principle, which is specifically tailored to work without error correction and a smaller number of qubits.
We are thrilled to follow the further development of Zapata and their algorithms and are grateful to continuously be intellectually challenged by such an incredible team of quantum tunnellers!
DISCLAIMER: Alán has in the aftermath of the investment joined Propagator Ventures as a technical advisor and has been invited to become a shareholder in the management company.