CARY – Atom Computing has raised $15 million in a Series A round, hired former Intel and Lenovo executive Rob Hayes as CEO, and debuted its first-generation quantum computer, Phoenix, which was built on first-of-its-kind technology developed by the company’s cofounders.

And, the company opened its executive offices in Cary.  Quite the week.

In an interview with Hays, WRAL TechWire asked for more details about the company, and about the promise of quantum computing.

The below Q&A is lightly edited for clarity.

WRAL TechWire: Tell us about how the company was created, and why it was founded.

Ben Bloom, CTO, and Jonathan King, chief scientist, are the co-founders of Atom Computing.

They founded the company in Berkeley, CA in 2018 with the purpose to apply their expertise in quantum mechanics to build a useful quantum computer out of naturally quantum materials.

Ben worked at Intel and Rigetti after getting his Ph.D. in 2014 from University of Colorado Boulder where he built the world’s most accurate atomic clock out of Strontium-87 atoms.

Prior to co-founding Atom Computing, Jonathan was doing his post-doc research on Nuclear Magnetic Resonance at University of California Berkeley, where he got his Ph.D.

The two of them came together three years ago when Ben was just getting the company off the ground, and they realized they could combine their expertise on building long-lived controllable quantum systems.

TW: Would you please tell us about quantum computing – how is it different, what are potential benefits, what are areas of opportunity?

Classical computers store and manipulate data as bits, 0 or 1. Quantum computers are more like analog computers, information is stored in an incredibly rich and continuous state space, where those bits are now called qubits.

Interestingly though, at the end of the computation, when you measure a qubit you only ever measure a 0 or 1.

It’s this act of performing the calculation on a very large state-space all at once that gives you this incredible speedup in calculation.  So it’s very important that applications and developers design algorithms that utilize this aspect of quantum computing.

To me the most exciting problems that naturally lend themselves to these calculations are in fact other quantum mechanics problems!

It’s probably fairly unsurprising then that the first applications for quantum computing are going to be about solving classically-intractable, quantum mechanics problems whether that is for chemistry, materials, or understanding the world around us.

Many people talk about future use cases from combinatorial optimization, engineering simulation and machine learning.

The fact is, we are at the beginning of something so radically new, it’s hard to see what the killer applications will be. People working on Boolean algebra probably didn’t picture anything close to the internet.

A quantum computing primer: What makes quantum different

 

TW: Let’s dig into the technology – what’s a nuclear-spin qubit – and what’s the element they’re coming from? Why’s that important? What’s the process?

Let’s take a step back and explain the bigger picture of what we’re doing before jumping into the technological breakthrough we’ve made.

We build quantum computers using optical traps, in this case optical tweezers, which cool and trap individual quantum particles.

We do this for a few reasons, first if you want more qubits you just add more light to the system (or use that light more efficiently) and second we don’t require wires, all our controls and readouts are performed optically in free space, kind of like a movie projector.

The other benefit of this is that the qubits are very small, 1 atom large, and we can pack a lot of them in a very tiny space.

So, what’s a nuclear-spin qubit? The normal problem people have when trying to trap and manipulate individual atoms is that the trapping itself causes decoherence, information being lost, between the internal states of the atom that form your qubit.

Even if those states involve the nuclear-spin, they usually couple that to the electron spin which interacts with the trap causing this decoherence.

At Atom Computing, we’ve built the first quantum computer of individual atoms that completely sidesteps this thorny issue by utilizing qubit states that are purely nuclear in nature. We use Strontium-87 to accomplish this, but our hardware, software, and methods are easily transferable across a range of elements.

Honeywell to invest up to $300M in new quantum computing joint venture with UK firm

TW: Tell us about Phoenix, the quantum computer the company has recently announced

Phoenix, our first-generation quantum computer, is our proving ground where we are showing off all operations necessary for a useful quantum computer, and allowing a select external pilot users on the system to test out our software stack.

Neutral atom systems are very flexible and scalable in terms of the number of qubits, building the first system it was just as easy to build a system that could control 2 qubits as it was to build a system that could control over a hundred.

In fact, after working with 50 qubits for quite some time on daily operations, it only took one day to “upgrade” to 100 atoms by changing some inputs to a python function and rerunning calibrations.

This is why we’re excited about atoms, imagine the time and effort that would have had to go into even classical computing technology to make a change like this.

So if we keep building bigger, more performant systems, why do we care about coherence time? As a back of the envelope guide to quantum computing, every operation on a quantum computer is going to have an error proportional to t_operation / T_coherence.

It doesn’t matter if it’s a gate, readout, classical error correction feedback, transit time from remote nodes, or ion movement all of these timescales are going to get compared to your coherence time. At Atom Computing we’re starting with that denominator being … very long … and we’re hiring engineers and physicists to make t_operation shorter and shorter with every upgrade, every new generation of the system.

Duke professors’ quantum computing startup going public in $2B offering

What’s next for the company and for Phoenix?  What else is important?

The company will grow and advance on a number of fronts in our mission to build the most scalable and stable quantum computers.

Over the coming year, we expect to continue to tune and optimize Phoenix to improve coherence times, develop error correction schema and gain experience running quantum circuits.

Meanwhile, we will begin building our second-generation system that will be even more powerful.

We will also build out our software stack to add features and libraries, improve usability in preparation for customer launch, and work with the open source and commercial developer ecosystem to integrate third party developer frameworks and application software with our quantum computing system.

We will hire key positions to help us with product management, marketing and business development. This team will build our customer and partner ecosystem to grow the business.