Can Algorithmiq solve the atom?

Last week we discussed a web3 company despite the fact I know almost nothing about web3. So, this week, I want to take a look at Algorithmiq, a Finnish quantum computing algorithm company. Because I also know almost nothing about quantum computing. 

But, as John Wheeler famously said:

 “If you are not completely confused by quantum mechanics, you do not understand it”

And, I do actually know something about the quantum computing industry. I’ve looked at more than a few companies in the sector.

What would probably surprise you most is that the industry is … not tiny.

(Almost) understanding quantum computing

The quantum computing market is valued around $13B, and it’s growing quickly. Most of that spending comes from big players like Google, IBM, and newer entrants like Rigetti. Smaller quantum rigs are as “affordable” as ~$15M, while larger installations can be $100M+. 

I won’t give a detailed explanation of how quantum computing works. But the venture-level metaphor goes like this:

A normal computer operates with “bits”, simple electronic logic gates which can either be “on” or “off”. With thousands of bits running in parallel, a computer can store information, run programs. Anything.

Quantum bits or “qubits” operate by quantum entangling multiple bits in parallel. Then quantum properties allow qubits to be both “on” and “off” at the same time. This means that a small number of qubits can perform far more calculations than a similar number of classical bits.

Whew.

If that made sense to you, congrats. You’ve watched Oppenheimer too many times.

Quantum computing so far: basically a very expensive proof of concept

It’s still early days in the world of quantum. 

As of 2018, there were only 11 quantum computers in the world. And there are challenges ahead before the tech is able to truly outcompete classical supercomputers.

Here’s a comparison:

In 1943, the Bell Labs Model II was one of the earliest “computers”. It had 440 bits of processing, and was used to test WW2 aiming systems. Today, my laptop has 16GB of ram, which is 1.28e+11 bits of processing. (A very big number).

In 2019, Google’s groundbreaking Sycamore quantum processor had 53 qubits. Earlier this year, IBM announced they had built a quantum computer with a historic 433 qubits. Against 128 billion in a laptop.

It’s not close.

Yes, computing development is exponential. And a single qubit is exponentially more powerful than a classical bit. But we’re probably still decades away from a truly revolutionary quantum breakthrough.

Gold panning in the desert

That being said, people are making money on quantum right now. Both Microsoft and IBM offer “quantum cloud services”. Because even if the industry is an incestuous circle of researchers, academics, and big tech. You can still make money selling to the academics.

Which brings us (finally) to Algorithmiq. Here’s how they describe what they do:

“Our mission is to revolutionise life sciences by harnessing the potential of quantum computing to solve currently inaccessible problems. We envision a world in which new drugs can be discovered and invented quickly, efficiently and cost-effectively, leading to affordable and precise medical treatments.”

The company has a solid team. Their CEO is Sabrina Maniscalco, a Professor of Quantum Information at the University of Helsinki. And one of their board members is former Nokia CEO Jorma Olilla. Which confirms my theory that literally the entire Finnish economy is based on indestructible brickphones.

The main problem I see for Algorithmiq though, is their market.

What is an atom anyway? 

Algorithmiq says they aim to solve the “electronic structure problem”, which sent me down a rabbit hole on quantum physics. Here’s my (simple) assessment of the situation:

• Atoms have a nucleus

• Electrons fly around that nucleus unpredictably

• The Born-Oppenheimer approximation assumes a non-moving nucleus

• Researchers built quantum physics software to further model molecular interactions

• This software is used in materials development and pharmaceuticals

That wild ride did more than just enable my second (2) mention of Oppenheimer. (Your welcome Bobby J). It also clarified what Algorithmiq is trying to do:

With modern quantum computing technology, and their own algorithms, the company believes they can more effectively model how atoms and molecules interact, enabling exponential innovation in pharmaceuticals development.

Pretty inspirational right?

Well. 

Wikipedia brought to my attention that there are currently more than 50 quantum chemistry software solutions. Let’s look at Gaussian, one of the biggest players.

The Quantum Competition (Quompetition?)

Gaussian was founded in 1970 by quantum researcher John Pople. They’ve been continually innovating ever since. 

But if you’ve been thinking the quantum computing software space might not be huge, I’d be inclined to agree with you. 

Estimates say Gaussian only has an annual revenue of $4M. But let’s assume that’s conservative. Pricing on their website ranges from $1,000 for a basic license to $35,000 for a broader government license. Let’s assume a $5,000 average.

Even if every one of the 4,800 universities in America subscribed, that would only be $24M in revenue. But in reality, only 766 schools have physics departments, which would imply a max total revenue of only $3.8M. 

Maybe $4M in revenue isn’t too far off. And this is the established competition in Algorithmiq’s industry.

The verdict: not good

I hate to be the pessimist saying the numbers don’t add up. But like, they don’t.

Maybe Algorithmiq has a new business model. Maybe they’ve built some innovation that enables genuine competitive advantage. Maybe they’re targeting the world’s 5,100 research pharmaceuticals, instead of academics. And maybe those pharma researchers are gonna pay a lot more than $5,000 per license. 

But that’s a lot of maybe’s. 

What I know for certain is that Algorithmiq just raised a $15M Series A in a very crowded space. In a quantum sector that still looks years from widespread adoption. 

That makes it a $30-60M company, in an industry where the winners make $4M in revenue. 

What’s the growth trajectory? What’s the exit? What would it take for this to be a $100M company and deliver something like a reasonable return?

Right now, I just don’t see it. 

But then again. Never bet against Finland.

Until next time. This has been,

The Weekly One Pager

—

P.S. - Sunday counts as part of the week now.

P.P.S - We also have our first interview with a CEO coming up sometime soon. So make sure to subscribe! Tell your friends. Tell your wife. Tell your boss (but only if you want them to know you’re smart and well read). And if these emails aren’t always showing up in your inbox, make sure to drag it to the “Important” folder so Google knows we aren’t spam. And if you recommend the Weekly One Pager to a friend, you will have a thousand years of good luck (trust me, it’s empirically verified)