What I Learned From One Of The World's Smartest Engineers
A summer at Antora Energy.
Hello 👋🏻
Over this summer I worked with Andrew’s Ponec, Co-founder and CEO of Antora Energy (a Bill Gates backed startup). I worked on fundraising, engineering, and operations.
I want to thank Andrew for taking a chance on me and giving me an unforgettable experience. He has a thinking capacity like I have never seen before in the climate space. They are one of the few companies that have a reasonable chance to make an impact on climate change. Antora is one of them.
What Attracted Me To Antora?
There are hundreds of companies working to solve climate change. Why did I decide to hound Andrew for a year to work at Antora?
Antora's approach focused on techno-economics. There are 2 components to techno-economics: Technology and economics. Antora’s approach is technology agnostic and aims to be cheaper than natural gas. To my surprise, Antora is one of the few companies that have both aspects to their approach.
Why is a technology-agnostic approach important?
Technology agnostic means not favoring any one technology over another. The common path for most climate companies (I have observed) goes as follows:
"I went to uni, got interested in {insert field}, got a Ph.D., worked on this concept for the last 10 years and now I have a company."
Imagine 50 climate companies in the same industry with the same type of path as mentioned above. There is no way that all 50 companies will have the best technology to solve the problem. Most build cool-sounding technology that does not scale. The Antora approach was intentional about not working to commercialize a specific technology. The team only cared about getting to the right answer.
Why are Economics important?
The world runs on money. If you are not cheaper than natural gas, most people will not buy from you. You can not scale a carbon-neutral/negative product without having superior economics.
How Did I Get Introduced To Antora?
About a year and a half ago, I learned solar and wind need energy storage to scale. During my research, I came across an Antora talk. The first point was the cost of natural gas to create a cost baseline for energy storage. From that point on, I knew the team was thinking about the problem the right way.
The Antora concept used basic analysis rather than relying on a breakthrough. The team started with the question "What are the main variables in an energy storage system?" The answers were materials, heat transfer, and heat-to-electricity conversion. I assume there was a deeper analysis, but this is the high-level perspective.
The materials thought process was "We need a material that can store 1,500C and is cheap." They looked up materials and put their properties in a spreadsheet. The team compared the options and picked the best one. The approach is obvious when you look back, but most energy storage companies didn’t do that. Most companies followed the status quo (e.g. molten salt). Another set of companies chose a material without evaluating all the options.
For the heat transfer method, the team looked at the options. They found radiative heat transfer to be the best. The method required no moving parts and worked with high-temperature carbon blocks.
For the heat to electricity engine, the team did the same thing. They looked at all the options and found thermophotovoltaic (TPV) cells. The team chose TPV because it had the scaling capacity of solar. You can work at small scales, medium scales, and large scales. This characteristic was not true of steam turbines (the existing method). TPV had no moving parts, steam turbines did.
Lesson 1: Be Solution Agnostic
Before talking to anyone from Antora, I learned a solid model for solving problems. Identify what a good solution looks like (e.g. solution requirements). Look at all the options and pick the one that aligns best with your requirements.
The Antora team had an advantage over other battery companies. None of the early people were battery scientists. They could not reason by analogy because they had no existing knowledge. The lack of knowledge helped them escape preconceived notions of what was right.
Antora did not fall into the trap of cool technology. They built up their approach from first principles.
Lesson 2: Build Your Cost and Performance Targets From First Principles. Don't Outsource Your Thinking.
Many founders say "{insert Big Name Organization} says energy storage should cost $50/kwh.” or "{insert Big Name Organization} says we need $1.5/kg of hydrogen“ or " {insert Big Name Organization} says carbon capture should cost $50/tonne of capture CO2.”
The founders use the number as their cost target. The issue with the reports is they set a target, but the reader does not know why the target is there. Not knowing how people came to their target may lead to working towards a goal that does not matter.
Imagine working on a technology for 10 years. You achieve your cost target made by the report. You later realize your cost target was wrong because customers are not buying. You would be better off spending a few hours on basic analysis instead of wasting your life on a bad target.
The same concept applies to product specifications. You have to understand why someone chose a particular cost/performance metric. If you do not seek understanding, you may waste your time on a target that does not matter.
Lesson 3: The Basic Tools For Analysis Are The Best Ones.
When calculating the cost of anything there are 3 important tools
Conservation of Energy
Conservation of Mass
The Magic Box
Lesson 3.1 Conservation of Energy
I knew the definition, but I never understood the practical use of the tool.
Andrew introduced me to the concept. I had a hypothesis for a profitable method of capturing CO2. Within 8 minutes, he invalidated something I thought about for months. Andrew showed me the energy input my process needed. I realized the economics did not work.
For every cost calculation, I recommend you look at the energy needed. If you don't know how to calculate the required energy, ChatGPT can help figure out what to calculate.
Lesson 3.2: Conservation of Mass
I knew you could not create matter or destroy it, but I did not understand the practical implications of the tool.
I found a company that takes CO2 and H2O (water) to make CH4 (natural gas). In theory, the concept sounds great. When I tried to figure out their cost, I used conservation of mass.
To make 1 mmbtu (unit of energy) of CH4, you need 58kg of CO2 and 5 kg of hydrogen.
The cheapest CO2 cost is $50/tonne (or $0.05/kg). 58kg * $0.05/kg = $2.9.
The cheapest hydrogen today is $1.5/kg. 5kg * $1.5/kg = $7.5.
Your final cost will be $10.4 of input per mmbtu. To compare, Henry Hub (natural gas price index) is at $2.55/mmbtu. By using conservation of mass, you realize the concept does not work.
Lesson 3.3: The Magic Box
When evaluating any company, Andrew uses what he calls The magic box. Inputs —> Magic box (on a whiteboard, he draws a box) —> Output.
Before believing any claim, you need to understand input and outputs. The magic box helps you identify any obvious gaps in a technology.
There was a company that claimed to use enzymes to capture CO2 to make cellulose. He started with his magic box. In the magic box, the only thing we know is the company uses enzymes. The first question Andrew asked was, “Where do they get their energy from?” The website and podcasts with the founders did not provide an answer. He pointed out to me enzymes are catalysts. Without the energy source, having enzymes doesn’t add much value.
As I spent more time looking into their technology, I found another problem using the magic box.
We start with CO2 and end with cellulose (many C6H12O6 molecules). They added hydrogen somehow. Where did it come from?
Another question is in their magic box, they capture CO2. Cell-free enzymes (one’s not living in bacteria or plants) can only create good yield in reactors. The founder mentioned they can do CO2 capture from the air. How do you capture CO2 from inside a reactor?
Instead of hearing a company's claims and accepting them at face value. Doing basic analysis helps find problems/gaps in the company's technology.
Lesson 4: Mechanics of Cost Reduction.
As you increase the volume of manufacturing, there are 2 ways cost reduction happens:
1. Higher volume cost reduction.
You buy a higher volume of inputs and manufacturers give you a better deal. The closer you are to their number 1 buyer, the better the treatment you will get. One component of better treatment is lower cost. There is no inherent cost reduction here. Manufacturers charge you more per unit when you buy small volumes.
2. Learning curve cost reduction.
The learning curve is an inherent cost reduction. The learning curve is cost reduction because of process optimizations. Improvements can be increasing efficiency, reducing material requirements, changing manufacturing processes, etc.
A common mistake is not separating cost factors based on the type of cost reduction. A learning curve model may tell you that a commodity (like Steel) costs less than the global market rate. For example, imagine the model said Steel should cost $50/tonne at scale. The global price is around $800/US ton. There is a clear mistake. These mistakes can give you a cost way below your real lowest cost.
I recommend you separate the cost reduction factors into the volume or learning curve.
For higher volume cost reductions, look up the global price index for the product. I found that ChatGPT can help you find the name of the index if you do not know it.
For learning curve cost reductions, you can apply a learning rate. Remember, the act of manufacturing more units does not guarantee process optimization. You have to find some way to reduce the cost. When making this model, ask yourself the question, how do I plan to get this cost down? You can start by breaking down the cost factors and creating hypotheses for how to reduce each factor.
To create a lower bound cost, the conservation of energy and mass should help. The energy and mass yield of each process step should help you get a more accurate final cost.
Lesson 5: People Took The Wrong Lesson From Wind And Solar
As explained in lesson 4. Increasing manufacturing reduces costs in 2 ways. You get better deals from suppliers and you improve your process.
In the case of solar, wind, and lithium-ion batteries this knowledge is enough. Investors miss an important component with technologies like electrolyzers or carbon capture. They are missing energy.
There are 2 main cost categories: Capital cost and operational cost.
Solar, wind, and batteries are capital-cost technologies. There is little to no money required to operate the devices. Electrolyzers and CO2 capture machines need energy for each unit of product.
Investors assume technologies that need energy have the same cost curve as solar. The assumption is incorrect. Increasing manufacturing volume has 0 impact on energy.
Investors miss energy because they reason by analogy (I will explain why in a future post). They saw a graph that showed lower cost after high volume. They assume every technology works like that. When you reason by analogy, you don’t learn what causes the outcome. When you don’t understand, you make investments based on an incorrect assumption.
This fundamental lack of understanding is why we had a bunch of money put into biofuels back in the 2010s. They did not understand the energy needed costs more than fossil fuels from the ground.
Highlight Moment
In my first week, Bijan taught me that hiking is glorified walking. After going on a hike with my roommates, I agree. Bijan introduced me to funnel cakes. They are good. The evil designer of funnel cakes knows that no one can finish one funnel cake alone. Bijan felt that torture.
Great insights Ahmed!