Launch HN: Atomic Alchemy (YC W19) – Manufacturing Nuclear Medicine
I’m Thomas Eiden, founder of Atomic Alchemy.
Atomic Alchemy will manufacture nuclear medicine and the radioactive materials used to make it, using several compact nuclear reactors.
These will be the first privately-owned nuclear reactors for nuclear medicine production. They are merely higher-powered versions of previously-licensed reactor designs that currently reside at universities. These reactor designs are passively safe and cannot melt down. With improvements in modeling and simulation that have occurred in the last few years, it is quicker and cheaper than ever to license and construct such a facility. But to be clear—we won’t be manufacturing reactors—we’ll be a chemical/drug manufacturer in the same way that Delta Airlines doesn’t manufacture planes or airports, but is a transportation company.
Before starting Atomic Alchemy, I was the lead reactor core designer at the Advanced Test Reactor at Idaho National Laboratory. My main job each operating cycle was to arrange the reactor core in such a way to allow the United States Nuclear Navy to run successful material experiments for their next generation submarines and aircraft carriers. I’ve always been interested in production and efficiency, and the issues currently plaguing nuclear medicine production have been of great interest to me all the way back to when I operated a reactor in college.
Nuclear medicine is used in a wide variety of diagnostic imaging procedures and cancer treatments. The most common procedure is the radiocardiogram to diagnose cardiological issues, and brachytherapy for cancer treatment.
Unbeknownst to many, there is a critical shortage of nuclear medicine worldwide--right now. The main failure in the supply chain is the fact that the entire world’s feedstock for nuclear medicine primarily comes from six government-run reactors, most of which are over 45 years old and will be retired in the next 10 years. Additionally, these government-run reactors are scientific research reactors and are not set up to efficiently produce these materials.
I have a design for a manufacturing facility that combines the entire supply chain, from irradiation, to chemical purification, to medicine production. This will allow Atomic Alchemy to fill the void as more aging reactors are shut down and allow those that remain to focus on their true purpose—science.
Currently, the reactors, chemical purification, and pharmaceutical manufacturing facilities are all in separate facilities, sometimes oceans apart. By shrinking the entire supply chain into a single facility, manufacturing costs can be slashed by up to 50%, as regulatory, shipping, and myriad other costs associated with the radioactive material decaying in transit, is reduced.
The market for radioactive feedstock alone is well over a one billion dollar market worldwide, and is constrained by the current supply. As the standard of living abroad continues to improve, developing markets will demand even more nuclear medicine.
Looking forward to your feedback and questions.
43 comments
[ 2.4 ms ] story [ 90.6 ms ] threadPeople are very hesitant about nuclear power, but medical isotope manufacturing requires little enough nuclear fuel that it should be very safe. You sound like the right person to tackle this, and the market is definitely there. Good luck!
1. https://www.nature.com/news/reactor-shutdown-threatens-world...
Chalk River shutting down was a huge blow to production capacity, as it alone supplied about 20% of the world's demand for radioisotopes. In fact, Chalk River shut down in 2016, but the license was extended for two years to keep it on hot standby solely because of its importance to nuclear medicine supply. When they finally permanently shut it down in March of 2018, I believe the reactor was over 61 years old!
You can find articles like this at least once a year, warning of the next shortage: https://www.cardiovascularbusiness.com/topics/cardiovascular...
NorthStar has started producing Mo-99 in accelerators without fissioning targets. Other producers may start soon:
Competitors abound to produce key medical isotope
After decades with no US producer, a plethora of hopefuls vie for a share of the molybdenum-99 market.
https://physicstoday.scitation.org/do/10.1063/PT.6.2.2019022...
Your FAQ says "As a private, production-oriented company, we intend to vertically integrate much of the radioisotope process to reduce inefficiencies and build a modular facility that can be expanded as a stable supply of radioisotopes encourages their use and increases demand." How much demand elasticity do you foresee, if Atomic Alchemy and competitors can easily produce more Mo-99 than was historically consumed?
Regarding potential competition such as NorthStar or SHINE Med: Basically, the non-reactor technologies have very poor unit economics. The product is less pure (which is fun to say in the context that we're talking about "molly"), and the electricity costs alone for operating accelerators is on par with just purchasing nuclear fuel. And, for the same "fuel" costs, eight of the SHINE accelerators will produce about 4% of a single Atomic Alchemy reactor.
NorthStar's approach of converting Mo-98 to Mo-99 works, but is very inefficient from a nuclear physics standpoint, and requires the same chemical processing, if not more, due to its low purity. It's a band-aid and won't really be able to compete once more players enter the market.
I guess another way to put it is this: the common thread between everyone in the article (except BWXT) is that they are relying on government $$$, and government money is never free. The govt. $$$ is given out on the condition it be used for developing "alternative technologies." And it just so happens that most of these alternative ways are either inferior from a unit economics standpoint, or are a costly regulatory nightmare because licensing new nuclear technology is simply a nightmare.
Regarding demand elasticity: Demand is actually dropping every year as the supply decreases. The weekly demand for the largest market share of radioisotopes is about 3/4 of what it was earlier this decade. As the population around the world grows and ages, the demand will continue to grow. If the myriad other radioisotopes used in nuclear medicine were more abundant and had a more stable supply, the demand for them, from what I can tell, would skyrocket. So there is a ton of room to grow JUST TO MEET CURRENT DEMAND. If China and India were to use nuclear medicine on the same per capita basis that the United States does, the world market would double and then triple, respectively.
The NorthStar chemistry advantage looks (IMO) to be the absence of fission products, which are diverse in chemical behavior and radiotoxicity. I have no idea if that can sufficiently offset the complications that come from lower specific activity in the product.
Good luck with your endeavor! I love to see startups that deliver physical products and not just software, especially a product as important as this one.
Yes--the scaled-up PULSTAR would still need regulatory approval for siting and operation, but would not need a new design certification (which is the real dream killer). I've deliberately chosen to use only technology that is commercially available to get a product to market as quickly and cost-effectively as is achievable.
However, if my seed raise goes as well, I will have some margin to get a software person to help couple together some of the various modeling and simulation codes we use for reactor design.
Further down the line, I will definitely need some software engineers though. I have some novel improvements I want to make upon how material is inserted and removed from the core to be processed, and I will definitely need a custom software control system that talks to hardware to do that.
I make a note to check back with you when I make a posting that requires your expertise, and I will preferentially hire people like you who are passionate and enthusiastic. It's definitely easier to learn how nuclear tech works than it is to shed decades of crappy nuclear industry corporate culture.
Regarding the simulation codes, I worked (as a hobbyist) with OpenMC; but you are probably using something else, as OpenMC is more a teaching aid than a production code.
Additionally, he or she (and with a few colleagues) will develop all of the processing procedures for the myriad isotopes that the facility will produce. A lot of what's publicly available for processing and purification has been optimized and is trade secret.
Another interesting problem is how to turn more of your waste into wealth? Currently, Mo-99 (the radioisotope with the largest market share) is primarily produced by extracting the Mo-99 from fission products. Mo-99 is produced in 7% of fissions. Well, if I process out the Mo-99, how many times and how many different processes on that waste can I perform to extract out other radioisotopes before I have to dispose of it?
Early on I'll need a radiochemist with a certain level of hands-on experience because their work will directly tie into the regulatory process, but as I raise more money and build my team, I'm certain that a computation chemist can help answer that last question which could be a huge innovation.
My question is, how equipped are you to handle regulatory hurdles for international customers? It's one thing for a small firm to deal with US civil nuclear regulations, but how are you going to deal with the regulators of each foreign market you intend to export to? I imagine that this is something that the larger providers in the field use economies of scale to deal with, but could be a hurdle for a startup.
Regardless, that is an important issue, but one I don't have to have a solid answer to for several years.
Great username, btw. ;)
I mainly just wanted to say "Hello", since it's a small industry, and if you have quality issues at some point we will probably end up talking anyways (one of my jobs is maintaining calibration traceability for a major dose calibrator manufacturer).
I sent you a LinkedIn contact; it appears we have a mutual friend.
Do you have any interesting innovations you are considering, other than consolidating production facilities? I presume you are planning on sticking with the normal "cows and pigs" methods of delivery to pharmacies, for compatibility.
As an example, most research reactors have a facility called a "rabbit tube" which is similar to the pneumatic tubes at a bank. You can basically shoot little capsules in and out of the core at will this way. But most of these reactors only have one of these tubes, or none, and every experiment they move in or out of the core is done by hand with 20 foot poles.
I want to automate the entire process by making every irradiation facility in the core accessible via a pneumatic tube system, and have a software system automatically eject material that had "cooked" long enough and send it straight to the chemical processing rooms via this shuttle system. This would drastically reduce human performance errors, reduce personnel requirements, reduce radiation exposure, reduce the capex required to move highly radioactive material around the facility, and allow me to move things in and out of the core without first needing to shut down the reactor.
In the short term, I'll most likely be sticking to the "cow" method, but as I learn more about that end of the supply chain, I'm sure I'll find more places to push for innovation.
If this is correct, can you bring an already formed startup to YC? I have always assumed YC was targeted towards very young "I have a million dollar idea" type entrepreneurs, given that they require you to move to SA for three months for $150k in seed money.
In any event, best of luck.
With YC, you only need the idea, and you can form the company after you are accepted. And in my case, you can certainly bring an already-formed company to YC. In fact, a small number of companies accepted into YC are already profitable.
Roughly half of the companies that join YC are new companies that are just getting started (< 6 months old), and roughly half are more established companies. There are even a significant number of companies each batch that are pretty large (say, > 20 people).
Due to the simplicity of the reactor design and facility, and the low power it will operate at, it should take about 18 months for each of these approvals, and both can be (mostly) done in parallel. I anticipate production starting in 2024 or 2025. There are a variety of things that can speed this up or slow it down, but this is my best projection to date.
The Advanced Test Reactor is regulated by the Department of Energy, which is way more stodgy than the NRC can be, so I definitely appreciate the importance of navigating the regulatory environment carefully. It can take ten years if done horribly wrong.
The NRC has two guides for licensing non-power reactors: NUREG-1537 Parts 1 and 2. Part one lays out the format and information required in a license application, and part two is the review guide that the NRC uses to evaluate the application. So really, all of the questions AND answers are published. A prospective applicant just needs to say what they're doing, show how it meets safety criteria, and prove they're doing what they said they'd do when they build the thing.
It's obviously much more complex than that, but that's the gist of it.
* existing facilities
* already funded 50m+
* existing staff
* ironed out ideas
Were unable to get NRC clearance for a variety of things for an extended amount of time. Granted, this was an accelerator based method, not reactor-based. However, this leads me to my second point.
What kind of fuel are you planning on using for these reactors? From my understanding, the NRC has expressed interest in removing reactor-based methods for mediso generation and instead aiming for domestic accelerator-based production. Companies like North Star, Nordion, and the Mallinckrodt nuclear spinoff are aiming for similar goals and seem to have expectations aligned with mine. If it were truly as easy as getting a reactor designed, a couple million dollars, and six years, don't you think these massive DoE/DoD contractors would have done it by now? As you said yourself, the market is limited but the potential for growth is massive if we can get a stable supply.
Again, sorry for all the questions, take the time you need. I don't mean to come off as arrogant or inflammatory, just curious.
I'm using commercially available fuel that is available from fuel vendors. UO2 with zircaloy cladding.
As far as I'm aware, the NRC has no stated mandate for removing reactor-based methods. Or, if you have a source for that, I'd love to see it because that's a big deal. What you may be thinking of is the DOE/NNSA grants that are being given out for encouraging a domestic Mo-99 supply, which have strings attached which mandate it be for developing an "alternative" technology. I've found that the correlation is very strong between nuclear startups which have failed and which ones rely on government money.
I'd be curious why you think those projects have failed. I've had many discussions with other nuclear startups and friends at INL on this topic, and it's pretty nuanced.
Having been a DOE contractor, I have some... strong opinions about how nearly everyone in the industry operates or their approach to regulation. To refrain from ranting, I'll just say that it is my strong opinion that pretty much everyone does things "they way we've always have done them" and that everything is over-engineered and costs way more than it needs to. Most people I've spoken to seem to think it costs $100M to build hot cells. I know it doesn't cost that much to build a glorified box of concrete, even if that's what cost-plus entities have historically paid for them.
We could take this discussion offline as well, if you'd prefer. Then we can both be more specific and protect anonymity. Feel free to shoot me an email at info@atomicalchemy.us if you wish.
Here in Silicon Valley, there are a few nuclear startups working on various non-LWR concepts. One of them has done roughly the equivalent amount of work that NuScale has done but has only spent on the order of a few hundred thousand dollars, and with a team of less than 20.
So to me, it is clear that there are better and worse ways to go about things.
I have proof of the desperate need of of these materials with two LOIs valued at up to $100M.
For the first phase of this project, which is mostly regulatory, I'm raising 1.5 on 12.
There really are no larger players who already operate reactors other than the power industry (power reactors generally cannot be used for radioisotope production) or government-run research laboratories. Realistically there is nothing preventing a government entity from joining the fun, but they will never be competitive. South Korea has spent several billion dollars designing a new reactor model and developing fuel for it (which I tested in ATR a few years ago, oh the irony!). It will be dedicated to doing the exact same thing I am. They've been at it since 2012 and they haven't broken ground on their facility yet. Last I checked the project caught a bad case of NIMBY.
The second part of the answer is that most of the processes are trade secret. If you dig deep enough, you can find a journal article or two from several decades ago, and even some newer ones where a different approach is studied. However, most of the existing players have an optimized process protected by trade secret. It takes a lot of domain expertise spread over several hard science fields to effectively get into the business.