Is Oklo a good bet?

Oklo is building a next gen small module reactor that reminds me of Ford's Model T. It won't be the first nuclear reactor, but there's a decent chance that it will be the first to go mainstream.

There are several reasons for this.

If you'd like to learn more, I recommend checking out the investor day deck and presentation. They made it super approachable and have a great way of telling the story. My goal with this post is to organize my own thoughts. These wree the questions I had and I think you'll find the answers are pretty exciting.

Industries like nuclear are incredibly difficult to stand up. However, there are a couple of reasons why today might be special.

Oklo is building an owner operator business where they sell electricity to customers directly through 20-40 year Power Purchase Agreements (PPAs). This is attractive to data centers and utilities who do not want to be in the Nuclear game and are willing to pay more for clean and reliable energy.

It also has other benefits as well:

The best way to compare costs across power plants is to look at the levelized cost of energy (LCOE) over the lifetime of the plant. Oklo believes that they can bring the LCOE on a per MWh basis down from $90 to $40 as they go from their first power house (FOAK) to the 20th (NOAK).

There are several assumptions which I list below, but the $40 MWh is exciting because it would be cheaper than renewables + battery and natural gas with carbon capture. Straight solar is in the $20s, onshore wind is in the $30s, and natural gass is in the $40s, so $40 MWh would put them in the ball park of natural gas and would be reliable in a way that renewables are not.

Importantly, their pipeline is now 2.1 GW and on good terms. For example, they recently added Equinix

they claim that these customers are not sensitive on price.

Time will tell though since these are non-commital, but it is easy to see how data centers who are running up against energy availability concerns would be eager to standup an alternative to renewables and natural gas as they scale their power demands.

Oklo's assumptions on LCOE:

When you listen to the CEO, Jacob DeWitte, talk about Oklo's future, he keeps coming back to the potential of fuel recycling. There are a couple of statistics that bring this to life.

Recycling has the potential to reduce their fuel costs by 80%. Given that 50% of the cost of a plant comes from fuel, you can see how important this might be, but there is a second point that's potentially more profound.

The U.S. currently has 90k tons of waste fuel and generates 2K tons of waste fuel a year. Fun fact, the US government currently pays a million dollars a day to store it. And here's the kicker, current reactors only use 5% of the fissile material, so if you scale's Oklo's design and fed it only waste fuel you'd have enough energy to power the US at current demand for a century. One year's waste is enough to power current demand for four years.

And it gets more fantastical, with the uranium believed to be underground, there would be enough energy to power 10 billion people for 10 billion years.

Coming back down to earth, the opportunity to reduce their fuel costs by 80% is the strongest concrete argument that they'll be able to significantly reduce their costs as they scale. Other arguments include standardization and partnerships opportunities which simplify operations and maintenance. When combined with financing opportunities, it's possible to see how each power plant could have a short payback time and help them scale fairly efficiently.

And then there's always the physics argument that fuel's like coal and gas have an energy density of 20 and 50 MJ/kg while uranium has an energy density of 3,900,000. If you follow this argument to its logical conclusion, it leads you to cheap abundant energy.

Over the long term, there's always the question of how well they'll be able to defend their margins. I don't understand the energy sector well enough, but if we're in a world where SMRs are replacing coal and natural gas as the primary baseload power, shrinking margins should be a first world problem andit was only 2013 when Exxon was the world's most valuable company!

It's easy to look at a company like Oklo and see many of the risks they face, but their february investor day webinar and deck did a good job of showing me that they're taking them seriously and there's a good chance that the market has priced in how much they have been de-risked.

Oklo is working on attaining a custom combined license (COLA) for design, construction, and operation of the Idaho National Lab demonstration plant. This application will serve as a reference COLA and will make it easier to receive subsequent COLAs in the future.

They have had a long relationship with the NRC starting in 2016, an initial application in 2020 for the Idaho National lab project which was rejected in 2022, and will be re-sumitted early next year with the goal of being approved early 2027. At this point 10% of Oklo's team is former NRC employees and there's no getting around the fact that planning and licensing is a major component of what they do.

With that said, there is evidence that the NRC has been reforming over the past decade with each administration bringing in more supportive leadership and updates they've been making to the licensing process.

Two examples of this are the combined licensing application which lets Oklo get permitting for the design and construction at the same time and their new subsequent review process which treats the initial application as a reference and avoids re-reviewing redundant sections. Because of Oklo's owner operator model where they both design and build the plants, they were the first to take advantage of the COLA benefits. And when it comes time to file the second COLA, they expect that it will take 6 months to complete.

At this point, there's significant evidence that nuclear is one of the rare bi-partisan issues. The IRA made it possible for nuclear firms to claim a 30% investment tax credit and introduced the idea of transferability which would make them re-sellable. There are additional benefits such as production tax credits, plants that are built along side retiring coal sites, etc etc. There's also a 900 million dollar Investing in America fund which serve as loans for Gen III SMR Light Water Reactors. Oklo is building a Gen IV plant, but this is a good example of additional financing opportunities that could become available to help facilitate the development of the first N plants [5].

And for what it's worth, Chris Wright, Trump's nominee to lead the department of energy, is on Oklo's board and the chairman of the board is Sam Altman so there's no shortage of friends in high places!

The Oklo team points to a couple of factors that re-assured me that public sentiment might be shifting towards YIMBYism.

I don't know the science well enough to claim to speak on the safety concerns, but the point that impressed me was that they inherited a safe-by-design from the ERB-II reactor which operated for 30 years and generated 20 MWe.

The important feature is that it does require an external shut-down actuation. The system will self-stabilize, self-control, and cool by natural forces. The CEO, Jacob DeWitte, talked about their safety by physics approach which is another way of saying that they did not need to add any extra safety systems in. As the core gets hotter the fuel will naturally separate and the reaction will slow down. If the vessel were to heat up, hot air will rise and cold air will flow down. Moreover, the plant does not even require an on-site operator.

While not directly related to safety, they also made the point that safety measures are often the biggest cost and compliance drivers so the safe-by-physics design is a big efficiency driver.

This is another area where I don't feel very qualified to speak, but here is my high level understanding.

My hunch is that the biggest risk that a company like Oklo has is time and their ability to raise on good terms if there are significant delays. Short of a Cruise style disaster which is minimal, the worst case scenario is that there are delays and cost overruns that would cause them to miss their 2027 timeline for completing the Idaho National Laboratory project and the three other projects which are in the works: two in Ohio and one in alaska.

At the moment, their balance sheet looks strong and they have a track record of hitting their goals, but that could always change [6].

There are two cautionary tales that are worth mentioning here. The first is Georgia Vogtle nuclear reactor which was expected to take 8 years, but ended up taking 14 years and was expected to cost 13 billion, but ended up costing 32 billion.

The second is NuScale's first deployment, the Carbon Free Power Project (CFPP), where they partnered with the Utah Associated Municipal Power Systems (UAMPS) to develop a small modular reactor (SMR)..

The project aimed to construct a 720-megawatt electric (MWe) facility at the Idaho National Laboratory near Idaho Falls, Idaho, comprising twelve 60 MWe nuclear power modules. The project was approved in 2020 when the DOE agreed to a 1.4B cost-share with the goal of being completed by 2030 and a target cost of $5.3b and LCOE projection of $58/MWh and was terminated in 2023 when the projection was revised to 9.3 billion and $89/MWh and there was less demand from the utilities than expected.

Both of these stories stress the risk of building a first of a kind nuclear facility. With the CFPP, some of the factors were post-pandemic supply chain disruptions, but many others came from engineering adjustments required by the regulatory process that increased the complexity and cost of the project.

Let’s put aside the speculative nature of the bet. It's a given that there are many assumptions here, revenue is at least two years out, and profits are many years after that.

The more interesting question is will nuclear be able to earn its place alongside other mature energy sources. Today, nuclear accounts for about 8% of US energy production, but there has not been a new nuclear project started after 1975 (the date when the NRC was founded).

Put another way, we're talking about going from 0 new gigawatts of energy to 200 in the next 30 years. To do this successfully, we'll need to build an entire ecosystem alongside the reactor and do it quickly. This will require significant government and private investment coordination.

It's totally possible that any one of these systems could fail and already mature technologies like Solar and Natural gas could eat nuclear's lunch. For instance, more efficient carbon capture or battery technology could further push their cost curves down.

On the other hand, one only has to look at how the cost for solar has plummeted over the past 50 years leading to nearly 600 new GW of power being produced every year to believe that scaling laws and economies of scale do work [7].

Oklo is a ten year old company that's built on the back of 30 year old tech. Getting in today, two years before they complete their first plant, could be the opportunity to get in on the ground floor of the next Exxon. And the good thing about markets is that they tend to pull future forward, so the price will not wait for nuclear to win to move.

1 DOE: Pathways to commercial liftoff

2 Oklo investor day deck

3 An ideal solution for replacing coal

4 Impact of GenAI on energy

5 Investing in america

6 Oklo balance sheet

7 A surprise solar boom

8 Nuclear Energy Could Contribute Meaningfully To Securing Net-Zero Emissions

9 Tech companies are investing in nuclear