Our Charts & Crafts episode on Ajinomoto hit 1M views in less than 24 hours yesterday. It’s a truly fascinating story. Many folks have asked for deeper reading on the topic, so here are a few things I thought you might enjoy:

1. Notes from Ajinomoto’s Earnings Call - coincidentally posted yesterday (I did not plan this, the universe is just conspiring on our behalf)

2. A behind-the-scenes look into my process creating a chart and a series like this

3. More reading (sources I used etc) if you’re into Ajinomoto

Let’s ride.

1. Ajinomoto Earnings - Q12026

   1. The revenue for the specific division that does all the chip film (ABF) grew 54% year over year when it was expected to do 30-40%. My video is already out of date because I said 35%…which was true for last quarter (AI industry moves FAST man)

      1. it was the biggest contributor to the whole company’s growth (from a profit POV) this quarter

They only expect this to continue to accelerate as the chips industry scales

Next, a little BTS (we’re going to call these BTC - Behind The Chart, ok?) on my video process:

When I posted the first video in this series (how Toto - the japanese toilet company - is winning in AI by making ceramic wafers inside 3DNand machines), someone commented with this chart:

Naturally, it piqued my curiosity about Ajinomoto so I looked into it a bit. Then, a very smart friend came over for dinner last weekend who also mentioned Ajinomoto’s sneaky role in this space, so I got to work researching.

For every Charts & Crafts video, I do multi-hours (sometimes days) of research. As I read and watch, I scribble notes and drawings in my dotted notebook like this:

Then once I feel like I have a thorough grasp of the topic in my brain and a set of visually-pleasing images to go with it, I practice some elements on scratch paper. This will surprise most people: I don’t often practice the whole thing - I really enjoy the thrill of doing it live on camera for the first time (I’m sick, I know. It’s my own version of sport). Here’s the practice page where I tested colors and made sure I could draw a convincing plastic bag:

I also really enjoyed getting my MSG prop for the video. I am dead serious - if anyone has great recipes, I’m all ears. I don’t often cook with MSG but I am excited to as umami is my absolute favorite flavor profile.

I am planning more videos in this series. Teasers on what comes next: a company with a name that you definitely know from the 90’s but I bet you don’t think is cool anymore (spoiler alert it is and the reason is underwater cables), and a German company that makes beautiful eyeglasses and also…really important chip stuff. You get the theme.

Finally - for those who want to go do their own research on Ajinomoto:

Sources & further reading

1. The origin story (Ajinomoto’s own telling mostly):

   • “Sprinting a Marathon”: Ajinomoto’s official ABF story. The 1996 knock on the door, and the researcher who gambled on a refrigerated resin and gave up his weekends to test it.

   • The ABF Innovation Story. Covers the 1970s amino acid research and why CPUs suddenly needed a new material.

   • Ajinomoto’s Unexpected Role in Semiconductors. Why ink failed, straight from the engineers.

   • Electronic Materials investor deck (PDF). How the film actually gets turned into substrates.

   • Shigeo Nakamura’s bio. The researcher who invented ABF now runs the entire company.

2. The monopoly and the money:

   • Palliser Capital’s activist campaign. The hedge fund pushing Ajinomoto to raise prices and break out ABF as its own segment.

   • “How an MSG Factory Holds Nvidia by the Throat”. ABF costs under 0.1% of a GPU’s price, and the supply gap is projected to run through 2028.

   • Ajinomoto’s record FY2025 results, an analyst deep dive. The segment numbers behind the 35% jump.

   • 50%+ margins and the new Gifu plant

3. The 2021 shortage:

   • Understanding the Global Chip Shortages, an SNV report (PDF)

   • The Unimicron factory fires.

   • How the ABF shortage hit GPUs

4. The AI supercycle:

   • Ajinomoto raises ABF prices 30%. Effective Q3 2026, with lead times stretching to 30 weeks.

   • The supply gap extends through 2027. Why AI chips need exponentially more film.

5. Earnings:

   • Ajinomoto Earnings Press Release

It’s called Until Labs. 

Until Labs is working on something called “reversible whole body cryopreservation”. 

That means cooling down the human body to down to negative 13 degrees Celsius and then, at any time in the future, warming it back up to a normal temp. 

While it sounds like science fiction, it’s not altogether new - we already cryopreserve human embryos during IVF then thaw and implant them in the future.

Scaling this process up to a whole body isn’t yet possible though. In the interim, the company is building toward something a bit simpler and very mission oriented: Organ transplants. 

Right now, there are more than 100,000 people in the US waiting to receive an organ donation. Of those, 17 die every single day. Transport of these organs is a huge issue that involves cargo planes, trains and automobiles because for some, like the lungs, the are only viable for 4 hours. Thousands of good donations go to waste while 100,000 wait. 

Until Labs wants to add time to the clock. 

But here’s what they are up against. When you take an organ down to negative 13 degrees, there’s a “danger zone” to traverse between -3 and -13. In this zone, ice forms on the organ that prevents it from functioning when rewarmed. 

Until’s tech aims to solve for this in a few ways: 

1. - Prepping the organ by replacing the blood with a protective fluid - this is called perfusion

2. - Protecting the organ from ice using “cryoprotective agents” or molecules that prevent ice from forming

3. - Reviving the organ during its thaw by rapidly re-warming it through the danger zone more quickly 

The most impressive part may not be the bold vision - its the prodigy behind it. CEO Laura Deming is only 30 years old, an MIT Dropout with a lifelong longevity fascination who started working in biology labs at age 8. She just raised $58million dollars - let’s see if she can pull it off. 

That’s what naysayers say to new American semiconductor company Substrate. Led by a Thiel fellow and backed with $100M, they’re tackling perhaps the hardest part of the chip making process - lithography - using a novel technology using x-ray light and particle accelerators that they hope will unseat both ASML and TSMC on the path to (re) building American chip foundry dominance. 

To understand what substrate is doing, we first have to understand lithography. 

Lithography is critical to making microchips. To understand lithography, we first have to understand their power source - light - which lives on a wavelength spectrum from radio to gamma rays, with the waves getting tighter all the way

Each silicon microchip we use works because it has been etched with patterns using one of these powerful light sources - that process is called lithography.

Today, we do EUV lithography: extreme UV light is beamed via high powered lasers inside a $25B bus-sized device to melt and direct drops of molten tin through mirrors and lenses to etch designs on silicon wafers.

There’s only one company -  ASML- that makes EUV lithography machines good enough to make advanced microchips, like NVIDIA Blackwells. ASML has 100% market share, virtually zero competitors, a $414B market cap, and are based in the Netherlands. Many think their global monopoly is an existential risk. 

For the last 5 years, the founders of Substrate have been working on a way to etch an even more complex pattern onto the same size piece of silicon - how? They use the tighter wavelengths from X-Ray light instead of UV. Using a particle accelerator instead of a laser, Substrate directs x-ray light to etch teeny tiny patterns onto silicon wafers. 

More patterns matter because it creates more information per nanometer -hyper-critical in AI chips.. 

Lithography is only one part of the chip making process, though. Chips go from design, to wafer prep, to lithography, to being built and finally tested and packaged. Eventually, Substrate wants to compete with an even bigger player than ASML - TSMC.

TSMC holds the vast majority of the global chip foundry market and boasts a $1.3T market cap. 

This creates a huge issue for the US though - TSMC is based on the island of Taiwan, a mere 100 miles from mainland China. 

Substrate has a bold vision to bring this process back to the United States and has raised more than $100M to try. The founder James Proud has said himself that they might only have a 1% chance of success, but that’s it’s a worthwhile bet for a prize this big and a national cause this important. 

It’s a nuclear company called Oklo that’s all over the news lately, if you know where to look.

They are building something crazy - a tiny nuclear reactor the size of a hot tub.

We all know we need clean energy, but nuclear power often gets a bad rep - let’s break down the difference between traditional nuclear fission power plants and what Oklo is planning.

Traditional nuclear power plants can cost $35B to stand up and take more than a decade to build. They span thousands of acres, require thousands of people to maintain and waste over a thousand tons of energy.

The Oklo Aurora reactor is planned to cost only $60M, be the size of a hot tub, built underground, with 0 human maintenance, and fueled by recycled nuclear waste - so turning something that was a huge cost into an advantage.

Sounds pretty good right? Sam Altman thought so too - he spent the majority of the last decade as the company’s chairman.

Here’s the timeline.

After founding in 2013, they went through Y-Combinator in 2014, met and got a seed check from Sam, who then became chairman in 2015. In 2024, he helped them IPO via SPAC, then stepped down in 2025. People speculate it’s not because he doesn’t believe, but instead so they can partner with OpenAI and others without conflict. This September, the company was selected as 1 of 4 companies to join a Department of Energy pilot to speed up nuclear power.

Then, on October 21, the Department of Energy sent out an application for any interested companies to apply for up to 19 metric tons (or 42,000 lbs) of plutonium - the stuff that’s inside nuclear weapons - that we keep in reserve.

Oklo is one of the companies that checked the box for yes.

They will recycle that plutonium as a temporary fuel for their reactors while they wait for the government to unlock the uranium supply they will ultimately need.

How about the business model. The back of the envelope math is: 1 reactor creates 50 mega watts of power, utilized at 80% all year, which means 350 megawatts of power a year. They’ll charge an average of $100/ megawatt, totalling $35M a year in revenue.

A single Aurora reactor can power 33,000 homes, 1 large data center, and 1 military base. That’s not for a year - that’s for 20. It promises to last that long without maintenance.

Since their debut, Oklo’s stock is up a staggering 650%, but this week it took a turn - it dropped 16% because famed public investor Cathie Wood’s ARK Invest sold off 53000 shares citing too much hype - no revenues, only LOIs and a long time horizon still to go.

The first reactor is supposed to go live in 2027. Let’s watch what happens next.

It’s a nuclear company called Oklo that’s all over the news lately, if you know where to look.

They are building something crazy - a tiny nuclear reactor the size of a hot tub.

We all know we need clean energy, but nuclear power often gets a bad rep - let’s break down the difference between traditional nuclear fission power plants and what Oklo is planning.

Traditional nuclear power plants can cost $35B to stand up and take more than a decade to build. They span thousands of acres, require thousands of people to maintain and waste over a thousand tons of energy.

The Oklo Aurora reactor is planned to cost only $60M, be the size of a hot tub, built underground, with 0 human maintenance, and fueled by recycled nuclear waste - so turning something that was a huge cost into an advantage.

Sounds pretty good right? Sam Altman thought so too - he spent the majority of the last decade as the company’s chairman.

Here’s the timeline.

After founding in 2013, they went through Y-Combinator in 2014, met and got a seed check from Sam, who then became chairman in 2015. In 2024, he helped them IPO via SPAC, then stepped down in 2025. People speculate it’s not because he doesn’t believe, but instead so they can partner with OpenAI and others without conflict. This September, the company was selected as 1 of 4 companies to join a Department of Energy pilot to speed up nuclear power.

Then, on October 21, the Department of Energy sent out an application for any interested companies to apply for up to 19 metric tons (or 42,000 lbs) of plutonium - the stuff that’s inside nuclear weapons - that we keep in reserve.

Oklo is one of the companies that checked the box for yes.

They will recycle that plutonium as a temporary fuel for their reactors while they wait for the government to unlock the uranium supply they will ultimately need.

How about the business model. The back of the envelope math is: 1 reactor creates 50 mega watts of power, utilized at 80% all year, which means 350 megawatts of power a year. They’ll charge an average of $100/ megawatt, totalling $35M a year in revenue.

A single Aurora reactor can power 33,000 homes, 1 large data center, and 1 military base. That’s not for a year - that’s for 20. It promises to last that long without maintenance.

Since their debut, Oklo’s stock is up a staggering 650%, but this week it took a turn - it dropped 16% because famed public investor Cathie Wood’s ARK Invest sold off 53000 shares citing too much hype - no revenues, only LOIs and a long time horizon still to go.

The first reactor is supposed to go live in 2027. Let’s watch what happens next.

Sign up now so you don’t miss the first issue.

Subscribe now

In the meantime, tell your friends!