>> read in browser

How J.R. Simplot went from the largest seller of french fries to McDonald's in the 1960s to the largest shareholder of Micron Technology in the 1990s. 

Thanks for subscribing to SatPost.

Today, we will talk about J.R. Simplot, a potato farming billionaire who went on to build what is now the $220B semiconductor firm Micron Technology (an all-time business pivot with some lessons for the current AI gold rush).

Also this week:

• ChatGPT x Walmart
• MrBeast Thumbnails
• …and them wild posts (…including Toy Story 5)

This issue of SatPost is brought to you by Bearly AI

Are you looking for a no-logging, encrypted and anonymized AI chat tool?

Then try the Bearly AI research app, which provides an easy-to-use UX for the newest models from OpenAI, Anthropic, xAI, DeepSeek, Gemini and more.

While some AI providers keep chat logs indefinitely even after a user deletes them...Bearly AI has a default no-logging policy and user chats are encrypted while requests to LLM providers are completely anonymized.

Try one month FREE of the Pro Plan using the code BEARLY1.

I recently finished Chris Miller’s book Chip War, which traces the geopolitical history of the semiconductor industry (and also serves as a 474-page reminder that I misallocated my portfolio during COVID to animal-themed NFTs and 15-leg parlays instead of Nvidia).

My favourite bit from the book was about J.R. Simplot, an Idaho potato farmer who used his understanding of commodity markets to spot an opportunity in the dynamic random-access memory (DRAM) segment of the semiconductor industry.

And by “an Idaho potato farmer”, I mean “the person who became a billionaire supplying McDonald’s with 50% of its french fries and used to drive around the city of Boise with a license plate that read SPUD”.

Just absolute boss moves.

Simplot was born in 1909 and lived nearly a century, during which he never owned a computer (when asked about it during the launch of Windows 95, Simplot said, "Hell, boy, I came before the goddamn typewriter"). 

In 1980, he invested $1m into Micron Technology for a 40% stake, which would eventually make him a billionaire again just from memory chips (or as Russ Hanneman would say, “re-billion-izing”).

Simplot’s unlikely career pivot was widely covered in the 1990s when his wealth hit $5B (at a time when the richest person was Slick Willy Gates at $15B-ish):

• 1992: The Idaho Angel On Their Shoulders (The Washington Post)
• 1995: The Simplot Saga: How America’s French Fry King Made Billions More in Semiconductors (Fortune)
• 1996: From Mr. Spud to Mr. Chips (The New York Times)

While Simplot passed away in 2008, Micron Technology remains one of the largest employers in Boise, Idaho and a major semiconductor player with a market cap of $220B+. In fact, Micron and two South Korean firms (SK Hynix, Samsung) own 90%+ of the DRAM market. 

You can read the full story here including:

• How Simplot Built a Potato Empire
• The Invention of DRAM
• The Founding of Micron 
• Simplot's Understanding of Commodity Markets 
• Micron vs. Japanese Manufacturers
• The AI Age (Micron, Samsung, SK Hynix)

ChatGPT x Walmart

On the topic of career pivots, we need to talk about how Sam Altman is taking his experience running Y-Combinator (fundraising, dealmaking, allocating capital) to financial engineer OpenAI's $1T play for AI dominance. 

It all starts with the fact that ChatGPT is the most important consumer AI product and AI is basically propping up the entire US economy right now: AI data centre buildout accounts for ~2% of US GDP and was responsible for ~40% of GDP growth in the first half of 2025.

With 800m weekly users, ChatGPT is gonna hit the billy club soon along with Facebook, Instagram, TikTok, WhatsApp, Telegram, YouTube and — damnit — LinkedIn.

This is 3 years after launch. Insane.

The math doesn’t look to be mathing on the financials, though. For 2025, OpenAI is projected to lose $20B on revenue of $13B while it’s committed to 26GW of AI data centers worth more than $1T.

This is why Altman is throwing Olive Garden Never Ending Pasta Bowl levels of spaghetti at the wall, per the Financial Times:

OpenAI is working on new revenue lines, debt partnerships and further fundraising as part of a five-year plan to make good on the more than $1tn in spending it has pledged to create world-leading artificial intelligence.

OpenAI is planning on deals to serve governments and businesses with more bespoke products, creating more income from new shopping tools, and new sales from its video creation service Sora and AI agents, said multiple people familiar with the start-up’s efforts.

It is also weighing ways to cash in on its intellectual property by developing new AI infrastructure, making forays into online advertising and plans to launch consumer hardware products, including a new AI-powered personal assistant device, with former Apple star designer Jony Ive.

Over the past few months, it’s clear that the market wants to believe that Altman’s high-wire act will pan out and many tech stocks have popped on OpenAI announcements:

• Oracle received a $300B data centre commitment from OpenAI (ORCL gained +36% on news)
  
  
• Shopify and Etsy were added to ChatGPT checkout feature (SHOP +6%, ETSY +16%)
  
  
• Nvidia intends to invest $100B into OpenAI (NVDA +5%)
  
  
• SK Hynix and Samsung agree to partnerships with OpenAI for advanced memory chips (SK Hynix +10%, Samsung +4%)
  
  
• AMD agreed to sell OpenAI 6GW worth of chips for AI data centers and also gave the startup warrants worth 10% of AMD (AMD +23%)
  
  
• Broadcom announced a partnership to build a custom chip for OpenAI (AVGO +10%)

The most recent major OpenAI partnership came last Tuesday: shoppers will soon be able to buy non-perishable Walmart items directly from ChatGPT.

Did Walmart’s stock pump on the news? Is the Pope Catholic?

Of course it did. The Bentonville, Arkansas retailer took Altman’s pixie dust and gained 5% to a record market cap of $850B.

It’s the latest tech move by the world’s largest seller of individual bananas (true story) to take on Amazon (which it has sharply outperformed in the past 5 years: WMT +131%, AMZN +34%):

• Walmart’s e-commerce sales hit $100B in 2024; while still short of Amazon’s $480B, WMT’s online business is growing 20% a year (vs. 10% a year for AMZN).
  
  
• Online ads were $4.4B in 2024, up 30% YoY (while less than 1% of Walmart’s total sales, high-margins on ads make it 10% of Walmart’s operating profit).
  
  
• Walmart has turned its supercenters (which usually stock 120k SKUs) into online distribution hubs (100s of millions of SKUs now).
  
  
• Wall Street is recognizing the change and Walmart’s P/E is ~40, more than that of Amazon, Apple, Meta or Microsoft (WMT has outperformed all of them in the past year).

Walmart’s digital success is bolted onto the largest real-world retail operation, per The Economist:

• “With $680B in revenue [for 2024] and 2.1m workers, it is the largest company in the world on both measures.”
  
  
• “In America it takes in a 1/10th of all retail spending, excluding cars, and a quarter of the outlay on groceries.”
  
  
• “Walmart has some 5,000 stores across America; 90% of Americans live within 10 miles (16km) of one.”
  
  
• Uses its size to get major discounts and items are 4-5% cheaper than Target (and 8-10% cheaper than other grocers).

How will Amazon respond?

The $2.3T behemoth founded by Big Daddy Bezos is actually in a tricky position. While ChatGPT is an important (and perhaps dominant future) distribution channel, Amazon has spent decades building an e-commerce walled garden and has a lot more at risk than Walmart.

Still, Ben Thompson believes the trade-off is worth it:

Now the big question is: Will Amazon follow Walmart’s lead? The answer is not obvious (and Amazon isn’t saying). Unlike Walmart, Amazon is the dominant e-commerce site, so doing anything to encourage shoppers to go elsewhere could be counterproductive. Moreover, working with ChatGPT or another chatbot would be an admission that its own AI features aren’t cutting it. (When was the last time you used Amazon’s e-commerce chatbot, Rufus?) Plus, having more shoppers search outside Amazon’s site would likely undercut its gargantuan advertising business, which is largely made up of merchants and brands paying to boost their products in search results on its website.

On the other hand, if you believe a big chunk of online shopping in the future will take place through chatbots, refusing to work with OpenAI in 2025 could be akin to refusing to let Google index your website in 2000. What’s more, if Amazon were to do a big e-commerce deal with OpenAI, it could pave the way for OpenAI to someday become a large Amazon Web Services customer, just as OpenAI has signed up to use Oracle and even Google Cloud. AWS and OpenAI took a tiny step in that direction in August, when AWS finally started offering OpenAI models to customers—more than two and a half years after ChatGPT became a household name.

Amazon has successfully navigated a similar “frenemy” relationship with TikTok—a large AWS customer but also an increasingly significant competitor in e-commerce. Though Amazon competes with TikTok in online shopping, it decided to start placing shoppable ads on TikTok. That followed an internal debate about whether doing so could hurt Amazon by further boosting TikTok’s appeal as a shopping destination, as we reported last year.

If Amazon was purely a retail company, then CEO Andy Jassy’s calculus of whether to work with ChatGPT’s Instant Checkout might be simpler. But he’s leading a huge conglomerate where what’s good for one business can hurt others. That makes it tough to handicap his final choice.

Either way, I can’t wait until OpenAI goes public and we get this announcement on CNBC.

MrBeast Thumbnails

Bloomberg’s Lucas Shaw has a great profile on Jimmy “MrBeast” Donaldson, the absolute gorilla on YouTube with over 400m subscribers.

MrBeast has taken his massive audience and started selling them products he directly owns under the banner of MrBeast Industries.

In 2024, it made $450m and that was evenly split between videos (300 employees) and his Feastables chocolate bar brand (150 employees). MrBeast Industries also does a Lunchables competitor (Lunchly) but it’s not at the same scale.

The operation was valued at $5.2B after a recent funding round but is still running at a loss:

Right now, however, Beast Industries is hemorrhaging money. It’s had three years of losses, including more than $110 million in 2024. The viral videos account for all of it, overwhelming the profits from Feastables. Donaldson has been spending between $3 million and $4 million on every video he produces for the main YouTube channel, most of which lose money. In 2023, Beast spent $10 million to $15 million shooting videos it never released to the public because they weren’t up to its standards. He also lost tens of millions of dollars producing Beast Games, a popular show for Amazon Prime Video in which 1,000 people competed for $10 million by, among other things, moving a 10,000-pound boulder.

MrBeast recently hired 55-year old Jeff Housenbold, a veteran of Silicon Valley (eBay, Shutterfly) to professionalize MrBeast Industries. In recent weeks, they’ve announced plans to get into banking and mobile phone plans. These make a lot of sense. The lifetime value (LTV) for these industries are high. The offerings between competitors aren’t too different. So, the ability to acquire customers cheaply is a major advantage.

And here’s my favourite example of Housenbold finding some low-hanging fruit:

Donaldson has learned, though, that he doesn’t always need to spend more money to drive viewership. There isn’t a huge difference between giving away $1 million and $100,000; an average viewer sees both numbers as big and will tune in for the results. In fact, giving away $1 million sometimes suppresses viewership. “We definitely found the point where more money doesn’t equal more views,” Donaldson says. “If I were to give a random person on the street a million dollars, a lot of people who see that video, especially because we have a very international audience—they just don’t think it’s real.”

Final fun fact: MrBeast employs more than a dozen people to make YouTube thumbnails…and they really are the best to do it. INSTANT CLICKS RIGHT HERE.

Links and Memes

Some other links for your weekend consumption:

RIP Diane Keaton: The legendary actress died at 79. Her 50+ year career included a range of roles from drama (The Godfather trilogy) and comedy (Annie Hall, Father of the Bride, Something’s Gotta Give). She won Best Actress opposite Woody Allen in 1977’s Annie Hall and probably has the most iconic ending shot in film history.

***

United Airlines is offering Starlink Wi-Fi…on its mainline fleet. Last year, WSJ’s Joanna Stern explained why Starlink’s Wi-Fi is so much better than existing airplane Wi-Fi. TBH, crappy airplane Wi-Fi was one of the best ways for me to smartphone detox and actually finish books. I’ll need to find a new reading hack.

***

The World’s first Waymo DDoS attack…”50 people went to San Francisco’s longest dead-end street and all ordered a Waymo at the same time”. Great piece of performance art but also idiotic for protesting automation when you realize how many lives will be saved by autonomous vehicles.

***

OpenAI top 30 token users…aka people who have used over 1 trillion tokens including Duolingo, Salesforce, Indeed and Shopify.

***

Drake sued UMG for promoting Kendrick Lamar’s “Not Like Us”…but the judge dismissed the lawsuit because when Kendrick called Drake a ped0, he did so in the context of a rap beef. No, for real, the judge literally went line-by-line through all the songs and opined on the relevant lyrics. It’s an incredible read.

***

Speaking of edgy songs…absolutely nothing in the world will prepare you for the lyrics (or music video) for a raunchy country song called “Country Girls Make Do” that has gone extremely viral on TikTok because people make reaction videos of their family listening to it for the first time.

***

“Using a Urinal Wearing Khakis”…is one of the funniest clips I’ve seen in 2025 via comedian Andrew Rousso.

***

Spotify will post video podcasts from The Ringer on Netflix…both of these companies are competing with YouTube (which just dominated in total view time). A comparison I’ve seen thrown around is YouTube is Amazon while everyone else in e-commerce (eg. Shopify, Walmart, Etsy, Instacart, Google Shop) try to work together to win share. The reality is that YouTube’s monetization and distribution is so good that posting exclusive to Spotify/Netflix isn’t worth it (even Joe Rogan went back to YouTube).

***

Since the Gold ETF (GLD) launched in 2004…it has outperformed the S&P 500 (SPY) ETF. Holy smokes. This is what happens when the money printer goes brrrrr, boosting asset values instead of stocks advancing on merits of innovation.

***

The battle over rare earths…Peter Harrell gives a great overview of how China is weaponizing its monopoly over rare earth refining (and what the rest of the world can do about it).

***

Apple won US F1 TV rights with a $750m bid over 5 years…the annual rights spend of $150m is significantly more than $90m a year that ESPN was paying. This news comes a few months after Brad Pitt’s film F1 made an Apple-record $630m at the box office. Vincenzo Landino has a great breakdown of the F1 deal numbers.

…and them wild posts (including how OpenAI says it will allow adults to generate erotica content in the same week that Google announced it had used AI to discover a “new potential cancer therapy pathway”):

Written by me, Trung Phan.

Workweek Media Inc. 
1023 Springdale Road, STE 9E
Austin, TX 78721

Want to ruin my day? Unsubscribe.