September 2025 Random Ramblings
Summary
Andrew Walker sees a switch from quarterly to semiannual reporting as a market-structure event, not a long-term-compounding event. If Q1 and Q3 releases disappear, May and November options could lose embedded earnings volatility while August and March gain it, creating potentially “free alpha” for quant shops and larger post-earnings overshoots for patient stock-pickers.
Elon Musk’s roughly $1 billion open-market Tesla purchase was extraordinary less for its bullish message than for its unprecedented scale. Walker could find no comparable corporate-insider transaction: Dustin Moskovitz accumulated roughly $1 billion of ASAN over time, while Berkshire’s $3 billion OXY Form 4 reflected its status as a greater-than-10%-owning financial buyer rather than a corporate insider.
The historical record of enormous insider purchases is surprisingly poor, so conviction should not be confused with foresight. ASAN fell from near $100—Walker thinks Moskovitz began buying in the $80s—toward $14, OXY lagged, and Patrick Soon-Shiong’s $45 million NantKwest investment was disastrous. CEOs can be “getting high on their own supply.”
Sector-wide insider buying is a lens Walker increasingly prefers to one spectacular transaction. He highlights regional banks after the 2023 failures and busted biotech earlier in 2025, where directors and executives bought meaningful amounts across the group: “You could throw a dart at any regional bank” and find insiders purchasing.
Time spent investing is not automatically improvement. Walker contrasts years of “mindless practice” in golf and video games with writing, publishing, and revisiting old work, arguing that embarrassment at something written two or five years ago is welcome evidence that the learning curve remains alive.
September 2025’s meme-stock surge raises a process-versus-outcome problem: how seriously should one take an investor whose implausible thesis has produced a five-bagger? Walker admits he discounts people who own companies whose technology appears to be “smoke and mirrors,” alongside ATM issuance and insider selling, yet worries about judging an entire investor by one “pot of magic beans” position.
Deep dive
1. Semiannual reporting would redistribute volatility—and create alpha opportunities
Walker would prefer quarterly reporting, partly because the UK’s semiannual regime can produce “BS trading updates” with selective disclosures. Yet for genuinely long-term investors, he does not believe quarterly versus semiannual reporting materially changes the value of a great compounder.
The investable question is how a moderate rule change alters the game. His analogy is basketball’s three-point line: teams that adopted it faster gained a structural advantage, and even moving the line by two feet changed strategy. Reporting frequency could similarly create alpha opportunities as the market structure changes.
His proposed trade: if Q1 and Q3 reports disappear, sell May and November call options whose prices still embed those earnings events, then buy August and March options before six-month Q2 and full-year disclosures. Those releases would cover six months rather than three and could be “lit,” while the removed quarters would lose volatility.
The second-order effects could reach multi-manager “pod shops” built around whisper numbers, pre-earnings positioning, and four annual catalysts. Even without trading options, Walker would hold more dry powder: when a favored stock next reports after six months of information, a miss and bad guidance update could make it “more likely to overshoot to the downside.”
2. Musk’s Tesla purchase sits in a category of its own
Musk bought roughly $1 billion of Tesla on the open market on September 12 and filed on September 15. Walker accepts the standard interpretation—“people don’t do that if they’re bearish”—but stresses that token insider buying and a billion-dollar personal commitment are entirely different signals, even given Musk’s wealth. Musk’s earlier Tesla purchases were smaller and generally alongside equity offerings.
The nearest comparison he found was Dustin Moskovitz at Asana, ticker ASAN: a $350 million purchase directly from the company in 2022 plus roughly $1 billion of open-market buying under 10b5 plans from late 2021 through the end of 2022. Even that total accumulated over about a year rather than in one transaction.
Berkshire reported a $3 billion OXY purchase on one Form 4 in early 2022 and may have bought $6 billion to $7 billion overall, but Walker rejects it as a direct comparison: Berkshire was a financial buyer filing because it owned more than 10%. Musk’s purchase was the largest corporate-insider transaction Walker could identify.
3. Giant insider conviction has often produced poor results
Moskovitz began buying ASAN near $100, Walker thinks in the $80s; by the recording it was around $14. Walker calls the program “some of the worst capital allocation of all time,” a warning that founder conviction does not guarantee a good outcome.
Berkshire’s OXY investment looked less disastrous but still unimpressive: the shares were roughly flat since early 2022 and had underperformed both the S&P 500 and the largest oil and gas stocks. Buffett may be “the goat,” Walker says, but the purchases do not exactly “cover him in glory.”
Patrick Soon-Shiong’s NantKwest transaction was another cautionary specimen. He invested $45 million through a private placement near $12 per share while the accompanying public offering priced around $9—a roughly 30% premium—yet the company subsequently performed terribly.
There are important winners. Jamie Dimon’s $26 million JPM purchase in early 2016 compounded to almost 7x with dividends versus roughly 3x for the indices; Harold Hamm bought $200 million of CLR around $17 during 2020 and later took it private in the low-to-mid $70s. Tilman Fertitta’s roughly $27 million Wynn purchase in late March or early April 2025 also looked well timed, though Walker says it is early and Fertitta is not really an insider. Walker’s conclusion remains deliberately mixed: “Big insider buys don’t always mean super-bullish stuff.”
4. Sector-wide buying is Walker’s preferred lens
Walker increasingly prefers clusters of transactions across a distressed sector to a celebrated purchase at one company. After Silicon Valley Bank failed and First Republic was absorbed by JPMorgan in 2023, regional banks traded below book value while directors and CEOs bought across the group.
The individual purchases were smaller but personally meaningful: multiple directors sometimes committed a full year of board fees, often making their first-ever open-market buys, while CEOs invested hundreds of thousands of dollars. That across-the-board insider buying reinforced his bullish regional-bank view.
He saw a similar pattern in busted biotech earlier this year, another of only two examples he says worked well, while conceding he “wish[ed] I had leaned harder into it.” He presents the broad pattern as a lens worth following rather than treating one billionaire’s confidence as decisive.
5. Deliberate feedback separates learning from repeated activity
Walker’s personal warning is “mindless practice.” In grade school he visited the driving range three or four times weekly, played 18 holes once a week, and walked nine holes once or twice a week, yet remained worse than peers practicing comparable amounts. Video games produced the same realization: time accumulated without much improvement.
An old Todd Combs–Ted Weschler interview sharpened the question. One of them, already in his mid-50s, described joining Berkshire and working with Buffett for a year or two as the steepest learning curve of his career. Walker wants to know the mechanism—perhaps the depth of Buffett’s questions—so he can model it rather than merely admire it.
The unresolved social challenge is pushing oneself and collaborators without becoming “a completely condescending know-it-all butthole.” Berkshire appears both demanding and collaborative; Walker is searching for a way to reproduce that combination, where people feel challenged, happier, and demonstrably better.
His best existing mechanism is public output. Preparing ramblings forces clearer thought, while rereading blog posts from ten, five, or even two years ago often embarrasses him. He welcomes that discomfort: the investor and writer he is now “would write something better” and think more rigorously.
6. Meme-stock winners still have to defend their process
September 2025 is, in Walker’s phrase, “shitco stock season”: companies with no revenue, a bad or nonexistent business model, and high short interest have gone parabolic, leaving short sellers despondent. He jokes that if he owns such a company, its stock has been getting killed.
He distinguishes legitimate controversy from situations where the technology appears to be “smoke and mirrors,” insiders continually sell, the company hits the ATM, accounting looks shady, or controlling managers appear to be bad actors. When serious investors own those names, he instinctively takes them less seriously.
Walker knows that response may be unfair. If nine positions support intelligent discussion but the tenth depends on “magic beans” growing into a beanstalk, should that single thesis discredit the investor—especially after it becomes a 5x, 10x, or 20x winner? “Maybe it’s a failure of understanding on my end,” he concedes.
The harder problem arrives with Q3 letters and victory laps, from retail claims of turning a YOLO position into 100x in 24 hours to serious investors saying they are up 400% or more. Outcomes cannot settle whether the original reasoning was sound, so Walker returns to disaggregating track records: separating repeatable judgment from one spectacular result.