July 2026 Random Ramblings

Summary

• Walker’s core frame: active investing is “a game of arrogance,” and the hardest skill is knowing when the arrogance is no longer deserved. The base rate for a stock purchase is a market-like return, so every purchase asserts you know something the market doesn’t in a technically zero-sum game — “any alpha that you take has to be taken from somewhere.” The recurring question is when to say “it’s not the market, it’s me,” whether on one stock or on the generic 15–20-year fund examples he cites, such as 8% annualized against the S&P’s 10%.
• He offers a concrete heuristic for individual positions: the 3-year rule. If a stock has gone nowhere three years after purchase, “the answer’s probably you” — and he sketches the classic loss pattern where a “really good company at 15 times earnings” degrades year by year into “a restructuring play” while the multiple compresses from 15x to 10x to 6x.
• His most candid self-criticism: he saw AI inflecting in late 2025 — “this is really starting to change my workflows” — and didn’t pull the trade, partly because “I’m a value guy. I’m an event guy.” He now asks whether that was discipline or “a mental block,” noting SanDisk spun out in March/April 2025 at “like 0.5 times next year’s earnings” and became “an absolute screamer.”
• A tradeable structural point: with Nvidia at 7.5% of the S&P 500, an active manager benchmarked to that index who avoids Nvidia is “technically short Nvidia” — and short a huge slice of the index once AI exposure is aggregated. The mirror image explains why value investors’ peak outperformance ran around 2000–2004: they owned none of the collapsing internet companies that had become large index weights.
• His sold-too-early ledger is brutal: uniQure, a net-cash biotech, posted surprising Huntington’s results and ran from roughly $5–6 to a peak around $60, while he says he bought below net cash and sold around cash; it now trades around $40. Nebius, the former Yandex, was bought around 18–19 and sold in the high 20s 15 months ago; it now trades at 230. The open question he can’t resolve: “are you churning through the portfolio too fast… or are you getting impatient and not letting these theses play out?” — with the honest counterweight that many biotech wins were Phase 2 and Phase 3 trials coming up “heads instead of tails.”
• On London — his “favorite little emerging market” — Mitie (MTO) just became, he believes, the 11th £1bn-plus FTSE 250 takeover this year, amid a headline warning “We’re going to run out” of listed firms. Private equity is paying big premiums (EasyJet was taken out by Apollo, he believes), suggesting private-market value is much higher than public-market value — yet 11 deals is only 4–5% of the FTSE 250, so a concentrated manager probably owned none, and Walker’s own three-to-four London names “go nowhere.”
• The London trap crystallizes his broader theme: “Every investor wants to be invested in an inefficient market until they’re actually invested in an inefficient market.” With insider ownership low in many firms and management potentially loath to sell for fear of losing pensions, salaries, and board fees, the takeout catalyst becomes a coin flip — leaving activism, capitulation, or evolving his value principles as the unappealing options.

Deep dive

1. Buying a stock is an act of arrogance — the skill is knowing when to revoke it

• Walker’s framing: the base rate when buying a stock is roughly the market return, and investing is technically zero-sum — “everyone can’t generate alpha. Any alpha that you take has to be taken from somewhere.” Every purchase claims “I am smarter than the wisdom of the markets. I know something the market doesn’t,” whether the edge is knowing fundamentals cold or spotting a forced seller getting margin-called.
• The mirror question runs at two levels: a single position (“if it was at 10 and it goes to eight, when is that just the dips of the market versus the market is telling you something?”) and the generic long-term fund examples he sees — managers running 15–20 years at 8% annualized against the S&P’s 10%, for instance. When must the manager admit, “I was arrogant to launch this”?
• His partial defense of persistence: “I’m a better investor today than I was not just yesterday, but particularly 2 years ago, 5 years ago… Maybe the next 10 are different.” He explicitly doesn’t know the answer — “I don’t know the answer” is the honest hedge, not a rhetorical flourish.

2. The 3-year rule and the anatomy of a value trap

• Walker’s heuristic: “If you invest in a company and it’s been 3 years and the stock has kind of gone nowhere, it’s time to really look in the mirror… the answer’s probably you.”
• His self-drawn loss pattern, worth keeping verbatim in spirit: buy “a really good company at like 15 times earnings,” a year later it’s “probably wasn’t as good as I thought, but now it’s at 10 times earnings,” then “an okay company, but it’s six times earnings and they’re buying back stock,” and finally “this is a restructuring play” — the thesis quietly rewritten at each step down.

3. Evolving vs. capitulating: the missed AI trade and the Fundsmith question

• The self-indictment: late last year he “saw pretty clearly that AI was inflecting and accelerating” and blogged about it changing his workflows, yet “didn’t pull the trade on any AI trades” — partly “outside my wheelhouse,” partly “I’m a value guy. I’m an event guy, and I don’t see the value. I don’t see the events here.” The AI names then “went on a generational run.”
• His evidence that the opportunities were visible in value terms: Meta in late 2022 “generationally cheap”; Netflix in 2015–2016 “crazy cheap” around when John Malone said it had “gone past escape velocity”; SanDisk spun out in March/April 2025 “at like 0.5 times next year’s earnings.” Best anecdote as told: Micron internally debated seeking a 100% price increase from Apple for 3 years of supply, expecting Apple to negotiate it down to 50% — “Apple just signed on the dotted line instantly. That’s very un-Apple-like.”
• He contrasts this with value investors who abandoned their principles in late 1999 or early 2000 and then got their faces ripped off by internet stocks. The Fundsmith letter he references said they’d been underperforming while “sticking to our hardcore value investing principles” and would “start being responsive to the market” — momentum, algorithms. It “got dumped on in a lot of places… if I’m giving my priors, I would probably say rightly so” — but he can’t dismiss the question: “When were the value investing principles holding them back versus when are they just throwing their principles out the window to chase the momentum?”
• His anti-modeling stance gets a caveat: he avoids extensive models because “you kind of miss the forest for the trees”; a model can show 20x cash or a 20% free-cash-flow yield while the business is falling apart and the cash is disappearing. He concedes, however, “maybe if I had done a little bit of modeling work, I could have said, hey, I’m seeing an inflection.”

4. If you’re benchmarked and don’t own Nvidia, you’re short Nvidia

• The mechanics: Nvidia is 7.5% of the S&P 500, and “whether you like it or not as an active investor, you’re probably getting benchmarked against S&P 500” — so no position means “you’re technically short Nvidia… Nvidia goes up and you’re starting from behind the eight ball.” If Nvidia falls, avoiding it helps.
• The historical proof of the inverse: value investors’ peak outperformance was around 2000–2004 because they owned none of the internet companies that had become large index weights — indexes were down 20%, while some of those value investors were up 20%. His unresolved question: “do I mean to be, quote unquote, actively short these by not having this exposure?”

5. Sold-too-early forensics: uniQure, Nebius, and the counterfactual problem

• Reviewing an April 2025-era portfolio full of “absolute bangers”: uniQure, one of the biotechs trading well below cash, posted surprising results suggesting its drug might successfully treat Huntington’s disease. The stock went from roughly $5–6 to a peak around $60 and was around $40 when he spoke; Walker says he bought below net cash and sold “around cash. Did okay. Wish I had held it.”
• The Nebius case in full: the former Yandex — “the Russian Google” — was sanctioned, delisted from Nasdaq, and frozen for 18 months before a major reboot. “This is kind of neat for Andrew.” His cost basis was around 18–19, and he sold the last shares in the high 20s 15 months earlier; “stock’s at 230 today.” Smarter friends valued the sum of the parts far higher, plus “free upside optionality of the neocloud” — “I think they’ve been kind of proven right.”
• His genuine uncertainty, both sides stated: “are you churning through the portfolio too fast… or conversely — the money that is yours to be made, the situation you were here to invest in, has kind of played out.” Many of the biotech outcomes depended on Phase 2 and Phase 3 trials coming up “heads instead of tails,” and “maybe I’m not counterfactually hard enough.”
• He also questions the comparison set: Nevro has substantial cash but can morph into an AI play, which works especially well in a world where AI works. Net-cash holdings may perform well in most scenarios, whereas the AI-linked play may work mainly in this one.

6. London: an emerging market where the only exit seems to be a takeout

• The week’s data point: Mitie (MTO) was taken over at a big premium — “the 11th £1 billion-plus takeover from the FTSE 250 so far this year” — under the headline “We’re going to run out” of London-listed firms. EasyJet was also taken out by Apollo, he believes, at a big premium. These deals suggest private-market value is much higher than public-market value.
• The supply-demand puzzle: shrinking supply should boost remaining multiples — his analogy is Australia’s structural premium from pension funds forced into a limited stock pool — yet “the remaining companies don’t seem to be getting much of a boost.”
• The practical trap: 11 takeovers is only 4–5% of the FTSE 250, so a concentrated 10-stock manager probably owned none; Walker holds three-to-four London names, having sold one “out of frustration,” and says his stocks “go nowhere, go nowhere, go nowhere.” In many firms insider ownership is low, while management may be loath to sell because a sale could cost them pensions, salaries, and board fees. The takeout catalyst is therefore a coin flip.
• His options are activism, leaving for another market, or evolving his value principles. His closer: “Every investor wants to be invested in an inefficient market until they’re actually invested in an inefficient market” — to which the old-timers say, “Welcome to the party, pal.”