2025-11-15
Every cycle has its moment of comic relief, and this one came the instant Berkshire’s 13F revealed a new Alphabet position. A small one. A late one. A 1.6% “why not” entry tucked between a mountain of cash and yet another Apple trim. But the crowd didn’t see nuance. It saw the halo. And suddenly the same people who scream “there is no bubble” every five minutes found their miracle: Buffett bought Google. See? This isn’t a mania. This is value investing.
Reality couldn’t be further from that fantasy.
The truth is simple: Berkshire didn’t buy Google because the AI future suddenly became clear, or because they discovered a new religion in LLMs, or because Buffett finally “gets tech.” They bought it because Alphabet was a laggard — a rare moment when an AI-adjacent megacap wasn’t already priced as if it would rule the galaxy. Around $200, Alphabet wasn’t cheap, but it wasn’t insane either. It was a stalwart trading at something close to sanity while the rest of the AI complex floated several layers above Earth’s atmosphere. That’s all it was: a reasonable business at a tolerable price, purchased by deputies whose circle of competence is different from Buffett’s own, and whose job is to keep some exposure in tech without torching the portfolio.
And even that small size tells its own story. I don’t think they stopped at 1.6% because they wanted to keep the training wheels on; they stopped because the stock ran too far, too quickly. Had Alphabet continued to languish—had the court ruling not triggered a rerate or the AI narrative not reattached itself—they might very well have added more. Berkshire is slow, but it is not blind. When the price starts drifting away from value, their willingness to size up drifts with it. In this case, the runway disappeared before they could take off.
There’s another possibility people never consider, simply because they’re too busy worshiping the headline to think about the mechanics: that GOOGL may not have started as a 1.6% position at all. Berkshire’s filings only show the ending snapshot, not the path. For all we know, they had begun building something larger, or intended to, before the court ruling jolted the stock upward and widened the gap between price and value. Berkshire never chases, not even for names they like. When a stock they’re accumulating sprints ahead of their cost basis, they stop cold. Sometimes they even trim it back down to the size that still fits the math. If Alphabet had stayed depressed—if the market had continued punishing it for “missing AI”—I have no doubt the deputies would have added more. The constraint wasn’t conviction; it was valuation running ahead of their hands.
But markets never read the footnotes; they worship the headline. So people are buying Alphabet now, after it has already run nearly 50% from Berkshire’s cost basis, after the court ruling, after the rerate, after the narrative shift, after the halo. They’re not copying Berkshire – they’re buying from Berkshire’s shadow, paying full price for someone else’s margin of safety, confusing lateness with wisdom. It’s the same absurdity we saw with Apple: Berkshire built its position when the stock was cheap enough to make sense, and then trimmed into strength. Meanwhile the crowd treated every dip as a sacrament, insisting Buffett would “never sell” — right up until he sold.
The “forever” myth died with Apple. And it should have died a long time ago with the Washington Post. Buffett doesn’t hold stocks for life; he holds them until the math stops making sense. When price outruns value, he sells. No drama. No ceremony. No vows. Just arithmetic. People forget that because they need a father figure in the market, someone to project their fantasies of stability onto. But value investing is not about loyalty. It’s about refusing to be a fool when prices detach from reality.
And reality is precisely what the AI bubble is doing everything it can to avoid. Everywhere you look, capital has been shoveled into data centers, GPUs, model training, power infrastructure — all of it capitalized as if demand will compound forever. It’s easy to celebrate that while revenue lines are rising and the keynote lights are bright. But slips don’t announce themselves politely. They show up as empty racks, idled clusters, price/mix compression, lead-time collapses, and utilization plateaus. Then come the big baths: the write-downs, the impairments, the admissions that a decade of future growth was booked too early. That’s what bubbles do. They don’t just reset multiples; they force companies to confess the truth of their balance sheets.
When that moment comes — the true “fire sales are the norm” washout we haven’t seen in years — I won’t be anywhere near the M7. Because if a selloff of that scale hits them, it won’t mean they’re “finally cheap.” It will mean the thesis that levitated them has cracked. Mascots of a failed theme aren’t bargains. They’re relics.
And this is where my own rule keeps me sane: with always-on wipeout risk, any single stock must offer at least the possibility of doubling within three years. Not as fantasy, but as plausible math. Meta under $1T fit that. Nvidia at $5T does not. Alphabet at $3T does not. For these names to meet my hurdle, they would have to reach science-fiction levels of valuation — outcomes so absurd they would drag the entire S&P into outer space. When a business requires a comic-book future just to make today’s price worthwhile, the only rational response is to walk away.
But most people don’t do that. They cling to the “heroes” of the bull market — the pundits, the momentum prophets, the chart magicians — all of whom vanish the moment a real bear market hits. Every cycle has its darlings who strike the right tone at the right time by pure luck, and every cycle buries them. Survivorship bias keeps the illusion alive; the crowd believes in talent where there was only probability.
Yet here we are again: another bubble, another wave of experts, another chorus of “golden buying opportunities,” another tidal surge of certainty right before volatility reminds us how fragile conviction really is. The only difference now is that I’m old enough, and tired enough, to see the game for what it is. And I no longer feel the urge to step into the quicksand.
So no — Berkshire buying Google doesn’t move me. It doesn’t change my read of the AI bubble, or my stance, or my hurdle rates, or my refusal to buy megacaps at cosmic valuations. If anything, it confirms what I already knew: the market will twist any fragment of news into a validation of whatever fantasy it needs to survive another week.
When the bubble finally cracks, and the impairments roll in, and the bathhouse is clean — then I’ll look around. Maybe there will be survivors worth owning. Maybe not. But at least the prices will speak the truth again.
Until that day, I stay where the ground is solid.
And ironically, if this whole AI bubble handed me one lasting asset, it wasn't a stock at all. It was solid sanity — a small anchor I built for myself through clear reasoning and disciplined thinking, the only thing that keeps me upright while the rest of the market spins itself into delirium. And I’ll admit this much: the very technology inflating the madness is also helping me stay sane.
How ironic, indeed.