What It Takes to Buy the Most Hated Stock — Five Arguments and the One Condition That Would Prove Me Wrong

When people you respect walk in opposite directions

What is happening around Adobe is unusual. Investors I genuinely respect are standing on exactly opposite sides.

Terry Smith, who runs one of the largest funds in Europe, says he cannot see how Adobe makes money from AI. Michael Burry — the Big Short skeptic — is long the stock. Oakmark, a legendary value fund, flatly calls Creative Cloud the de facto standard for professionals and students alike. Mohnish Pabrai argues that established software companies like Adobe have a real head start in the AI shift and are far more likely to benefit than get killed, and that the market is likely dead wrong about them.

I also have people around me telling me plainly that Adobe's business is worse. I understand why they say it. The growth rate came down.

In a situation like this, most people simply pick whichever side they already wanted to hear. I decided to fix the order of judgment instead. These are the five tests I run before buying a hated stock.

1. Can I reconstruct both arguments accurately?

The first gate isn't buying. It's understanding.

My standard is this: I have to be able to restate the opposing case so precisely that someone who holds it would nod along. For Adobe, that's three sentences. If AI produces the finished output directly, the value of the tool evaporates. The generation starting today grows up without ever learning Adobe. AI operating costs explode while customers expect those features included free.

I don't dispute any of the three. They are real risks. But there's a critical distinction hiding here: acknowledging a risk and assuming that risk has already materialized are completely different acts. The market is currently pricing the second one.

2. Is this a fear story or a broken business?

Separating the two is simpler than it sounds. Look at the results.

A broken business leaves fingerprints in the numbers. Revenue rolls over, margins collapse, customers leave. A fear story only leaves fingerprints on the stock chart.

Adobe fell 66% from its all-time high while posting record revenue and record profit quarter after quarter. Revenue grew 13% year over year last quarter, annual recurring revenue passed $27 billion, and AI product revenue more than tripled in 12 months.

This is not a company falling apart. It is a company priced like it is falling apart. That gap is precisely where opportunity lives.

The results could of course roll over later. But at that point the data changes, not the story — and when the data changes, the judgment has to change with it. That's test five.

3. Does the balance sheet absorb mistakes?

What kills companies in a major transition is not being wrong. It's not having the money to survive being wrong.

That's why I keep coming back to Blockbuster. People remember it as the company that missed streaming, but Blockbuster was actually doing well on its streaming pivot. What buried it was a debt-loaded balance sheet colliding with the financial crisis. With a stronger balance sheet, you might be watching something different tonight.

Run Adobe through the same lens and the picture inverts. Market cap $91 billion, enterprise value $104 billion, so roughly $13 billion of net debt. Free cash flow last year was $10.3 billion. A little over one year of cash wipes out the entire debt load.

What that cushion actually buys is time. Adobe doesn't have to be right about AI on day one. It can be wrong several times, keep funding AI research through the whole stretch, push free apps like Express to pull in new users, acquire companies like Semrush, and repurchase its own stock while it's cheap. Weak companies die during big transitions. Rich ones adapt.

4. A margin of safety is the price you put on being wrong

This is where I get misread most often. A margin of safety isn't timidity. It's a formal admission that your own estimate has an error band.

The assumptions I ran on Adobe sit below consensus. Revenue growth of 3/6/9% is less than half of what the company did last year, and free cash flow margins of 37/40/43% bracket a 10-year average of 39%. Even so, the output is a $400 to $890 range with a $595 midpoint. The stock is $227.

Here is the key point. I did not build a structure that only pays off if the optimistic case is right. The price has already fallen far enough that the most pessimistic 3% growth case still implies a 17.5% annual return. Meaning the bears can be partially right without me getting badly hurt. The full calculation is in Adobe at 9x Free Cash Flow.

Put inversely: the reason I demand a margin of safety is that the bears might be right. Rather than dismissing that possibility, I price for it.

5. Write down what would prove you wrong — before you buy

The last test is the most important. An investment thesis without a falsification condition isn't a thesis. It's a belief.

My condition on Adobe is specific. Decelerating growth is not enough. If revenue and profit start clearly and steadily declining — not for a quarter or two, but consistently over a meaningful stretch — that is when I concede that AI was a much bigger force than I expected.

Writing this down in advance matters because you can never write it down later. A falsification condition set after the stock has fallen is just a sentence that justifies the loss. The condition has to be fixed before you buy.

Not one stock — thirty

One last framing I want to leave.

I am not arguing that I'm right about Adobe. My claim is far more modest: if I hold thirty companies that look like this, I will probably do well. That's the whole thesis.

Buying hated stocks means being wrong repeatedly by design. What matters is not hitting each one, but building a structure at the price level where being right pays more than being wrong costs. When a name nobody would touch — Micron around $90, for instance — eventually delivers an outsized return, it isn't because someone nailed that single call. It's because the same filtering process kept running over and over until candidates like it surfaced.

In the same vein, I wrote about being laughed at for buying Intel at $17 in I Bought Intel at $17 While Everyone Called Me Stupid: What Contrarian Investing Actually Means, and about how I screen for these candidates in How to Find Stocks the Market Is Practically Giving Away: Hunting Near 52-Week Lows.

The hard part of investing is the stomach, not the numbers. Buying when everyone is scared is a question of temperament, not spreadsheets, which is exactly why I put the five steps above into writing before I buy. It removes the space where emotion would otherwise operate.

Right now everyone is fearful about Adobe. So I'm happy to be the one being greedy.

FAQ

Q: When investors I respect land on opposite sides, whose lead should I follow? A: Neither. Use their arguments as raw material, but reach the conclusion through your own assumptions. Terry Smith's cost concern and Pabrai's incumbency-advantage case are both testable hypotheses, and the next several quarters of margins and AI revenue growth will answer which one holds.

Q: Isn't "decelerating growth alone isn't a sell signal" a dangerous standard? A: That's exactly why the condition has to be written concretely. Mine is revenue and profit declining consistently over a meaningful stretch. If you write down something vague like "sell if the business gets worse," you'll either never sell or sell at random.

Q: Is contrarian investing just buying whatever is hated? A: No. Cheap and hated are different things. You want companies that are hated while their cash flow and moat remain intact. Plenty of companies are hated for entirely valid reasons, and those keep getting cheaper on the way to disappearing.

Q: I can't manage thirty positions. Does this approach still work? A: What matters more than position count is whether any single position bets everything on one assumption. Five names work fine if each has a clear falsification condition and a real margin of safety. Just size them more conservatively, since the impact of any one being wrong lands proportionally harder.