Here is one of the most stunning charts an investor is ever likely to come across.

What this chart shows is simple, yet baffling. The top line tracks the monthly revenue growth of a key Taiwanese company, a maker of the servers that power the AI revolution. You can see the growth curve bend upwards, accelerating past 100%, then 150%. The business has undergone a significant transformation, to the point where the company is compelled to relinquish some of the orders due to capacity constraints.

The bottom line shows what analysts predict for its earnings growth in the following year. It remains stubbornly flat. Stuck in the mid-teens.

It’s almost as if no one was watching. As if the daily reality of explosive sales had no bearing on the outlook for tomorrow. The forecasts seem automated, disconnected from the seismic shift happening within the company itself.

This strange habit of looking away from overwhelming evidence is not an isolated case. To truly understand its impact on how companies are valued, we can look at an even more telling example. This time, from Korea.

This second chart shows SK Hynix, the HBM (High Bandwidth Memory) king, whose 2025 earnings estimates have been revised up by 3x over the past two years. The business has delivered with a revenue growth trajectory similar to WiWynn above, and the analyst community has been forced to change numbers with every actual number climb. The current year’s numbers could still rise another 20%, but that’s not the point.

What is striking is that the consensus forecast for 2026 is barely 8% above 2025. This is a mega-cap covered by dozens of analysts with capacity fully booked and an almost monopoly in a key component going into the AI servers.

This is a company trading at 7–8x suppressed forward earnings, despite margins and growth that would have given any U.S. or Indian peer a 50–80x multiple. If nothing else, the company’s valuations make for an entertaining reading of analysts trying to justify their target price multiples (with “Samsung fear” liberally inserted to justify 2x PB).

Déjà Vu, Humbly Told

These charts remind this writer of earlier years in these same markets.

When I was heading research in Taiwan two decades ago, we did a study that was almost embarrassingly simple. Taiwan companies publish their monthly sales usually around the 8th to 10th. We took the annual growth of last month’s numbers, did no filtering, and simply picked the 10 fastest-growing firms for investments for the following four weeks until the next monthly picks.

No forecasts, no narrative, no model assumptions. 

And the result? A portfolio of those 10 firms outperformed the Taiwan Weighted Index by more than 15% per year going back 15 years.

Why? Because, as is evident in the charts above, forecasts in some markets don't lead. They follow.  In a market where forecasts mean-revert by default, consensus numbers move only when the facts change. The surprises come not from deep analysis, but from actual results. The market never prices in growth but keeps raising or lowering the current years’ forecasts with the next year's numbers at a standard rate away from the current year’s. It was true then. It appears true now.

Korea, however, had a different lesson.

When I led a research team in Seoul around the turn of the Century, I found something stranger: analysts didn’t try to lead or even react. They mirrored. They deferred. Korean corporate management, especially at the time, rarely showed concern for share prices. The one thing they expected, though, was loyalty and glowing recommendations from the analyst community. And they got it.

Quantitative proof of this arrived later. A global study I was part of revealed that Korea was tied for the lowest percentages of "sell" recommendations in the world. More importantly, it also had the lowest dispersion in earnings forecasts. It was as if analysts were perpetually clustered around corporate guidance, none daring to stray too far.

They weren’t wrong or right. They were just… close. That was the goal.

The Curious Case of Corporate Apathy

And that brings us to a broader puzzle. Why don’t these companies care more?

In Korea and Taiwan, you’ll rarely hear a CEO obsess over valuation. Compare that to the U.S., India, or even China. In these markets, executives know that valuation is oxygen. It fuels acquisitions, lowers the cost of capital, and attracts talent or corporate raiders. It matters.

In contrast, most large Korean and Taiwanese firms don’t raise capital regularly from equity markets. There’s little M&A. Valuation is not a strategic concern. And so the analysts don’t have incentives to project ambition either. They forecast defensively. They talk cautiously. Everyone under-promises.

But there is a deeper, more behavioral reason. It comes down to where you fall on the optimism-pessimism scale.

The Other Land of No “Sell”s - the US

The other country that shared the title for the lowest number of "sell" ratings was, surprisingly, the United States. But the reason was the polar opposite of Korea's. It wasn't deference to management. It was inherent, boundless optimism.

This cultural divide is crystal clear today. Taiwanese analysts see a company’s sales triple and still pencil in a quick reversal to mid-teens profit growth. In the US, for similar companies like Supermicro or Dell, analysts project high growth to continue, and often accelerate. The gap between the forecasts for Korea's Hynix and America's Micron or Nvidia is simply astonishing. One worldview expects the party to end abruptly; the other expects it to last for years.

Different Yardsticks for the Same Reality

There’s another problem here. The same analysts, same global banks, often use different yardsticks depending on where a stock is listed.

I've seen salespeople pitch the same semiconductor stock with different narratives: if it’s U.S.-listed, it’s a structural winner. If it’s Korea-listed, it’s a cyclical memory play.

In private conversations, they’ll admit this is just how the ecosystem behaves. The valuation spread isn't rational; it’s habitual. The result is persistent mispricing. And for patient investors, occasional windfalls.

But let’s be clear: this doesn’t mean the gap will close.

You can’t expect Korean or Taiwanese corporates to suddenly embrace Wall Street-style guidance. Nor will local analysts change overnight. These behaviors are shaped by decades of incentive design, capital market structure, and institutional memory.

Here’s the part that stings, at least for those of us who love a good narrative twist: in markets like Korea and Taiwan, unearthing these distortions rarely earns you a round of applause. No CNBC interview. No LinkedIn victory lap. No finance bro reposting your chart with emojis.

In contrast, if the same asymmetry were uncovered in the U.S., Europe, or India? You’d have thinkpieces, TV debates, and 40-slide decks circulating before lunch. There’s a quiet indifference in some markets to stories like these. Proofs will be there next week when the article fails to find even a fraction of the total audience in Korea or Taiwan, the topic point of this piece!

So yes, the gaps remain.

However, this means investors must discern which forecasts to trust and when to adjust them in light of the same global data, such as the sudden sovereign AI surge of the current year. In some markets, the signal is barely in the current results. In others, it’s already in the story.