**The most surprising finding is that the biggest price multiples in this dataset are not mostly the famous AI chip names. Semiconductor equipment and niche makers in Japan and Korea rival or beat them, while the giants barely moved.** I measured each stock's highest close divided by its lowest close from 2012 to August 2025 (results in steps 4–6): - **NVIDIA** is first, at about 687× (0.26 to 179.27). That fits the AI-surge story. - **Lasertec (Tokyo)** is second, at about 443×. It is ahead of Wolfspeed (355×), AMD (130×) and Broadcom (149×). - **HANMI Semiconductor (Korea)** reached about 284×, and **Rorze Corporation (Japan)** about 283×. Both are obscure next to AMD or Broadcom. - **Advantest** (85×), **Disco** (75×) and **Tokyo Electron** (50×) also outran most well-known names. - **Samsung** moved only about 5.6× over the same period. **SK Hynix** was about 17×, and Renesas about 17×. So the largest "AI-era" winners include the picks-and-shovels equipment makers, which I would not have expected to outrank AMD and Broadcom by this measure. The largest memory and foundry-type names were comparatively flat. **Caveats:** - **The ratio is not a return.** It is the max close over the min close, and the high may come before the low. A drawdown stock can still show a large ratio. - **Prices are in local currency** (KRW, JPY, TWD and others), so absolute levels are not comparable. The ratios are unit-free and are comparable. - **Some rows look unreliable**, which may inflate a few extremes (step 0): - 461 rows have a close outside the day's high–low range. - 4,888 rows have zero volume. - 6,840 rows are flat days. - SK Hynix's maximum close is exactly 300,000.0, a suspiciously round number that could be a data artifact. - BrainChip's minimum of about 0.009 is a penny-stock price that makes its 230× ratio easy to exaggerate. - **The composition of the dataset is itself notable.** It holds 152 distinct stocks, one company per ticker, with no duplicate stock-date keys, and mixes global exchanges and very different price scales.