Published November 18, 2026 · by Troels Højbjerg, Founder
Almost any company can call itself an AI company. A list of "AI stocks" is therefore always somebody's choice. I wanted a group I could define the same way every time, so I took the chip makers and the makers of chip equipment among the 200 US stocks I cover on this site.
That gives 16 stocks, from NVIDIA and Broadcom to smaller names like Astera Labs and Credo. Some are foreign companies listed in the US, such as Taiwan Semiconductor, ASML and SK Hynix. I left out Microsoft, Alphabet and the other big cloud companies: they spend heavily on AI, but most of their profit comes from other things.
Here are all 16, measured from November 17, 2025 to November 17, 2026. "Biggest drop" is the largest fall from a high to a later low inside those 12 months.
| Stock | 12 months | Biggest drop | Below 12-month high | Signal today |
|---|---|---|---|---|
| NVIDIA NVDA | −17% | −35% | −35% | Strong Sell |
| Taiwan Semiconductor TSM | +29% | −27% | −26% | Sell |
| Broadcom AVGO | 0% | −39% | −29% | Strong Sell |
| SK Hynix SKHY | −18%* | −39%* | −38% | Strong Sell |
| AMD AMD | +109% | −38% | −23% | Hold |
| Micron Technology MU | +194% | −46% | −41% | Sell |
| ASML ASML | +36% | −32% | −30% | Sell |
| Intel INTC | +59% | −63% | −61% | Strong Sell |
| Applied Materials AMAT | +54% | −56% | −52% | Strong Sell |
| Lam Research LRCX | +43% | −57% | −51% | Strong Sell |
| Arm ARM | +31% | −62% | −58% | Strong Sell |
| Marvell Technology MRVL | +78% | −58% | −53% | Sell |
| Teradyne TER | +40% | −56% | −52% | Strong Sell |
| Astera Labs ALAB | +50% | −63% | −56% | Strong Sell |
| Credo Technology CRDO | −15% | −63% | −59% | Strong Sell |
| Skyworks Solutions SWKS | +40% | −32% | −6% | Strong Buy |
| S&P 500 | +10% | −9% | −6% | – |
Source: own calculation on daily closing prices, November 17, 2025 to November 17, 2026. *My price data for SK Hynix in the US starts on July 9, 2026, so its numbers start there. Signal, momentum and quality as shown on the stock pages on November 18, 2026. Momentum and quality are hidden on small screens; they are on each stock page.
Measured stock by stock, the median gain was 40%. Micron rose 194% and AMD 109%. NVIDIA, the best-known name, is down 17% over the same 12 months, and Credo is down 15%.
The S&P 500 never fell more than 9% from a high in the period. The median chip stock fell 56% at some point, and Intel, Astera Labs and Credo each fell 63%. Intel is up 59% over the year and still went through that drop.
The daily swings tell the same story. The median chip stock had 60 trading days out of 252 where it moved 5% or more, up or down. The S&P 500 did not have one. With moves like that, the day you buy matters a lot for how the year ends.
Most of the group topped out together. Eleven stocks set their 12-month high in June 2026, and 8 of those 11 are now more than 50% below it. NVIDIA, AMD and Taiwan Semiconductor kept rising until October 5-6, 2026, and have fallen since.
Across all 16, the median stock is 46% below its 12-month high. The S&P 500 is 6% below its own. The odd one out is Skyworks Solutions, only 6% below its high.
The signal on each stock page counts four things: is the price above its 50-day and its 200-day average, and is the MACD (a trend measure built from two moving averages) above zero on a daily and on a weekly basis. Each yes counts +1, each no −1. That puts 10 of the chip stocks on Strong Sell and 4 on Sell. AMD is on Hold, and Skyworks is the only one on Strong Buy.
A sell signal is not a forecast of a crash. When I recalculated the signal for 195 US stocks since 2023, the median stock was still up 1.7% sixty trading days after a new sell signal. The difference was in the risk: big drops came more often. The details are in When should I sell a stock?
The quality score ranks a company's profitability and financial health from 0 to 100 against all the stocks I cover. Nine of the 16 are at 90 or more, and Micron and SK Hynix are at 100. Intel is the weak one at 38.
Momentum ranks how strongly the price has moved compared with other stocks, also from 0 to 100. Here the picture turns: 15 of the 16 are under 60, and NVIDIA is at 29. Only Skyworks is high, at 89. NVIDIA can easily be a great company even if the stock is weak right now. The two are different questions.
I cannot hand you a list of the best AI stocks, because nobody can say which ones will rise next. What I can show is how the 16 US-listed chip stocks behaved. Over 12 months the median one rose 40%, four times the S&P 500, but it also fell 56% from a high along the way. Eleven of them peaked in June 2026, and today 14 of the 16 show a sell or strong sell signal. Most of the companies still score high on quality, while their momentum is weak.
One thing to check: for each chip stock you own, look at how far it is below its 12-month high and whether the price is above its 200-day average. Both are on the stock page.
All the US-listed chip stocks I track, ranked by score and momentum and updated after each close, are on Best semiconductor stocks, free and with no sign-up.
There is no official list of AI stocks, so I measured the 16 US-listed chip makers and chip equipment makers among the 200 US stocks I cover. Over the 12 months to November 17, 2026, the median one rose 40%, against 10% for the S&P 500. Along the way the median biggest drop was 56%. Today 14 of the 16 show a sell or strong sell signal on their stock page. The companies still score high on quality (9 of the 16 are at 90 or more out of 100), but only Skyworks Solutions has a buy signal right now. A strong company is not necessarily a good stock to buy today.
In the 12 months to November 17, 2026, the median chip stock in my group had 60 trading days where it moved 5% or more up or down. The S&P 500 had none. Eleven of the 16 stocks peaked in June 2026, and the median stock is now 46% below its 12-month high.
Read also: Best semiconductor stocks · When to sell a stock · What is a good P/E ratio? · How to compare two stocks · Buy signals today · Methodology
This is information, not individual investment advice. The numbers are our own calculations on market data and may contain errors. Past results are no guarantee of the future. You decide for yourself — and remember to spread your risk.