Tech Sell-Off: This All-Star Artificial Intelligence Stock Is Down 50% but Up 127,300% Since Its IPO. Here's Why It's a Screaming Buy Heading Into the Midterms.

The four-month sell-off in artificial intelligence (AI) chip stocks has been severe. Since June 30, the iShares Semiconductor ETF (NASDAQ: SOXX) has been down over 25%, with many components down much more.

Recently, one of the biggest winners of the AI boom sold off severely after its recent earnings report. Shares of this longtime outperformer are now down almost 50% from their all-time highs back in June.

But this winning stock has had similar drawdowns before. Yet shares are still up over 127,000% since its 1984 initial public offering. And that doesn't even include the company's growing dividend, which management just raised by 27%, bringing the current yield to 0.57%.

Lam Research is at the heart of AI

Semiconductor equipment manufacturer Lam Research (NASDAQ: LRCX) lies at the heart of AI. Its advanced etch and deposition equipment is critical to producing cutting-edge logic and memory chips. In particular, Lam specializes in producing "vertical" structures that require the delicate stacking of materials at the atomic level.

Years ago, the NAND flash industry began stacking memory cells on top of each other, and Lam went on to dominate machine sales to the NAND sector. Today, advanced logic semiconductor structures are increasingly "vertical," such as gate-all-around transistors. Even DRAM (dynamic random-access memory) is going vertical through both high bandwidth memory (HBM) and "3D DRAM" structures.

AI-level growth on display in September

Lam Research's strength was on display last quarter. Revenue surged 54.5% to $8.22 billion, while adjusted (non-GAAP) earnings per share rallied 77.4% to $2.20. Both figures were ahead of analyst expectations and were records for Lam.

Yet the stock fell after the report. This may be due to guidance of "only" $8.6 billion in revenue and $2.33 in earnings per share for the quarter ending Dec. 27.

That would normally be really strong guidance, but in the current AI environment, investors might have been hoping for more.

Why this quarter's guidance shouldn't matter

The reason for the so-called "light" guidance isn't lower demand, but rather tight supply. Management pointed out that there isn't enough clean room space available for its machines, and shortages of certain components for Lam's machines will prevent the company from reaching its full potential this quarter. CFO Doug Bettinger said guidance would have been "a bit higher" if not for these constraints.

Meanwhile, management forecasts industrywide wafer front-end equipment (WFE) spending will rise to over $200 billion in 2027, up from about $155 billion this year. Keep in mind, Lam's management sees this projection as a supply-constrained number as well. Moreover, management didn't say how much more than $200 billion. Given that we are still in October 2026, this is likely a very conservative number.

A WFE figure of over $200 billion implies about a 30% growth rate for Lam next year if it grows in line with the industry; however, growth could be much higher. Lam should also grow faster than the overall WFE, given its importance in the execution of vertical chip structures. Finally, Lam has a record of expanding margins during up cycles, and it's also buying back stock. Therefore, Lam's earnings per share should rise even more, as it did last quarter.

Image source: Getty Images.

And Lam has another growth driver ahead: New NAND

Interestingly, much of the growth Lam has seen over the past couple of years has come from advanced logic and DRAM, as these have been the main beneficiaries of the AI build-out. However, NAND flash, where Lam has the strongest position, may soon have its moment to shine.

Over the past decade, NAND flash makers have been packing as many bits as possible into a memory cell and then stacking as many modules as feasible onto a chip.

Lam has benefited from advanced stacking of NAND modules, usually through machine upgrades. However, this production efficiency has limited overall wafer growth in the NAND industry and, therefore, the need for new NAND machines.

However, AI infrastructure companies may soon move toward something called SLC NAND (single-level cell NAND), which stores just one bit per memory cell. AI companies are doing this because SLC NAND has lower latency, meaning it's faster. For applications in which speed is paramount, such as agentic AI, SLC may be the best option, even though it stores fewer bits.

The emergence of less-dense NAND may then require more NAND wafer production. Lam's management said that every 1% shift of NAND capacity to SLC requires a 2% increase in NAND wafer volume.

On the recent conference call with analysts, Lam CEO Tim Archer said, "NAND is yet to emerge as to its full potential from our perspective, and I think you are starting to hear about it and the strategic role it can play in AI."

In the buy zone again

After this drawdown, Lam currently trades at just 23.9 times the average analyst estimate for Lam's fiscal 2027, which ends on June 28, 2027. Based on earnings per share for the full calendar year 2027, which should be higher, Lam's valuation may well be under 20, or below that of the overall market.

Calendar 2027 should be a massive year for the semiconductor equipment industry. Though it's unclear what will happen in 2028 and beyond, given recent commentary about industrywide supply constraints, and given the fact that NAND investment may soon reaccelerate, Lam's discounted valuation seems like a low price for this all-star chip stock. After this summer's pullback, Lam Research is once again a buy.

Billy Duberstein and/or his clients have positions in Lam Research. The Motley Fool has positions in and recommends Lam Research and iShares Trust-iShares Semiconductor ETF. The Motley Fool has a disclosure policy.