Editorials

Lam Research filed its fiscal 2026 10-K on August 7, 2026, capping a year of genuine operational accomplishment. The company generated $23.2 billion in revenue and $11.7 billion in gross profit, yielding a gross margin of 50.5%. Operating income reached $8.2 billion, pre-tax income $8.26 billion, and net income $7.3 billion. Diluted EPS came in at $5.76, up 39% year-over-year. These are the numbers of a company executing well in a favorable industry environment.

What matters now is whether investors are paying a reasonable price for this strength, and whether the company's capital allocation reflects the discipline of true value creation.

Lam's gross margin of 50.5% is the headline. Across four quarters, gross profit ranged from $2.65 billion to $3.48 billion--a spread of less than 6% in dollar terms. This consistency in absolute gross profit dollars, against quarterly revenues ranging from $5.32 billion to $6.72 billion, tells a disciplined story: the company is neither desperately maximizing volume at thin margins nor pricing opportunistically above what its customers will bear. It is extracting steady, predictable value from its equipment sales and aftermarket services.

Semiconductor equipment is a capital-intensive, specialized business. Lam's customers--memory makers, foundries, IDMs preparing for AI and advanced node migration--have limited alternatives and are willing to pay for equipment that advances their roadmaps. That dynamic shows in the gross margin. A 50% gross margin leaves room for substantial operating expense (R&D and SG&A together totaled $3.5 billion, or 15% of revenue), yet still yields an 35% operating margin. That is legitimate operational leverage.

EPS grew 39% year-over-year, and net income grew 35%. Revenue growth was 26% in the same period. These are impressive figures. The fiscal 2025 comparable generated $17.3 billion in revenue, $9.0 billion in gross profit, $5.4 billion in pre-tax income, and $5.4 billion in net income (basic EPS of $4.17, diluted of $4.15).

But here is the uncomfortable arithmetic: revenue growth of 26% generated EPS growth of 39% only because share count shrunk. Diluted weighted average shares were 1.261 billion in FY 2026 versus 1.275 billion in FY 2025. That is a 1.1% reduction. The real driver of 39% EPS growth was not pure operating leverage; it was buybacks layered on top of 26% organic growth.

Further, free cash flow fell 9.7% year-over-year. Operating cash flow was $5.86 billion (down from $6.47 billion), and capex was $966 million (up from $658 million). The company returned $5.72 billion to shareholders ($3.85 billion in buybacks, $1.27 billion in dividends), against free cash flow of $4.89 billion. That is sustainable in a strong year, but it leaves little cushion if semiconductor demand softens.

Lam trades at a P/E multiple of 54x diluted earnings. The company's net income has grown from $4.2 billion (FY 2024) to $5.4 billion (FY 2025) to $7.3 billion (FY 2026). If we assume flat earnings from here, the company would need to double earnings in the next three years just to trade at a market-average multiple of 18x--a path that requires perpetually accelerating revenue growth in a sector known for feast-and-famine cycles.

The bull case rests on AI-driven semi demand sustaining for a decade. The bear case, quieter but real, observes that semiconductor capex cycles have surprised investors before. Memories of the 2000-2001 collapse and the 2008-2009 recession loom. Lam is a leader, certainly, but it is also a leveraged bet on a single thesis: that AI infrastructure capex will not revert to historical means.

Earnings yield is 1.85%, owner earnings yield 1.73%. These are thin margins of safety. If pre-tax margins compress by just 5 percentage points (to 30% of revenue, still respectable), the earnings support at current prices evaporates. History suggests that at some point, they will compress.

Lam returned $5.72 billion to shareholders in FY 2026. Of that, $3.85 billion (67%) was share repurchase. The shares were bought back at an average price well above $300 per share, in a company with a book value of $9.97 per share. The price-to-book multiple of 31x is among the highest in industrials.

A conservative capital allocation discipline would ask: at 31x tangible book value, is $3.85 billion in annual buybacks a prudent use of free cash flow? The honest answer is that it depends on what the company could do instead. If organic growth continues at 25%+ and returns exceed cost of capital, buybacks at such multiples are defensible. If growth slows and returns normalize, they will be viewed, in retrospect, as value-destructive.

The dividend, by contrast, is measured. At $1.27 billion on $7.3 billion in net income, the payout ratio is 17%. It grew 13% year-over-year. This is prudent capital allocation--modest, predictable, and leaving room for surprise if business conditions change.

Lam Research did what strong companies do: it converted revenue growth into operating leverage, maintained pricing discipline, and generated substantial free cash. The 50% gross margin and 35% operating margin are genuine competitive moats in action.

But valuations at 54x P/E, 31x price-to-tangible-book, and 47x EV/EBIT compress all of that excellence into a single point-in-time calculation. The implicit assumption is that earnings will remain stable or grow further. Semiconductor cycles have historically moved in 4-7 year waves. We are, by most accounts, three to four years into the current upswing. When it ends--and history says it will--Lam's earnings will contract, perhaps sharply.

The stock fell from $433 in late June to $311 by early August. That 28% decline is the market's way of asking whether perpetual growth is a reliable assumption. It is not a judgment on Lam's operational excellence in FY 2026. It is a judgment on price.