⚠ Heavy R&D Is a Bet That Can MissLow threat

Enormous R&D spending is the engine of the moat, but it is also a standing risk: research on the outer edge of physics is expensive, long-dated, and not guaranteed to work. ASML's EUV bet happened to pay off spectacularly, but it took twenty years and could have failed; the next set of bets — pushing High-NA, and whatever succeeds it toward the ultimate physical limits of lithography — carries the same uncertainty. Each generation is harder and more expensive than the last, and the laws of physics do not guarantee that a viable successor to today's technology exists at an economic cost.

The subtler risk is that lithography eventually approaches limits that even ASML's research cannot economically overcome, or that the industry finds ways to keep advancing chips (new architectures, advanced packaging, chiplets) that lean less on ever-finer lithography and more on other techniques. ASML's spending keeps it ahead so long as the game remains 'print smaller features,' but the billions poured into research are a bet that the game continues on that axis and that the physics keeps yielding. It has always yielded so far, and betting against ASML's engineers has been a losing trade for two decades. But heavy, long-horizon R&D is a wager, not a certainty, and the sums at stake grow with every node. The precedent took two decades: EUV nearly didn't make it¹.