Yucca Moth

In 2013, TSMC's chairman Morris Chang made a $10 billion bet on a single customer. He committed to building 20nm fab capacity on Apple's promise to fill those lines. 'I bet the company, but I didn't think I would lose,' he later said. By 2020, neither company could leave the partnership. Apple had no alternative—Samsung's 3nm yields ran 30-40% versus TSMC's 80%+. TSMC couldn't afford to lose Apple—iPhone chips brought 25% of total revenue and filled 70%+ of 3nm capacity. This is obligate mutualism, and the yucca moth perfected it tens of millions of years ago.

The yucca moth is the only insect that actively pollinates its host. Most pollinators stumble into pollen transfer accidentally while feeding on nectar. The yucca moth does neither. It evolved specialized tentacle-like mouthparts (modified maxillary palps) specifically to collect yucca pollen, roll it into a ball, and deliberately pack it onto the plant's stigma. This is active, intentional pollination—the insect equivalent of a dedicated contract manufacturer.

The relationship is obligate in both directions. Yucca plants cannot be pollinated by any other species—their flower structure excludes generalist pollinators. Yucca moths cannot reproduce anywhere else—their larvae feed exclusively on developing yucca seeds. If either partner disappeared, the other would face extinction. Over 22 species of yucca moths have co-speciated with over 35 species of yuccas, creating matched pairs of biodiversity.

Darwin recognized this immediately. In an 1877 letter to botanist Joseph Hooker, he called the yucca-yucca moth interaction 'the most wonderful case of fertilisation ever published.' Modern research (Smith & Leebens-Mack, 2024) confirms it remains the archetypical example of coevolved intimate mutualism after 150 years of study.

The Apple-TSMC relationship exhibits identical dynamics. Apple designs chips but cannot manufacture them—no other foundry matches TSMC's yields. TSMC manufactures chips but needs Apple's volumes—no other customer provides comparable demand predictability. Apple orders are known three years in advance, allowing TSMC to plan capital expenditure with utility-company confidence. The switching cost for Apple to leave: $2-5 billion in redesign and requalification alone.

Both systems enforce honesty through mutual vulnerability. The moth could theoretically deposit eggs without pollinating, but the plant would abort those fruits (host-sanctions). Some yucca moth species have evolved into 'cheaters' that lost pollinating mouthparts entirely—but these cheaters remain parasites on the mutualism, dependent on legitimate pollinators maintaining the system. In business terms: a supplier that cheats a critical customer destroys the relationship that makes cheating profitable.

Obligate mutualism creates path-dependence that looks like lock-in but functions as alignment. TSMC cannot betray Apple to Samsung—Samsung is a competitor. Apple cannot threaten to leave TSMC—there's nowhere to go. When partners can only prosper if the other prospers, cheating becomes self-defeating.

Two-thirds of yucca moth species pollinate only a single yucca host. This extreme specialization seems risky—extinction of one means extinction of both. But specialization eliminates monitoring costs. The moth doesn't need to evaluate multiple partners; the plant doesn't need to screen multiple pollinators. In a world of uncertainty, committed partnerships reduce transaction costs to near zero.