Will Taiwan Semiconductor Be Worth More Than Nvidia by 2030?

The market is still treating TSM as a pure “AI picks-and-shovels” beneficiary, but the more important second-order effect is that it is becoming the toll booth for capex across the entire hyperscaler ecosystem. That creates a stronger earnings convexity than the article implies: even if any single AI platform winner rotates, the foundry still monetizes the aggregate spend cycle. In practice, the near-term upside is less about unit growth and more about mix shift toward advanced nodes and packaging, which can expand dollar revenue faster than end-demand growth.

The biggest competitive implication is not that TSM will beat Nvidia in absolute market cap; it is that Nvidia’s supply-chain dependence increases TSM’s strategic pricing power over time. If leading-edge capacity remains tight, incremental benefits flow disproportionately to the manufacturing layer, while chip designers absorb more of the execution risk and working-capital burden. That said, the market may be underpricing how much of the current multiple already assumes multi-year AI capex persistence; if hyperscaler spend normalizes even modestly in 2026, TSM’s growth rate decelerates faster than the narrative suggests.

The key risk is not AI demand destruction in the abstract, but timing mismatch: TSM’s valuation can stay elevated for months, while customer capex digestion can hit sooner. Watch for any slowdown in cloud capex commentary from Amazon, Alphabet, or Microsoft over the next 1-2 quarters; that would pressure the stock before fundamentals roll over. The contrarian view is that TSM is probably the cleaner long than NVDA on risk-adjusted basis, but not necessarily cheap enough to be a high-conviction outright overweight if the AI trade broadens and then compresses.