Samsung has moved beyond writing a check into Dutch startup Euclyd, now reportedly carrying the two-year-old company's design to the test-chip stage inside Samsung Foundry. The shift from investor to manufacturing partner is the real story.
The strategic logic is customer acquisition disguised as venture activity. Samsung Foundry sits a structural distance behind TSMC in leading-edge market share, and it cannot close that gap by waiting for mature clients to switch allegiance. TSMC's moat is trust: proven yield, predictable ramps, and a design ecosystem that mature customers rarely leave. So Samsung is fishing upstream, embedding itself with startups before they have a foundry relationship to defend. Fund the company, absorb the design, run the test silicon, and you convert an equity bet into a captive process-node customer that grows with you. If Euclyd scales, Samsung owns both the upside and the wafer volume.
The risk is that early-stage bets are noisy. Most startups never reach volume, test chips routinely fail to translate into production tape-outs, and a foundry stretched across too many speculative engagements dilutes the engineering attention its anchor customers demand. This model only pays off if a handful of these companies become real volume, which is far from guaranteed for a firm that has yet to disclose a shipping product.
For Japan, the read-through lands squarely on Rapidus. Tokyo's foundry ambition faces the same cold-start problem Samsung is trying to engineer around: leading-edge capacity is worthless without committed design customers, and Japan's fabless base is thin. Samsung's playbook, co-investing to manufacture, is a template Rapidus and its backers should study rather than assume that 2nm capability alone will summon demand. Japanese chip-equipment and materials suppliers, meanwhile, benefit from any credible second source to TSMC, since a healthier Samsung Foundry means more diversified tool and chemistry orders.
For Japanese enterprises and SIers designing custom silicon for edge AI or automotive systems, the practical signal is that foundries are becoming willing partners to smaller design houses, not just hyperscalers. That widens the field of viable manufacturing routes, but it also means procurement teams must weigh a foundry's financial entanglement with its customers when assessing long-term supply reliability.
