Samsung's 2nm yield sits at 40 percent at the back-end wafer-test stage, up from 20 to 30 percent in early 2025. The industry mass-production threshold is 70 to 80 percent. TSMC's comparable 2nm initial yield is 60 percent. The Exynos 2600, the first design on Samsung's 2nm GAA node, enters mass production in November 2025.

The 40 percent figure is a back-end number, not a fab-floor yield. No front-end figure appears in the document, so the gap between wafer-start and test-pass is unquantified. Samsung's foundry share fell from 11.3 percent in Q4 2023 to 8.1 percent in Q4 2024 while TSMC climbed to 67.1 percent. A buyer evaluating a 2nm contract in 2026 is looking at a node that has not yet cleared the 70 percent line, and the only proof point Samsung can offer is a phone it ships to itself. TSMC's reported 15 percent price increase on advanced nodes is the one variable that could pull a design house toward Samsung before the yield story is finished.

No front-end 2nm yield, no wafer-start volume, no external customer name, and no Exynos 2600 clock or core count appear in the document. The Galaxy S26 launch in Q1 2026 is the next measurement.

The 40 percent reading is a back-end test number, not a fab-floor yield, so the true wafer-start loss is still invisible. Samsung's only 2nm proof point before an external customer signs is a phone it ships to itself in Q1 2026.

The back-end 40 percent figure does not tell a buyer how many wafers are lost at the front end. Until a front-end number is published, the 20-point gap to TSMC's 60 percent initial yield is a floor, not a ceiling.