Samsung Q2 2026 — Preliminary Guidance (Pre-Print)

Core Thesis

Samsung's preliminary Q2 2026 guidance is the clearest memory-market proof point that AI datacenter buildout demand is showing up as hard pricing power for HBM/DRAM suppliers, not just GPU vendors.

Key Numbers (Samsung's own preliminary guidance, per this report)

• Revenue (Q2 2026 preliminary): ~₩171 trillion.
• Operating profit (Q2 2026 preliminary): ~₩89.4 trillion — Samsung's largest-ever quarterly operating profit.
• ~19x YoY increase in operating profit; +56% QoQ vs. Q1 2026.
• Driver: "strong AI-related demand for high-bandwidth memory (HBM) and conventional DRAM, which has tightened supply and pushed prices higher" — Device Solutions (semiconductor) division is the primary profit engine.
• Labor-cost caveat: a recent labor agreement allocates 10.5% of operating profit to employee bonuses, which will reduce headline reported profit versus this preliminary figure.

Gaps / What's Not in This Source

• No foundry-specific detail (advanced-node utilization, HBM4 base-die ramp) — full segment breakdown is due with the complete Q2 report on July 30, 2026 (note: this trails the Jul-23 headline print date referenced elsewhere; treat Jul-23 as the preliminary/guidance-adjacent date and Jul-30 as the full segment-detail release per this source).

Why It Matters for Datacenters

HBM is the direct input that gates accelerator output (every GB200/GB300-class GPU needs HBM3E/HBM4 stacks) — Samsung's pricing power here is corroborating evidence, alongside TSMC's CoWoS sellout, that the AI-infra supply chain's binding constraint has moved from "capital" to "physical component availability" across multiple nodes at once (packaging + memory + power).

Source: Samsung Q2 2026 Guidance Signals Record Quarter — InsiderFinance, 2026-07-07. Full-text fetched.