Autoresearch: Samsung Foundry Q2 2026 — Profit Path, 2nm Yield Divergence, Qualcomm Risk
Samsung DS division Q1 2026 profit surged 48x (HBM-driven); foundry specifically targets full 2nm utilization Q2 2026 and 2T won profit FY2026. Critical: TrendForce Apr 14 flags SF2 first-gen yields still ~55% (below mass-production threshold), Qualcomm may revert to TSMC. SF2P second-gen at 70% is the relevant node for Apple/Meta deals.
Autoresearch: Samsung Foundry Q2 2026 — Profit Path, 2nm Yield Divergence, Qualcomm Risk
Generated by
/autoresearchon 2026-05-20. Synthesized across 2 rounds from web search snippets (WebFetch universally 403; early exit after round 2). No Grokipedia anchor (HTTP 403). Treat as raw material — review before promoting. Context: vault/projects/stock-market
Summary
Samsung Electronics' Device Solutions division posted extraordinary Q1 2026 results (operating profit surging 48x YoY to ~53.7T won) driven overwhelmingly by HBM/DRAM demand, not foundry. The foundry unit is targeting full utilization of advanced-node lines in Q2 2026 and an internal profit goal of 2 trillion won for FY2026. A critical new yield data point from TrendForce (April 14): Samsung's first-generation SF2 yields are reported at ~55% — still below mass-production threshold — which may cause Qualcomm to revert to TSMC rather than committing to Samsung 2nm. The SF2P (second-generation 2nm, used for Apple/Meta deals) remains at 70% (January 2026 data), creating a meaningful intra-node divergence. The Samsung vs. Intel vs. TSMC story is now three-tier: Intel on 18A-P for Apple M7, Samsung SF2P for non-Apple hyperscalers, TSMC N2 as the dominant fallback.
Findings
Samsung Q1 2026 Overall — HBM Drives Surge, Not Foundry
Samsung Electronics Q1 2026 consolidated results (reported April 30, 2026):
- Total revenue: 133.9 trillion KRW (record)
- Operating profit: 57.2 trillion KRW (beat consensus of 55.3T won)
- Device Solutions (semiconductor) division: 53.7T won operating profit — vs. ~1T won in Q1 2025 (+48x YoY, 4,800% increase)
- Q1 2026 semiconductor profit > Samsung's entire FY2025 semiconductor profit
Driver: HBM (high-bandwidth memory) for AI data centers, not foundry revenue. Memory chip prices recovered sharply while foundry remains in ramp mode. Per Samsung Global Newsroom Q1 2026 and Data Center Dynamics Q1 analysis.
Foundry specific Q1 2026: Samsung guided that the foundry business would "reach full utilization of advanced-node production lines in Q2 2026" while targeting "earnings improvement" — suggesting Q1 foundry contribution was still modest within the overall DS blowout. The 53.7T won figure is primarily DRAM/HBM. Per TradingKey Q1 analysis.
Samsung Foundry FY2026 Profit Target — 2 Trillion Won
Samsung Foundry's internal 2026 operating profit goal: 2 trillion KRW — its first meaningful profit target since foundry independence. Key components:
- Tesla AI5 and AI6 autonomous driving chips: confirmed foundry win; distinct from the Tesla Austin Intel 14A R&D program
- H2 2026: Mass production of second-generation 2nm (SF2P) for mobile products; 4nm memory products; LPU (large processing unit) chips for AI/HPC customers
- Taylor, Texas fab: Targeting first production in 2026; Samsung claims Taylor is 90% production-ready (April 2026)
Per TechNetBooks Samsung Foundry 2026 profit goal.
2nm Yield Divergence — SF2 vs SF2P: Critical Distinction
Two different Samsung 2nm nodes have materially different yields as of May 2026:
SF2P (second-generation 2nm):
- 70% yield confirmed as of January 2026
- Represents 4 years of GAA transistor telemetry advantage from 3nm experience
- This is the node used for Apple evaluation visits and high-profile customer deals
- Competitive with TSMC N2 (which entered HVM Q4 2025 with "good yield"; TSMC claims 70–80%)
SF2 (first-generation 2nm):
- ~55% yield as of April 14, 2026 (TrendForce report)
- Below the ~70% threshold typically considered mass-production ready
- Qualcomm risk: TrendForce (Apr 14) reports Samsung 2nm yields at ~55%, flagging that Qualcomm "may opt for TSMC" rather than Samsung for upcoming Snapdragon chips
- SF2 is the lower-cost, higher-volume baseline; SF2P is the optimized variant
Per TrendForce Apr 14 SF2 yields and FinancialContent SF2P 70% milestone.
Implication: The wiki's "70% yield (SF2P, January 2026)" figure is correct but represents the premium variant. The mass-market SF2 node is still below threshold. Qualcomm committing to Samsung 2nm would be a significant catalyst that appears at risk.
Samsung as Third Alternative — Updated Position
As of May 2026, the competitive framing has evolved:
- Apple M7: Intel 18A-P (preliminary deal confirmed; production testing underway)
- Apple long-term: Samsung remains an exploratory second source; Digitimes May 12 flagged Intel deal puts Samsung Apple aspirations "under pressure"
- Qualcomm: In final negotiations with Samsung 2nm; at risk due to SF2 yield issues (may revert to TSMC)
- Tesla AI6: Confirmed Samsung Taylor win (~$17B automotive AI chip deal)
- OpenAI HBM4: Samsung supply deal confirmed
The bilateral TSMC-vs-Intel framing overstates the case; Samsung is a real third competitor, but within Apple specifically, Samsung holds a weaker position than Intel.
Contradictions and open questions
- SF2 vs SF2P yield confusion in market reporting: The 70% figure (SF2P) is frequently conflated with Samsung's 2nm overall. TrendForce April 14 specifically flagged SF2 at ~55%. Watchers should distinguish which node is being cited.
- Qualcomm decision timing: Qualcomm CEO's Korea visit (April 2026) and "final negotiations" framing — if Qualcomm chooses TSMC, Samsung loses a major 2nm anchor customer and the 2T won profit target becomes harder to hit.
- Samsung foundry profitability isolation: The 48x DS operating profit figure is largely HBM; foundry's standalone P&L is not publicly broken out. The 2T won target is an internal goal, not a segment-reported number.
Provenance
Rounds run: 2 of 3 (early exit — round 2 resolved the SF2/SF2P distinction; no material new threads for round 3)
Sub-questions by round:
Round 1 (broad survey):
- Samsung foundry Q2 2026 profit path, 2nm SF2 yield, major customer wins
- Samsung Electronics Q1 2026 overall earnings — foundry segment performance
Round 2 (drill-down):
- SF2 vs SF2P yield distinction — which yield figure applies to which customer pipeline
Anchor source: no Grokipedia anchor (HTTP 403 in this environment)
URLs fetched: 0 successful (all 403); all synthesis from search result snippets
Search queries:
- "Samsung foundry Q2 2026 profit breakeven 2nm SF2 yield customer wins May 2026"
- "Samsung foundry Q1 2026 earnings operating profit loss April 2026"
Tools used: WebSearch (2 queries), WebFetch (attempted, all 403). Generated: 2026-05-20