SANM · Sanmina Corporation

Current Thesis

SANM is a re-rate story wearing a contract-manufacturer's clothes. FQ2 FY2026 (reported 2026-04-27) detonated the legacy frame: revenue $4.01B vs $3.29B consensus (+22.8%), non-GAAP EPS $3.16 vs $2.40 (+31.8%), an +11% after-hours move, a fresh $600M buyback, and a FY2027 revenue target of $16B+ roughly double FY2025's $8.1B. Sell-side is still behind the tape (JPM PT $145, Susquehanna $135, both under a ~$190 print), the analyst-lag this playbook exists to front-run. The catch: the post-earnings leg already ran (RSI 83 in mid-May, off a $74.90 52-week low to a $193.53 high), and the FQ3 guide of $3.2–3.5B sits below the $3.522B consensus on management's own "orders pulled forward" language. Narrative intact; the easy momentum leg is spent until the next print.

Bull Case

• ZT Systems converts a low-margin EMS book into an AI-server manufacturing franchise with AMD as anchor customer and committed strategic partner; company guided the deal accretive to non-GAAP EPS in year one (deal closed 2025-10-27).
• FQ2 FY2026 (2026-04-27): revenue $4.01B (+22.8% vs consensus), non-GAAP EPS $3.16 (+31.8% beat), Q2 free cash flow $342M, non-GAAP pretax ROIC 34.7% vs 23.0% a year prior.
• FY2027 revenue target $16B+ against FY2025's $8.1B, with management "increasingly confident." FY2026 guide lifted to adj EPS $10.75–$11.35 (vs $10.02 est) on sales $13.7–14.3B.
• Cheap on the new earnings base: NTM EV/EBITDA ~11x vs Jabil 14x and Celestica 29x; forward P/E ~18.6. That peer discount is the re-rate fuel if AI revenue proves durable.
• Core Communications Networks (IP switching, optical pluggables) grew 22% YoY independent of ZT the legacy book is inflecting too, not just the acquisition.
• Balance sheet supports a dip: net debt ~$940M, net debt/EBITDA 1.48x (inside the 1.0–2.0x target), plus the open-ended $600M repurchase authorized at the print.
• Sell-side targets sit BELOW the market (mean PT $173.75 vs ~$190). Upgrade catch-up toward the tape is the un-played catalyst.

Bear Case

• AMD concentration: ZT is a single-customer-anchored rack-manufacturing relationship. JPM's Neutral explicitly flags "AMD rack concentration risk" if AMD second-sources NPI/rack manufacturing or its accelerator share slips, the growth engine stalls.
• The FQ3 FY2026 guide is soft: $3.2–3.5B revenue vs $3.522B consensus, a sequential DROP from $4.01B, attributed to "accelerated compute orders pulled forward from H2." A pull-forward spike is not a run-rate, and the next print can disappoint even this lowered bar.
• Structurally thin margins: FY2026 op-margin guide just 6.3–6.6%, and JPM notes Sanmina "has historically lagged peers in gross margin expansion." Doubling revenue at thin margins justifies a very different multiple than the market now pays.
• Stretched: ~$190 sits against the $193.53 52-week high after a >150% run from $74.90; RSI hit 83 in mid-May. Wall Street mean PT $173.75 implies ~10% downside. Buying the 52-week-high candle with no near-term catalyst is the stretched-above-MA trap.
• Cyclicality: ~63% of revenue is tied to data center and cloud, so an AI-capex air-pocket hits SANM disproportionately.

Setup & Price Structure

The structure is post-catalyst consolidation, not a fresh breakout. The 2026-04-27 earnings gap drove the parabolic leg into mid-May (RSI 83); price has since digested near the 52-week-high band ($188–193). With the FQ3 print roughly seven weeks out and nothing inside the next 30 days, fresh upside needs one of two things: a sell-side upgrade cluster (targets are $135–145, far under tape — upgrades would be real fuel) or a pullback that resets the 20-EMA. The constructive re-entry is a higher-low retest of the post-earnings breakout shelf near $160 on cooling RSI, not a chase of the $190 high. A weekly close back inside the pre-earnings range would mark the demand pull-forward as a one-quarter event. This is a MATURING setup: the bulk of the re-rate has happened, and the next leg is gated entirely on the late-July print confirming the $16B-FY27 trajectory.

Catalyst Calendar (next 30 days)

• No company-specific binary inside the next 30 days (through ~2026-07-07).
• ~2026-07-27 (est.): FQ3 FY2026 earnings THE binary. Tests the $3.2–3.5B guide and whether ZT/AMD demand is run-rate versus pull-forward. Just outside the 30-day window but the only catalyst that moves the thesis.
• Ongoing: sell-side re-rating watch JPM $145 / Susquehanna $135 vs ~$190 tape; an upgrade cluster toward market price would confirm narrative acceleration ahead of the print.
• Second-order tells: AMD AI-accelerator (MI-series) demand signals and hyperscaler capex prints feed SANM's order book directly.

What Would Change Our Mind

• FQ3 (late July) revenue below the $3.2B guide floor → confirms pull-forward, not run-rate; thesis breaks.
• Weekly close back below the post-earnings breakout shelf (~$160) / loss of the rising 20-EMA → momentum leg over, MATURING flips toward SATURATED.
• AMD signals second-source rack/NPI manufacturing or loses accelerator share → the concentration bear case activates.
• FY2027 $16B target walked back, or op-margin guide cut below ~6% → the re-rate premium is unearned.
• Bullish reversal: a sell-side upgrade cluster lifting targets through the $190s plus a clean higher-low retest would re-arm an ACCELERATING entry rather than a watch.

Correlation Notes

SANM trades as a high-beta proxy for the AI data-center buildout, tightest to AMD (anchor customer and partner) and to EMS/ODM peers Celestica (CLS), Jabil (JBL), and Flex (FLEX). Celestica's ~29x EV/EBITDA is the bull's re-rate comp; Jabil's ~14x is the base case. A break in AMD's accelerator narrative, a hyperscaler capex pause, or a Celestica/Jabil guide-down drags SANM regardless of its own order book. It also tracks optical/networking names through the Communications segment (IP switching, optical pluggables). Idiosyncratic protection is low ~63% data-center revenue means it moves with the theme, not against it.