Sanmina (SANM) Q2 2026 Earnings Transcript

Sanmina delivered a major Q2 beat, with revenue of $4.01 billion up 102% year over year, non-GAAP operating margin at 6.4% (+80 bps), and non-GAAP EPS of $3.16 (+125%). Growth was driven by ZT Systems accelerated-compute shipments, with ZT revenue at $1.88 billion and management raising confidence in FY2027 to "$16 billion plus" revenue. The company also generated $399 million of operating cash flow, repurchased $160 million of stock, and authorized an additional $600 million buyback, though component shortages remain a supply-chain headwind.

Analysis

SANM is turning into a cleaner AI supply-chain beneficiary than the market likely modeled at deal close. The key second-order effect is that this is no longer just an EMS re-rating story; the acquired platform is expanding the company’s addressable mix toward vertically integrated rack/system assembly, which should lift wallet share with hyperscalers even if unit growth normalizes. The more important signal is management’s confidence that capacity, not demand, is the binding constraint — that usually leads to a multi-quarter backlog conversion tailwind rather than a one-quarter beat.

The near-term setup is asymmetric because the quarter pulled demand forward, but the call also made clear that the real revenue inflection is later in the year and into FY27 as next-gen platforms ramp. That means the stock can de-rate on any perceived Q4 “air pocket” even while the underlying order book improves, creating a window for volatility around guidance resets rather than fundamental deterioration. The key question is whether investor focus shifts from headline revenue to revenue quality; if a larger share becomes consignment-heavy, top-line could look less impressive while economics remain intact.

The underappreciated risk is supply-chain friction. Memory and custom ASIC shortages are not just a timing issue — they can compress customer schedules, delay design wins into revenue, and force SANM to carry more working capital exactly when it is trying to scale. If shortages ease, the business can outrun guidance; if they persist into 2027, the market may be overestimating how quickly the ZT thesis translates into durable growth. AMD is the cleanest read-through beneficiary in the near term, while NVDA remains oddly less directly exposed given management’s explicit comment that the current shipments were AMD-based.

Consensus appears to be underestimating buyback support and overestimating balance-sheet risk. With leverage still low and a fresh authorization in place, any post-earnings pullback is likely to be absorbed faster than a classic hardware cycle name, especially if cash generation stays strong while capex ramps modestly. The more contrarian view is that SANM may be evolving into a higher-quality AI infrastructure compounder, but the multiple expansion will likely lag until the market sees a second consecutive quarter of next-gen program execution rather than just pull-forward shipments.

Key Decisions for Investors

• Go long SANM on any 5-8% post-earnings pullback over the next 1-3 weeks; risk/reward is attractive because near-term volatility is likely driven by timing noise, while FY27 optionality remains underappreciated.
• Pair trade: long SANM / short a basket of lower-quality EMS or low-growth industrial hardware names for the next 3-6 months, betting that AI-linked mix shift and buybacks drive relative multiple expansion.
• Add a tactical long AMD vs. short NVDA basket for 1-2 quarters; current ZT shipment mix indicates nearer-term direct exposure to AMD compute ramps, while NVDA read-through is weaker here.
• Sell out-of-the-money SANM puts into any volatility spike, targeting 2-4 month tenor; implied downside should be cushioned by cash flow, low leverage, and repurchase authorization.
• If SANM breaks below its post-earnings support on no negative fundamental news, reduce exposure rather than add — the main risk is not demand collapse but a temporary revenue air pocket before next-gen schedules convert.