Earnings call transcript: PowerCell Group sinks on weak Q2 2026 results

PowerCell’s Q2 2026 revenue fell 46% YoY (ex-one-off licensing) and gross margin dropped to ~10% vs a normalized ~25%, pressuring profitability despite improving demand signals. Order intake rose to SEK 86m from SEK 50m in Q1 (+72% sequential), including an SEK ~30m ECL PS190 (5 MW) reference order and 9 MW signed in late June/early July, while management flagged ongoing lumpy revenue timing. The stock fell 8.2% to $18.35 premarket, with management emphasizing cost cuts already underway for H2 2026–2027 and expecting revenue swings to continue, as they build a second growth pillar in power generation for data centers.

Analysis

TSMC’s update reinforces that the AI cycle is becoming a capacity-allocation story, not just a demand story: the spillover alpha is increasingly in the toolchain, advanced packaging, and test/inspection names that monetize every incremental wafer and node transition. That makes semicap exposure more attractive than chasing the largest-cap pure-play foundry at already-fuller valuations; the cleaner risk/reward is in names with operating leverage to capex follow-through. A near-term reversal would be any sign that AI customers are front-loading less than expected or that management starts hinting at digestion rather than acceleration.

PowerCell reads as a very different setup: the equity is being priced like an option on a future market, but the quarter says the cash conversion engine is still too brittle for that premium. The key second-order risk is financing overhang — if revenue remains stepwise while gross margin stays well below normalized levels, the company may be forced to fund growth before the “second pillar” is real, which caps upside even if the story improves. The order-book bounce is directionally helpful, but until repeat orders show up, this is not yet a durable re-rating catalyst.

Contrarian-wise, consensus may be underestimating how quickly TSMC’s capex revision can lift the entire AI supply chain into another leg up over 1-3 months. For PowerCell, the market may be over-interpreting the strategic narrative: the right lens is not the reference win, but whether the company can prove margin normalization and avoid dilution before 2027 cost actions matter. If China policy or data-center demand creates follow-through, that is a 6-18 month thesis — not a fix for the next two quarters.

Key Decisions for Investors

• Long AMAT or KLAC on any 1-2 day post-print dip; TSMC capex guidance should translate into higher tool spend and better earnings revisions over the next 1-3 months. Falsify if TSMC commentary turns to utilization digestion or capex pause.
• Pair trade: long TSM / short INTC for 1-3 months. The thesis is continued share gain in advanced nodes versus a lagging competitor still fighting execution and foundry credibility; risk/reward favors the market leader if AI demand holds.
• Avoid chasing PCELF strength; use any bounce to initiate a small short or buy put spreads into the next 4-8 weeks. The setup is weak revenue leverage, subscale margins, and a non-trivial dilution risk if order conversion stalls. Cover if management shows two consecutive quarters of meaningful margin recovery and repeat orders.
• Relative-value watch: long BE versus PCELF as the cleaner listed proxy for data-center prime power adoption. If the theme is real, better capitalized incumbents should capture the first wave before smaller peers can finance scale.