Andfjord Salmon Q2 2026 slides: first material revenue in a year

Andfjord Salmon returned to revenue generation, reporting Q2 2026 customer sales of NOK 51.3M (vs effectively zero in the prior four quarters) after its first commercial post-smolt sale of 450,000 fish to Eidsfjord Sjøfarm. However, losses persist: operating EBITDA (ex fair value) improved to -NOK 17.5M (from -NOK 29.7M) and the company recorded a net loss of NOK 20.2M, alongside weak liquidity (current ratio 0.28) and cash burn (levered FCF of -$193M LTM). The company is expanding capacity at Kvalnes to 11,000 tonnes in Phase 1 (up from 6,000 tonnes) with financing support via a NOK 385M private placement and increased undrawn bank facilities to NOK 200M, while management targets meaningful financial improvement from Q4 2026 onward as first harvests approach.

Analysis

NVDA’s print is less about a one-day beat and more about confirming that AI infrastructure is still supply-constrained, not demand-constrained. That matters because the next leg of upside accrues to the highest-beta enablers with pricing power and long lead times — TSM, AVGO, ANET, MRVL, and select power/cooling names — while lower-quality “AI exposure” trades can lag if investors rotate toward the names with real order visibility. Over the next 1-3 months, the market should keep widening the winner set, but the relative winner is still NVDA if gross margin and backlog stay intact.

The contrarian risk is that the consensus may already own the “AI demand is strong” conclusion, so the stock needs guide-up or supply normalization to extend materially from here. What would reverse the move is not a single weak quarter, but evidence that hyperscaler capex growth slows, custom ASIC adoption accelerates faster than expected, or channel inventory builds in the ecosystem. Over 6-18 months, the bigger threat is margin compression from competition and customer concentration, but that is a later-cycle issue unless management signals pricing discipline is slipping.

The aquaculture piece is a different setup: operational improvement does not fix balance-sheet fragility. For capital-intensive growers, equity value can become a financing instrument until cash generation is self-funding, so any delay in harvest timing or softer salmon pricing can quickly overwhelm the operating progress and force dilution. That argues for skepticism on any small-cap “turnaround” rerate until the market sees repeated free-cash-flow conversion, not just biological success.

Key Decisions for Investors

• Buy NVDA on any post-print pullback over the next 1-2 sessions; use a 4-8 week horizon and trim into strength if the move becomes purely multiple expansion without estimate revisions.
• Pair trade: long NVDA / short AMD for 1-3 months, betting that best-in-class execution and supply priority keep relative performance wide; thesis fails if AMD wins incremental design-ins faster than expected or NVDA guide disappoints.
• Add exposure to AI infrastructure beneficiaries on weakness via TSM or ANET rather than chasing the broad semiconductor ETF; better risk/reward if the market keeps rewarding scarcity and order visibility.
• Avoid long exposure to capital-intensive aquaculture turnarounds until there is evidence of self-funded growth; if forcing a trade, treat it as a financing-risk short on any rally, with the thesis invalidated by two consecutive quarters of positive operating cash flow.