The Pick-and-Shovel Phase of AI Has Arrived: 3 Stocks to Watch

Astera Labs, Iren and Nokia are positioned as ‘pick-and-shovel’ suppliers to the AI buildout and are reporting strong demand-driven results: Astera Labs Q3 revenue rose 104% YoY to $230.6m with Q4 guidance of $245m–$253m versus prior-year Q4 revenue of $141.1m. Iren, which converted GPU-heavy Bitcoin mining infrastructure to hyperscaler capacity, secured a $9.7bn five-year deal with Microsoft and delivered Q3 revenue of $240.3m (+355% YoY) and Q3 net income of $384.6m, with total assets of $4.3bn (cash $1bn) and convertible notes of $964.2m. Nokia reported Q3 sales of €4.8bn (+12% YoY), driven by a 114% jump in optical networking to €782m as it pivots toward edge/6G and AI data-center demand.

Analysis

Market structure: Winners are infrastructure and power providers—Astera Labs (ALAB), Iren (IREN) and Nokia (NOK)—plus hyperscalers that outsource capacity. Direct losers include legacy on‑prem vendors, smaller colo operators without renewable/scale advantages, and segments exposed to GPU supply bottlenecks. Expect multi‑year demand for AI connectivity, optics and green power; pricing power will concentrate with vertically integrated owners of capacity and specialized component suppliers, tightening supply/demand for NICs, optics and grid capacity for at least 2–4 years.

Risk assessment: Key tail risks are an AI spending pullback (model moderation), export controls on GPUs/ASICs, or counterparty concentration (IREN’s material Microsoft exposure). Time horizons: immediate (days) — sentiment/volatility spikes on headlines; short (weeks–months) — earnings vs. guidance; long (quarters–years) — build cycles, 6G standardization and data‑center rollouts. Hidden dependencies include energy prices, copper/optical supply, and foundry capacity; watch semi lead times and contracted backlog as catalysts.

Trade implications: Prefer concentrated long exposure to ALAB and IREN (infrastructure pick-and-shovel) and selective NOK exposure to capture edge/optical demand, using staggered entries over 2–12 weeks to average execution. Use 9–15 month LEAPS or 6–12 month call spreads to capture upside while limiting premium decay; sell covered calls or take profits on >25% rallies. Reduce portfolio duration by ~0.5 year and increase commodity/energy exposure modestly to hedge physical power/copper risk.

Contrarian angles: Consensus underestimates margin compression if hyperscalers internalize stack or component pricing eases as supply scales; conversely, market may underprice IREN’s vertically integrated energy advantage if power costs spike. Historical parallel: cloud infra cycles (2010s) saw pick‑and‑shovel vendors outperform but with multi‑quarter drawdowns. Set explicit deterioration triggers (revenue guide misses, backlog erosion) to avoid structural drawdowns.

Key Decisions for Investors

• Establish a 2–3% long position in ALAB within 2 weeks; scale to 5% if Q4 revenue meets or exceeds the midpoint of guidance (~$249M) or if non‑GAAP gross margin expands >200 bps; exit/trim if next quarter revenue growth falls below 20% YoY or guidance is cut.
• Initiate a 3–4% long position in IREN today and buy 12‑month 20% OTM puts covering 50% notional as downside insurance; increase to 6% if company reports contracted hyperscaler backlog implying >50% YoY revenue growth for FY2026 or if MSFT milestone revenue recognition accelerates.
• Accumulate a 2% position in NOK with a 12–18 month horizon; take profits on a >25% share rally or trim 50% if optical segment YoY growth decelerates below 20% in any quarter.
• Rotate 4–6% of consumer/streaming exposure (example: NFLX-sized positions) into an AI‑infrastructure basket (equal‑weighted ALAB, IREN, NOK) over 4–8 weeks; simultaneously reduce portfolio duration by ~0.5 years and buy 3–6 month ATM puts on the basket if VIX >18 or if any company issues a negative backlog/update.