Zacks Investment Ideas feature highlights Ark Invest's, SK Telecom, NVIDIA, Nebius and Sandisk

OpenAI raised $122 billion in committed capital, driving a post-money valuation of $852 billion and reported ~$2 billion in revenue per month (enterprise = 40% of revenue, targeting 50% by year-end), with an ads pilot at a $100 million annualized run rate. Anthropic’s valuation jumped to $600 billion with a $14 billion annualized revenue run rate (14x YoY growth) and is expected to pursue an IPO in Q4 2026; SK Telecom provides public investors pre-IPO exposure while ARKK holds ~3% exposure to OpenAI. NVIDIA shows its lowest forward P/E in seven years, and AI infrastructure names Nebius and Sandisk are forecast to see triple-digit revenue growth in 2026, suggesting broad positive implications for AI-related equities.

Analysis

The private AI funding surge is reshaping public market multiple and procurement dynamics without changing the basic compute-to-revenue math: vendors that control high-margin inference stacks and management of enterprise data will extract disproportionate economics, while commodity compute and storage providers face margin pressure as buyers push for integrated solutions. Expect enterprise contracting to shift from capex-heavy GPU purchases toward multi-year service agreements (consumption + software), lengthening sales cycles but increasing revenue visibility for platform players that can offer full-stack SLAs.

On the supply side, sustained demand for high-density accelerators and HBM will buoy chip and substrate suppliers, but it also tightens the bottleneck on advanced packaging and lead times at foundries — a two-speed market where incumbents with secured supply contracts win near-term, and nimble software/architecture optimizers win medium-term by reducing compute per inference. That bifurcation creates an arbitrage: hardware vendors with fixed-asset leverage are levered to capex cycles, whereas software-first firms are levered to ARPU and deal cadence.

Key risks are asymmetric: near-term headline-driven reratings and private-to-public arbitrage around IPOs can spike prices in days, while a 9–18 month horizon threat is meaningful model efficiency gains or regulatory limits on high-capacity models that reduce marginal compute demand. From a positioning standpoint, prioritize optionality and sell limited upside to fund longer-dated convex exposure rather than taking naked beta into a potentially mean-reverting sentiment wave.

Key Decisions for Investors

• Long NVDA Dec-2026 1.5x LEAP call spread (buy deeper ITM, sell higher strike) sized 2-3% NAV — R:R ~4:1 to capture continued data-center tightness while financing via call sell; delta-hedge quarterly and trim 50% on 60% realized gain.
• Buy SKM equity (1.5% NAV starter) ahead of anticipated Anthropic-related corporate catalysts over 6–12 months; pair with 2% NAV purchase of 9–12 month puts to cap downside to ~15% — asymmetric play on re-rating from private-market exposure.
• Long NBIS 9–12 month call spread (small allocation 0.75% NAV) to capture pick-and-shovel growth without funding full vega — exit into any >50% rally or on signs of customer churn; stop-loss if quarterly bookings miss by >20%.
• Short small-cap/storage cyclicals vs long SNDK (pair trade) over 3–9 months: short a basket of high-inventory legacy vendors (size 1.5% NAV) and go long SNDK (1.5% NAV) funded by short proceeds — R:R targets 2:1 capturing industry consolidation and premium for AI-optimized flash suppliers; unwind if sector inventories rise >30% YoY.