2 AI Stocks That Didn't Get the Memo That the Bull Market Hit a Speed Bump

Palantir has seen accelerating revenue growth over the past 10 quarters, including a 70% increase last quarter driven by its AIP product and strong U.S. government (largest customer) demand amid the Iran conflict; the stock trades at an expensive forward P/S of 51x. Sandisk, spun out of Western Digital, is benefiting from surging NAND prices due to supply constraints and strong SSD/HBM demand and trades at a forward P/E of ~8x on fiscal 2027 estimates but remains cyclical/speculative. The author is cautious on chasing Palantir at current valuations; Motley Fool's Stock Advisor did not include Palantir in its top-10 picks.

Analysis

The market is pricing asymmetric optionality into two tech names that sit on opposite sides of the AI value chain: one as a data/ontology layer that can increase model utility and stickiness, the other as a commodity-adjacent supplier exposed to a capacity/cost inflection. That bifurcation creates distinct margin and capital-cycle dynamics — software firms can reprice on outcomes and enjoy high gross margins, while memory suppliers' returns will be driven by capex cadence and time-to-market for new node/packaging tech.

Second-order winners include systems integrators, cloud storage/software partners and testing/validation vendors who capture implementation dollars and recurring professional services; conversely, hyperscalers that choose to internalize data plumbing or bundle it with models are the latent threat. On the hardware side, a shift toward AI-specific memory (HBM/HBF) creates an interstitial market that benefits specialized fabs and materials suppliers (substrates, controllers) while compressing economics for general-purpose NAND if adoption bifurcates.

Key risks are asymmetric and timeframe-dependent: in the near term, quarter-to-quarter inventory swings and budget timing can reverse revenue momentum for suppliers within 1–3 quarters; medium-term (6–24 months) risks include competitive bundling by cloud providers, open-source model commoditization of software layers, or rapid capex reallocation by memory producers that collapses price realizations. Investors should size positions to these horizon-specific catalysts and prefer structures that limit downside while retaining convex upside to adoption surprises.

Key Decisions for Investors

• Long SNDK via 12–18 month LEAPS (buy Jan-2028 calls, delta ~0.45–0.60) sized as 2–3% portfolio exposure. Rationale: capture upside if AI-driven flash demand and HBF adoption outpace supply rebuild; risk limited to premium, target 2–4x payoff if tightness persists and product mix shifts favor higher ASPs within 12–24 months.
• Relative-value pair: long SNDK / short WDC (equal dollar exposure) for 6–12 months to isolate pure NAND tightness vs broader disk/flash exposure. Expect asymmetric return if NAND rallies further; target relative outperformance of 20–40% while capping macro beta.
• Hedge a directional PLTR equity exposure with a calendar or vertical call spread (buy 9–12 month call, sell shorter-dated nearer-term call) to capture long-term adoption optionality while funding some premium via near-term time decay. Keep size small (<=1% AUM) and treat as tactical core growth exposure — scenario: 2x return if multiple expands on material enterprise wins, max loss limited to debit paid.
• Event hedge: buy 8–12 week puts on PLTR (~3–5% notional) ahead of major government budget/contracting dates or quarterly results to protect against downside from contract timing/renewal misses. This preserves upside while capping drawdown risk over the near-term catalyst window.