2 Ridiculously Cheap AI Stocks That Could Turn $1,000 Into at Least $5,000 by 2028

Micron reported a fiscal Q2 2026 blowout on March 18 (revenue roughly tripled QoQ; EPS +682% YoY) but its stock is down ~30% since the print; it trades at ~8.2x sales and a forward P/E of ~7.6, with a modeled $98.91 FY EPS implying a potential $2,275 share price (~6x current). CoreWeave is down ~39% over six months, trading at ~7x sales, with a $66.8bn revenue backlog (+342% YoY) and FY revenue of $5.1bn (+168% YoY); capex is forecast to jump from $14.9bn (2025) to $30–35bn (2026) as it targets 3,100MW capacity by end-2027 (vs 850MW). Implication: both firms are well positioned to benefit from AI-driven HBM and data-center demand, offering substantial upside in the author’s view, but carry significant execution and capex risks — likely to move the individual stocks rather than markets broadly.

Analysis

Micron and CoreWeave are benefiting from concentration effects: memory that is wafer‑intensive (HBM) creates non‑linear scarcity because each AI GB pulls multiplex wafer capacity versus commodity DRAM. That amplifies pricing power for incumbents with available capacity and first‑mover process tech, while increasing marginal economics for advanced packaging and test suppliers one or two tiers down the chain. Expect asymmetric profit capture: node/packaging winners see steady margin expansion even if DRAM commodity cycles normalize.

CoreWeave’s business is effectively a long build‑option on external demand for third‑party capacity, but the option’s value is execution‑and‑funding dependent. Large capex ramps compress near‑term FCF and create meaningful dilution or refinancing events if markets reprice; conversely, bringing capacity live ahead of competitors creates multi‑year revenue visibility. Key catalysts are capacity turn‑ons and multi‑year contract commencements; key risks are GPU supply, financing terms, and hyperscaler vertical integration decisions.

Practical trade framing: own the structural supply‑constrained asset (Micron) via time‑spreaded, delta‑positive optionality to capture multi‑year margin reversion while hedging re‑rate risk, and express CoreWeave as a financed growth position (equity + short dated call financing or collars) to monetize high conviction around capacity ramps but limit dilution/tail risk. Pair trades that isolate HBM exposure (long Micron / short broad DRAM index or lagging foundry) de‑levers macro and isolates structural memory tightness. Monitor customer concentration signals and any large financing notices as binary de‑risk events.

Key Decisions for Investors

• Tactical LEAP for Micron (MU): buy MU 24–36 month LEAP calls equal to 3% NAV (delta ~0.35–0.45) to capture multi‑year HBM tightness; size with a hard stop at -35% and a profit‑taking plan to trim half at +100% and the rest at +250% (target horizon 12–36 months).
• CoreWeave equity with downside protection (CRWV): accumulate CRWV common for 2–3% NAV but purchase 9–12 month puts to cap downside at ~40% or implement a collar sold against near‑dated calls to finance puts; target 3x upside in 24–48 months if execution and funding succeed, downside capped by protection.
• Pair trade to isolate HBM/memory scarcity: long MU (as above) paired with short a broad DRAM/commodity exposure (or short a legacy DRAM peer) sized to neutralize semiconductor beta — net position expresses HBM skew while hedging cyclical inventory rebalances (timeframe 6–24 months).
• Event trade for CoreWeave capacity announcements: buy CRWV ahead of announced data center turn‑ons (allocate 1% NAV per event) and sell into live announcements; if capacity brings start missing targets or financing terms worsen, flip to protective puts within 2 weeks.