Navitas Semiconductor Is Flashy. This Boring AI Stock Might Make You More Money.

Navitas surged ~376% in 2023 to >$17 but traded near $9 as of Mar 19 (up 23% YTD, +250% 12M); median analyst price target is $8, implying ~9% downside. The Nvidia partnership to supply GaN/SiC for AI data centers (deployment from 2027) and a strategic pivot to data-center, EV and industrial power markets underpin long-term upside, but near-term revenue is expected to decline and profitability remains inconsistent. IBM grew revenue 8% and adjusted earnings 12% in 2025, lifted gross margin by 170bps to 59.5%, expects ~5% revenue growth and +$1B free cash flow this fiscal year, acquired Confluent, signed a watsonX deal with Nvidia, pays a 2.67% yield and has a $340 median price target (~36% upside).

Analysis

Navitas is a classic binary growth-reset: structural upside if GaN/SiC scale and qualification into hyperscaler stacks proceeds on schedule, but the next 18–30 months are execution-heavy and cash-flow negative catalysts. Customer concentration risk (single large design win) amplifies downside if qualification slips or Nvidia backfills with alternative suppliers or in-house designs; any yield or thermal surprises in GaN/SiC will compress gross margins quickly given current scale. IBM is the defensive complement: high free-cash conversion and enterprise data assets make it a vector to monetize AI without binary product qualification risk, and its data-streaming + hybrid-cloud footprint creates a low-beta conduit for AI revenue capture. Second-order winners from an NVTS successful ramp are test-and-pack vendors, substrate suppliers and specialized fabs — conversely, incumbents in silicon power conversion and commodity power-IC suppliers face margin erosion over several product cycles.

Key Decisions for Investors

• NVTS directional, event-weighted: Buy NVTS Jan-2028 $12–$20 call spread (long Jan-2028 $12 call, short Jan-2028 $20 call) sized at 0.5–1% of NAV to target asymmetric upside into 2027–2028 product ramp. Hedge with a 25–35% OTM Jan-2026 put to protect near-term binary downside. R/R: capped upside (~2–4x premium) vs loss limited to premium paid; exit or roll if sample-qualification news misses market windows in 2025.
• Pair trade — long IBM / short NVTS (dollar-neutral): Size long IBM to 1–2% NAV and short NVTS equal notional to harvest volatility and quality spread over 6–12 months. Target relative outperformance of 20–30% if NVTS re-rates down on execution or IBM delivers steady FCF beats; stop-loss if IBM underperforms core software peers by >8% on execution shock.
• Tactical options hedge around NVDA supply dynamics: Buy NVDA 9–12 month 5–7% OTM puts financed by selling the same-duration 15–20% OTM calls (skew capture) sized to cover NVTS exposure. This protects the fund from systemic AI-power demand shocks that would impair both NVTS adoption and NVDA sell-through. Expected cost: small net premium; protects portfolio delta to AI hardware demand shocks.
• Event-driven watchlist: Monitor fab yield, first-silicon performance reports, and hyperscaler qualification timelines (quarterly). If NVTS misses two consecutive milestones, convert short leg from the pair trade into a more aggressive short (increase by 50%) or buy-to-cover IBM if IBM shows cross-sell fatigue — re-evaluate at each milestone with strict time-based stop (6–9 months).