This AI Stock Could Soar 47% in 2026, According to 1 Wall Street Analyst (Hint: It's Not Nvidia)

Evercore ISI lifted its ON Semiconductor price target from $68 to $80 (implying ~47% upside from the Dec. 18, 2025 close), highlighting new vGaN power semiconductors and the Aura/Vcore acquisition as potential 2026 catalysts. The company reported Q3 revenue down 12% year-over-year and diluted EPS down 32%, management targets revenue CAGR of 10–12% to 2027 against a stated TAM of $44 billion growing ~18% CAGR, and shares trade at a forward P/E of 18.5. Market consensus is far more muted (average analyst upside ~9%; 15 of 34 analysts rate it buy/strong buy), and the author views Evercore’s 47% thesis as overly optimistic given mixed fundamentals and slower EV adoption.

Analysis

Market structure: Onsemi (ON) stands to win if vGaN adoption by hyperscalers/data centers occurs—50% energy-loss reduction implies value capture and pricing power in power semis versus legacy silicon suppliers. Direct beneficiaries: ON, data-center operators (lower TCO), GaN materials/supply chain; losers: legacy MOSFET/IGBT suppliers and EV-dependent chipmakers if EV demand stalls. Short-term supply tightness for qualified vGaN could support margins; implied-vol in ON options should stay elevated around product milestones.

Risk assessment: Tail risks include failed vGaN qualification, Aura integration delays, or regulatory-driven EV demand shocks (California/EPA/IRA changes) that reduce auto TAM by >5–10% over 12 months. Timing: days–weeks for earnings/analyst revisions, 3–9 months for Aura integration signals, 12–24 months for broad vGaN design wins to move revenue materially. Hidden dependency: top-customer concentration and hyperscaler in-house sourcing; catalysts are design wins, hyperscaler qualification, and 2026 guide updates.

Trade implications: Tactical plays favor option-based long exposure to ON with defined risk—LEAPs or 12–18 month call spreads—to capture H2 2026 vGaN adoption while capping premium. Pair opportunities: long ON vs short EV-exposed chipmakers (Infineon/STM) if regulatory signals damp EV growth. Reallocate 2–4% from broad auto suppliers into AI/data-center beneficiaries (ON, NVDA) while keeping portfolio-level stops.

Contrarian angles: Consensus underestimates execution/qualification runway—SiC history (multi-year design cycles) suggests a slow, binary upside if ON nails design wins; conversely upside may be underpriced (current forward P/E ~18.5) if ON captures even 3–5% incremental TAM by 2027. Watch for unintended consequences: hyperscalers internalizing GaN or rapid commoditization that would compress ASPs; absence of third-party qualification by Q3–Q4 2026 would invalidate bullish thesis.

Key Decisions for Investors

• Establish a tactical 2–3% portfolio allocation long ON via Jan 2027 LEAP calls (buy-dated 12–18 months) or a 1.5–2% cash stake if options unavailable; target +40–50% upside (Evercore $80) within 12–18 months, exit or cut to zero if company guidance misses consensus by >5% or shares fall 25% from entry.
• Conservative income-financed approach: buy ON shares sized 1–3% and sell 3-month covered calls ~20% OTM to fund position; concurrently buy a 6–12 month 15% OTM put as downside protection ahead of next four quarterly reports.
• Relative-value pair: go long ON (1–1.5%) vs short Infineon (IFNNY) or STMicro (STM) (1%) to express differential exposure to data-center vGaN upside vs EV/auto cyclicality; increase short leg if regulatory/IRA rollbacks confirmed within 90 days.
• Hard triggers to act: within next 90–180 days, increase exposure to ON if (a) published hyperscaler design wins or vGaN qualification announcements occur, or (b) management raises 2026 revenue guide >5%. Reduce/close positions by 50% if no material design-win announcements by Sept 30, 2026 or if California/EPA rules rollback reduces projected EV adoption assumptions by >5–10%.