3 High-Conviction AI Stocks With 10x Potential by 2036

The piece highlights three AI-related stocks: AMD, CoreWeave and Upstart, each positioned to benefit from accelerating AI demand. AMD projects a long-term revenue CAGR of 30% (60% for its data-center AI segment), has an upcoming MI450 accelerator, and trades at a forward P/E of ~53 after a >70% one-year gain. CoreWeave reported revenue up 204% year-over-year to nearly $3.6 billion in the first nine months of 2025 but saw costs rise 263%, interest expense jump to $841 million and a nine-month net loss of $771 million (improved from $857 million); its P/S is just above 7 and the stock sits ~60% below its 52-week high. Upstart posted $685 million revenue in the first nine months of 2025 (+57% YoY) and returned to profitability with $35 million, while its AI-driven underwriting model automates ~91% of decisions and the stock trades at ~5x P/S despite being down significantly from peaks.

Analysis

Market structure: AI demand is bifurcating winners — GPU incumbents (NVDA) retain scale advantages while second-tier silicon (AMD) and AI-specialized clouds (CRWV) capture share where price/performance or service specialization matters. CoreWeave’s 204% y/y revenue jump to ~$3.6B (first 9 months 2025) signals severe GPU-driven capacity tightness and strong pricing power for AI-tailored hosting, but its 263% cost growth and $841M interest expense highlight capital-intensity and margin pressure. Upstart’s AI underwriting (2,500 variables, 91% automation) threatens FICO’s static model in a $1T addressable market, but credit-cycle sensitivity ties its valuation to rates and loss trends.

Risk assessment: Key tail risks are (1) export/regulatory curbs on advanced accelerators (China access) within 3–12 months, (2) execution failure on AMD’s MI450 benchmarks within 6–12 months, and (3) CoreWeave refinancing stress if quarterly cash burn >$1B or interest costs remain elevated. Short-term (days–weeks) moves will be headline-driven (benchmarks, hyperscaler contracts); medium-term (3–12 months) depends on earnings and FCF; long-term (1–3 years) depends on secular share shifts and foundry capacity (TSMC). Hidden dependencies include Nvidia driver/stack lock-in and hyperscaler bilateral GPU deals that can unwind expected share gains.

Trade implications: Tactical longs: AMD for secular GPU share upside if MI450 shows parity (target 6–12 month re-rate), CoreWeave for platform exposure to AI infra growth but hedge refinancing risk, and a small speculative Upstart position tied to improving delinquency metrics. Pair trades: long AMD vs. short NVDA tail-protection to capture relative catch-up; buy LEAP call spreads on CRWV to limit capital at risk while capturing asymmetric upside if GPU pricing stays tight. Entry/exit: enter on event triggers (MI450 benchmarks, CRWV quarterly cash burn < $600M, Upstart 90+ day delinquency down 150bps) and cut 15–25% on missed milestones.

Contrarian angles: Consensus underestimates execution friction — historically CPU/GPU share shifts take multiple product cycles (AMD vs Intel CPU ramp took years), so price in at least 12–18 months of volatility for AMD. CoreWeave’s ~60% off 52-week-high may underprice refinancing and operating-leverage risk; conversely NVDA’s dominance is arguably over-crowded and priced for perfection — a single missed sell-in cycle or hyperscaler deal favoring AMD could produce rapid relative re-rating. Watch for hyperscalers vertically integrating or Nvidia softening pricing as anti-competitive defense, which would compress expected upside for second-tier plays.

Key Decisions for Investors

• Establish a 2–3% portfolio long in AMD (ticker: AMD) sized for medium risk: buy 6–12 month call spreads (buy ATM LEAP, sell 15–25% OTM) targeting a 30–80% upside if MI450 benchmarks and first product shipments meet expectations; stop-loss: trim if MI450 independent benchmarks miss or stock rises/falls 25% without fundamental confirmation.
• Initiate a 1–2% tactical long in CoreWeave (CRWV) via Jan 2027 call spreads (limit premium); hedge 40–50% of notional with OTM puts or buy-to-close on weakening cash flow — reduce/exit if quarterly cash burn > $1B, interest expense stays > $700M, or revenue growth decelerates below 40% y/y.
• Take a small 0.5–1% speculative position in Upstart (UPST) using 3–6 month call spreads, conditional entry only after two consecutive quarters showing net charge-off improvement and 90+ day delinquency falling >150bps; unwind immediately if delinquency trends reverse or revenue growth falls below 15% y/y.
• Run a relative-value pair: long AMD (2%) vs short limited NVDA exposure (1%) via buying AMD calls and funding with selling a 3–6 month NVDA call spread — aim to capture AMD re-rating should MI450 validate while capping downside if NVDA retains outsize leadership; exit if NVDA introduces material price cuts or AMD execution misses.