My 3 Favorite Artificial Intelligence Stocks to Buy Right Now

Three AI investment themes are highlighted: enterprise (Palantir), applied (SoundHound AI), and big-swing/full-stack (Tesla). Palantir’s AIP, integrated into Foundry and Gotham, has driven rapid new client wins including TWG Global, the U.S. Navy (shipbuilding and supply-chain work) and France’s domestic intelligence agency, and the article notes the stock is up >960% since Jan 1, 2024. SoundHound reported Q3 revenue of $42.05 million (+68% YoY), has $269 million in cash and no debt but is not yet profitable, and its stock is up 141% in 2025. Tesla’s AI efforts (widespread Robotaxi availability in the U.S. and Canada and long-term Optimus robot ambitions tied to Elon Musk’s incentive plan) are framed as sizable optionality despite the stock being up 14% in 2025.

Analysis

Market structure: The article reinforces a bifurcation—winners are AI platform owners (PLTR), voice/application layer players (SOUNW), and AI infrastructure providers (NVDA/MSI exposure); losers are legacy analytics consultancies and smaller pure-play consumer ad names that can't monetize AI. Rapid contract wins (PLTR up ~960% YTD) suggest pricing power for data orchestration and higher demand for GPUs, tightening supply of high-end compute and lifting related commodity demand (copper for EVs/robotics) and potentially pushing 10y yields +10–50bp as risk appetite grows.

Risk assessment: Key tail risks are regulatory (data/privacy and export controls on GPUs), operational (Tesla FSD/robotaxi safety incidents), and concentration (single-vendor GPU dependency). Immediate (days) risk is momentum reversal; short-term (1–3 months) risks center on earnings/contract disclosures; long-term (12–36 months) risks hinge on gov budget cycles, regulatory rulings, and Tesla execution on robotaxi/Optimus.

Trade implications: Favor concentrated but hedged exposure to PLTR (enterprise AI) and NVDA (infra), use SOUNW as a high-beta applied-AI punt sized small. Implement pair trades (long PLTR, short legacy consulting or low-monetization consumer tech) and use options to cap downside—e.g., 3–12 month put hedges and 12–24 month LEAP calls to capture structural optionality.

Contrarian angles: Consensus underestimates regulatory drag and execution timelines—robotaxi/Optimus are multi-year optionality, not 2026 cash flows; PLTR’s rapid rerating may already price in broad commercial rollouts. A single-point GPU supply shock or export control would reprice semis/AI names sharply; look for early signs (export policy language, large-scale contract cancellations) before levering exposure.

Key Decisions for Investors

• Establish a 2–3% long position in PLTR over the next 2–6 weeks for a 6–12 month horizon, financed by selling 1% notional 3-month 10% OTM puts as income; if PLTR prints a quarter with <25% YoY commercial revenue growth, trim to 1% or hedge with 3-month 8% OTM puts.
• Increase AI infra exposure with a 1.5–2% long position in NVDA (or NVDA call spread 9–12 month ATM+25% strikes) to capture GPU tightness; cut exposure if US export-control language appears or NVDA implied vol >50% for more than 5 trading days.
• Size a tactical pair: long 1.0–1.5% SOUNW vs short 1.0–1.5% SNAP (equal notional) for 6–9 months to express voice-AI monetization vs ad-monetization risk; unwind if SOUNW revenue growth <40% YoY or SNAP shows margin-accretive ad recovery.
• Buy TSLA 18-month LEAP calls (notional ~1% portfolio) to capture robotaxi/Optimus optionality while selling 90-day covered calls on existing TSLA holdings to finance premium; stop-loss: reduce LEAPs by 50% if regulatory-related negative headlines increase implied vol >60%.