2 Artificial Intelligence Stocks That Could Soar in the Next Bull Market

Symbotic and Applied Digital are presented as high‑growth AI infrastructure plays: Symbotic reported full‑year 2025 revenue up 26% to nearly $2.3 billion, a contracted backlog of roughly $22.5 billion (≈10x annual sales), a Walmart stake of ~15% and a turnaround to about $788 million in free cash flow from negative $102.45 million the prior year, while expanding recurring revenue via its GreenBox JV with SoftBank and acquiring Walmart’s Advanced Systems and Robotics unit. Applied Digital operates as an "AI landlord" and GPU‑as‑a‑service provider with roughly $16 billion of contracted revenue over 15 years (≈$11 billion from CoreWeave), an 84% quarterly top‑line increase to $64.2 million, a $5 billion Macquarie financing facility, a 4 GW development pipeline and shortened build times (12–14 months), though it remains net‑lossing from upfront construction and depreciation. Together the firms highlight structural demand for automation and purpose‑built AI data‑center capacity, supporting long‑run revenue visibility but with execution and capital‑intensity risks that investors should weigh.

Analysis

Market structure: Symbotic (SYM) and Applied Digital (APLD) are direct beneficiaries — SYM via a $22.5B backlog (~10x FY sales) locking multi-year revenue; APLD via ~$16B contracted revenue (CoreWeave ~$11B) locking long-term demand for purpose-built AI facilities. Losers are labor-intensive 3PLs and generalist colos without liquid cooling or cheap power; pricing power shifts toward specialists who can deliver lower total cost of ownership for AI and automated warehousing. Commodities impact: sustained buildout lifts copper, transformers and power demand (power prices could be a margin swing), and heavy capex will pressure corporate bond issuance in the sector.

Risk assessment: Key tail risks are GPU export controls or supply shocks (weeks-months impact), utility interconnection or permitting delays (3–18 months), and client-concentration blows (CoreWeave/Walmart exposures). Immediate risk: earnings or contract updates in next 30 days; short-term (3–12 months): construction cadence and FCF conversion; long-term (2–5 years): backlog conversion and recurring service revenue. Hidden dependencies include GPU availability, local grid capacity and financing tranche covenants (Macquarie facility triggers), any of which can delay revenue recognition and increase depreciation.

Trade implications: Favor selective, size-constrained long positions in SYM and APLD to capture secular AI/automation demand but hedge operational concentration. Use pair trades: long APLD vs short generalist data-center REITs (e.g., EQIX) to capture AI-specific premium compression risk. Options: use LEAP call exposure to play long-term capacity builds and short-dated puts to define risk; expect volatility around GPU-policy headlines and large customer milestones.

Contrarian angles: Consensus understates concentration risk — CoreWeave and Walmart stakes create single-client dependencies that can swing earnings sharply; conversely, SYM’s jump to ~$788M FCF suggests upside underappreciated by the market if FCF sustains. Historical parallel: 2016–19 hyperscaler land grabs where early infrastructure owners locked economics; unintended consequence: rapid buildouts could create localized power scarcity and political pushback (rate increases or permitting moratoria).

Key Decisions for Investors

• Establish a 2–3% long position in SYM (equity exposure) immediately; plan to add up to 5% total if trailing-12-month FCF stays >$500M or if quarterly revenue growth exceeds 15% YoY for two consecutive quarters. Hedge initial position with 6‑month puts 20% OTM sized at 25% of position cost.
• Initiate a 1.5–2% exposure to APLD via 12–18 month LEAP calls (near‑ATM) to capture capacity buildout optionality; scale to 4% if APLD brings ≥1 GW of revenue-generating capacity online or recognizes >$1B of contracted revenue into billings within 12 months. Cap downside by buying 9–12 month 25% OTM protective puts at 30% notional.
• Run a 6–12 month pair-trade: long APLD vs short EQUALLY‑sized EQIX (or another generalist data-center REIT) sized 1:1 notional to exploit AI-specific performance differential; take profits if spread widens >15% or compresses to 0. Monitor utility interconnection timelines weekly and trim the short if APLD reports material permit delays.
• Set active risk triggers to act within 90 days: if US/ally GPU export controls are tightened (public guidance change or new rule), immediately reduce APLD exposure by 50% and increase protection on SYM; if either company misses revenue conversion cadence by >20% vs guidance, cut position to half and reassess backlog realization assumptions.