Wall Street Analyst Warns of AI Bubble and Urges Investors to buy SaaS Stocks. 5 Stocks to Buy if He Is Right.

Benchmark GP Bill Gurley warned of a potential AI infrastructure bubble but recommended rotating into beaten-down SaaS names; five stocks highlighted are ServiceNow, Salesforce, Workday, UiPath and Adobe. Each name is down roughly 25%-35% YTD (Workday ~35%, others ~25%), with valuations shown: ServiceNow forward P/S 7.5x and P/E 28x; Salesforce P/S <4x and P/E <15x; Workday P/S <3.5x and P/E <13x; UiPath P/S just above 3.5x and P/E 15x; Adobe P/S 4x and P/E <11x. Growth callouts: ServiceNow revenue >20% YoY and new Tower AI product; Salesforce expects >10% annual revenue growth through 2030 and launched Data 360; Workday expects mid-teens growth and reported AI ACV doubled to $100M; UiPath ARR growth is re-accelerating as it builds an agentic orchestration platform.

Analysis

The durable edge in this cycle will be data-positioned orchestration layers — platforms that control telemetry, policy and routing of work to the right compute asset. That creates a feedback loop: orchestration vendors can reduce expensive LLM calls by substituting cached logic, rules engines or cheaper bots for a material slice of customer spend, which compresses incremental demand growth for GPU-heavy infrastructure even as AI feature adoption accelerates across enterprise apps.

A second-order bifurcation of compute demand is emerging: ongoing large-scale model training remains concentrated and hiring-capital intensive, while day-to-day inference economics will polarize into high-frequency small inference (edge, real-time agents) and occasional expensive calls. Winners are those who can monetize orchestration (subscription + consumption) without destroying margin; losers are narrow infra vendors whose TAM depends on per-token consumption growth continuing unabated.

Key risks and timing: expect the first visible re-rating window in 3–12 months as next-quarter ARR disclosures and AI contract dosages show conversion into revenue; a macro slowdown, rapid shift to consumption billing, model-quality regressions or a wave of enterprise reversions to on-prem will flip the trade within quarters. Conversely, a surprise enterprise consolidation (large ERP/CRM migration that bundles orchestration) would re-rate these SaaS incumbents over 12–36 months.

Given this topology, prioritize durable orchestration and monetizable AI features while hedging GPU-convex tail risk — position sizing and option structures should reflect potential binary outcomes from model economics and cloud pricing moves.

Key Decisions for Investors

• Long Adobe (ADBE) equity or buy 12–18 month call calendar spread (size 1–3% NAV). Rationale: highest probability of monetizing creative AI usage via consumption while retaining pricing power; target 25–40% upside in 12 months; set hard stop at -20% or hedge with 20% notional in 12-month puts.
• Long ServiceNow (NOW) 9–15 month call options (or buy shares financed with a modest sell of covered calls). Trade size 1–2% NAV. Rationale: control plane optionality and stickiness make it a takeover/roll-up candidate; reward skew >2x if orchestration monetization accelerates; protect with 10–15% downside stop or buy 9–12 month protective puts.
• Relative value pair: long Workday (WDAY) vs short a small tranche of NVDA convexity using 9–12 month put spreads on NVDA to cap cost. Allocate net 1–2% long WDAY with NVDA hedge sized to cap portfolio drawdown at ~3%. Rationale: WDAY captures HR/finance data moat while NVDA remains exposed to a sentiment-driven allocation; this reduces outright beta to the AI infra narrative.
• Convex punt on UiPath (PATH): buy 12-month out-of-the-money calls (small allocation 0.5–1% NAV). Rationale: asymmetric upside if RPA orchestration meaningfully displaces token-invocations; loss limited to premium, upside captures re-acceleration in ARR growth or M&A interest.