Why ServiceNow Stock Was Drifting Lower Today

Anthropic introduced a Claude feature that can remotely operate a computer from smartphone prompts, prompting a software-sector sell-off (ServiceNow down ~4.4% intraday; IGV down ~4%). ServiceNow is down nearly 60% from its Dec 2024 peak and still trades at a GAAP P/E of 63, raising valuation concerns despite solid company results. The piece flags rising AI agents as an ongoing competitive threat to enterprise software but notes that continued strong performance could eventually reward the stock.

Analysis

Advances in agentic AI are functionally shifting value away from GUI-driven workflow licensing toward compute, integrations and model orchestration layers; that reallocates margin capture to providers of high-throughput inference and integration tooling rather than legacy SaaS UI vendors. The second-order supply-chain effect is an accelerated capex cycle for datacenter GPU capacity and cloud interconnects, compressing time-to-benefit for companies that already own or control inference stacks.

For incumbents that sell process orchestration, the practical risk is not binary displacement but a steady erosion of incremental seat/license elasticity: each successful agent use case reduces the marginal willingness to pay for formalized ticketing and low-value process steps. That makes near-term KPIs (retention, ARPU expansion) stickier than headlines suggest, but introduces a sustained multi-year headwind to new-book pricing power if customers internalize automation via agent templates.

Key catalysts that will determine winners: (1) reliability/accuracy regimes of agents (enterprise tolerance for failure is low), (2) vendor partnerships that bake models into existing ERP/ITSM backplanes, and (3) regulatory or vendor-imposed guardrails that either slow or standardize adoption. Any of the three can swing adoption curves from gradual (3–5 years) to fast (12–24 months).

From a portfolio perspective this is a classic structural re-rate trade: own compute/integration exposure (convex to usage) and express skepticism on licensing-based growth at current multiples. Near-term flow-driven derating can continue, but the true test is revenue mix change over the next 4 quarters — not daily headlines.

Key Decisions for Investors

• Pair trade (12 months): Short NOW vs Long NVDA — equal-dollar notional. Size 1–2% NAV. Rationale: express secular margin shift to compute. Target 30–50% relative outperformance in 6–12 months; stop-loss if relative performance reverses by 15% or if NOW reports two consecutive beats-with-raise events.
• Long NVDA (24 months): Buy LEAP calls (12–24 month tenor) representing ~1–2% NAV exposure to capture convexity of increased inference demand. Reward asymmetric: 2–4x upside if adoption and pricing persist; downside limited to premium (plan for full premium loss scenario).
• Tactical contrarian (6 months): Initiate a small long position in NOW sized 0.5–1% NAV lower-cost-basis if fundamentals stay intact, funded by selling 30–60 day OTM calls to finance protective puts (buy 15–20% OTM puts). This hedged-bull structure profits from mean reversion while capping downside.
• Event hedge (3–9 months): Monitor enterprise earnings/guidance for integration partnerships; if multiple vendors announce deep model integrations, accelerate NVDA exposure and trim NOW shorts. Conversely, if regulators impose strict controls on agent actions, reduce NVDA call exposure by 30–50% within one trading week.