Fastly Stock Has Already Doubled This Year. Can It Keep Climbing?

Fastly reported Q4 revenue of $172.6M, up 23% year over year; trailing-12-month net retention rose to 110% and RPO increased 55% YoY to $353.8M (current portion +37% YoY). Management guided FY2026 revenue to $700–$720M (~14% YoY at the midpoint) and expects infrastructure capex of ~10–12% of revenue in 2026 (vs 5% in 2025). The company achieved non-GAAP profitability and positive free cash flow in Q4 but remains GAAP-loss-making with a $15.5M Q4 net loss (improved from $32.9M). At ~5x price-to-sales, valuation leaves little room for execution missteps despite AI-driven demand tailwinds.

Analysis

Edge-AI demand is real but it morphs the cost base: inference-heavy traffic amplifies bandwidth, egress and peering costs and shifts value toward nodes with accelerators. That creates a two-sided pressure for an edge operator — revenue per session can rise, but unit economics deteriorate unless hardware mix and pricing both re-rate in their favor over 12–24 months.

Competitive dynamics favor vendors that can bundle software and hardware or monetize security/management layers (bot mitigation, agent governance) as differentiated SaaS margins. Legacy CDNs and hyperscalers that can underprice pure-network capacity while offering integrated developer platforms are the clearest secular threats; conversely, silicon vendors and colo partners stand to capture the capital footprint if operators outsource the hardware layer.

Execution risk centers on the capex cadence and contracted backlog conversion: a miss in either will compress multiples quickly because the story is now priced for near-flawless execution. Near-term catalysts to watch are conversion rates of contracted bookings into billings and any shift to hardware-accelerated footprints; either will materially change free cash flow profiles over the next 2–4 quarters.

The asymmetric trade is therefore optionality rather than naked exposure — structured positions that limit downside but leave upside if Fastly converts backlog and controls capex are higher-probability plays than outright long equity. Simultaneously, being long the supply-side beneficiaries of edge AI (silicon/accelerator and colo) is a less binary way to capture the secular tailwind without dependent execution risk on a single, capital-intensive operator.

Key Decisions for Investors

• Buy FSLY 3–6 month put spread (buy puts at-the-money, sell 20–30% lower strike) sized to 1–2% portfolio risk — objective: capture a >2x payoff if execution/capex misses; cut if contractual conversion rates accelerate on next quarter.
• Pair trade: short FSLY equity (or synthetic via options) / long AKAM or NET dollar-neutral, 6–12 month horizon — play relative execution: if FSLY stumbles on capex or gross margins, expect 20–40% relative underperformance; trim on 15% absolute adverse move.
• Long NVDA 9–12 month call spread to play edge-inference acceleration (buy nearer-term calls, sell higher strike) — targets 2–4x payoff if edge GPU demand materializes, with defined premium at risk; reduce position if inventory or pricing relief is announced by silicon vendors.
• Event trigger: set alerts for quarterly guide, disclosed capex cadence, and contracted backlog conversion — if backlog-to-billings conversion slows sequentially, add to short exposure; if conversion accelerates and capex is offset by third-party partnerships, rotate into capped long/options exposure.