
Craig-Hallum downgraded Fastly to Hold from Buy and set a $24 price target, citing rising capex, memory and network hardware costs, and tougher comparisons as Edgio tailwinds fade. The firm still sees 13% fiscal 2026 top-line growth, but the current 35x EV/EBITDA valuation is viewed as fully reflecting the recent improvement. Fastly’s fundamentals remain mixed: gross margin is 57.08%, revenue is up nearly 15%, and AI traffic is rising, but margins and monetization may be pressured by AI-driven infrastructure demand and security-related pricing challenges.
The downgrade reads less like a fundamental bearish call and more like a valuation reset after a violent rerating. Fastly’s setup is now classic “good story, expensive stock”: the market has already paid for margin recovery, renewal leverage, and AI traffic mix improvement, so incremental evidence has diminishing upside unless growth re-accelerates materially. The more important second-order issue is that AI traffic may be helping utilization but not necessarily monetization; if bots and agents increase bandwidth while security and value-add features face pricing pressure, revenue can rise while operating leverage stalls. Competitive dynamics are worsening at the margin. As large incumbents and adjacent infrastructure vendors bundle edge, security, and app-delivery functionality, Fastly’s differentiated pricing power likely narrows just as capex intensity rises due to memory/network hardware inflation. That combination is dangerous because it forces the company to choose between preserving margin or defending share, and the market is implicitly assuming it can do both. The Edgio churn benefit also becomes a harder comp over the next few quarters, so the next leg of results may show less “easy” growth and more proof-testing of retention quality. The near-term catalyst path is asymmetric: upside from continued AI-driven traffic growth is slower and requires multiple clean quarters, while downside can arrive quickly if guidance implies gross margin or EBITDA expansion is plateauing. At ~35x EV/EBITDA, the stock is priced for sustained execution, so any sign that minimum-commit extraction is reaching saturation or that hardware inflation compresses 2026 margins could trigger another de-rating. The contrarian angle is that the selloff may overshoot if the company can convert AI traffic into higher committed spend and security attach, but that is a longer-dated thesis than the current multiple implies.
AI-powered research, real-time alerts, and portfolio analytics for institutional investors.
Overall Sentiment
mildly negative
Sentiment Score
-0.25
Ticker Sentiment