AI Infrastructure Trio Showdown: Datadog, Cloudflare, and JFrog

The AI boom has done more than create GPU and model companies — it has fueled an entire "behind-the-scenes monitoring" infrastructure chain. From training and deployment to protecting AI applications, every stage demands specialized tools. The three companies compared here — Datadog (DDOG), Cloudflare (NET), and JFrog (FROG) — cover the three core pillars of observability, network security, and software supply chain, making them some of the most direct beneficiaries of the AI infrastructure wave.

1. Datadog: AI Observability Leader Continues to Accelerate

Datadog posted Q3 2025 revenue of $886 million, up 28% year-over-year, marking yet another quarter of growth sustained above 25%. In recent years, the company has aggressively pushed LLM Observability, extending its traditional APM, logging, and infrastructure monitoring into AI model inference costs, token usage, and response latency — precisely the visibility enterprises need most when deploying generative AI at scale.

Even more noteworthy is Datadog's net dollar retention rate, which has stayed above 110% for an extended period, reflecting existing customers expanding their usage. As AI customers move from POC to large-scale production, the company's customer ARPU has clear room to climb. The Q4 revenue guidance midpoint of $914 million came in above the consensus of $887 million, and the market responded with a 23% single-day gain.

That said, Datadog's valuation has remained rich, with a price-to-sales ratio persistently above 15x. The moment AI observability demand growth cools, valuation pressure will surface immediately.

2. Cloudflare: From CDN to AI Application Platform

Cloudflare posted Q3 2025 revenue of $562 million, up 31% year-over-year — the second consecutive quarter of acceleration. On the surface, growth was driven by cybersecurity and CDN revenue, but the deeper shift is the rapid expansion of the Workers developer platform and AI Gateway. This quarter alone, single-customer upsells across Workers and application services reached the $6.8 million level.

Management noted that RPO (remaining performance obligations) grew 43% year-over-year, signaling strong forward revenue visibility. Cloudflare's strategy is to converge networking, AI inference, and security into a single global edge network. Enterprise customers can call Workers AI directly on Workers for inference, paired with Magic Transit and Bot Management, forming a hard-to-replace closed loop.

For full-year 2025, Cloudflare's revenue ultimately reached $2.168 billion, a 34% increase — continuing its high-speed expansion. However, it is worth noting that the company remains at a small GAAP operating loss. The market is ultimately watching whether non-GAAP operating margin (about 15.3% in Q3) can keep expanding, and how quickly Workers AI monetizes.

3. JFrog: Software Supply Chain Meets the AI Security Tailwind

JFrog posted Q3 2025 revenue of $136.9 million, up 26% year-over-year, with cloud business (JFrog Cloud) growing even faster at more than 50% year-over-year and a net dollar retention rate of 118%. The AI wave has done more than drive model training — it has also driven enterprises to demand strict "software provenance." Once malicious model dependencies or backdoors are embedded in code, the consequences are far more severe than traditional vulnerabilities.

JFrog is positioned precisely as the "system of record for software supply chain" in the AI era. The company integrates Artifactory repositories, vulnerability scanning, and signature verification into a single platform, and has added AI/ML model artifact management. Management has explicitly positioned JFrog as the "authoritative system of record for software supply chain in the AI era" — a narrative that resonates strongly in an environment of tightening compliance requirements.

That said, JFrog remains much smaller in scale, and still has a long road ahead before it can match the size of Datadog or Cloudflare. The valuation premium the market gives it is largely for the long-term imagination space around AI security.

4. Comparing the Investment Theses

Datadog's core pitch is "essential AI observability" plus "high retention," making it suited to investors who believe the pace of AI application deployment will continue. Cloudflare's advantage lies in its platform breadth and its natural position at the edge for AI inference, fitting those who favor the long-term theme of decentralized AI inference. JFrog is the "AI security" small-cap play — high upside potential but also high volatility, better suited to portfolio allocation than a heavy concentration.

All three stocks share a common risk: elevated valuation and high sensitivity to interest rates and macro liquidity. Should the AI capex cycle slow, the market's optimism toward "AI infrastructure" could quickly recede. Add to that the fact that all three operate on a SaaS subscription model, making retention rates and customer concentration changes the key leading indicators — worth tracking closely around each earnings release.

Overall, AI infrastructure monitoring remains one of the few software segments still able to sustain high growth into 2026. Datadog, Cloudflare, and JFrog each represent a different path — observability, networking, and supply chain — and combining them in a portfolio helps diversify single-company risk while more fully capturing the dividend of AI application proliferation.