Tale of the tape
| Akamai (AKAM) | Cloudflare (NET) | |
|---|---|---|
| Price, Sep 25, 2026 | ~$115 | ~$355 |
| Market cap | ~$17B | ~$126B |
| FY2025 revenue | $4.21B (+5%) | $2.17B (+30%) |
| Latest quarter | Q2 2026: $1.10B (+5%) | Q2 2026: $696M (+36%) |
| GAAP profitable | Yes ($452M net income, 2025) | No (-$102M net loss, 2025) |
| Free cash flow, 2025 | $699M | $261M |
| Cash and securities | $4.62B | $4.1B |
| EV / revenue (LTM) | ~4x | ~47x |
| Network | 4,300+ locations, 130+ countries | 330+ cities, 120+ countries |
| Share count trend | -18.5% since 2015 | Diluting (SBC >20% of revenue) |
| CEO | Tom Leighton, co-founder | Matthew Prince, co-founder |
| Founded | 1998 | 2009 |
Prices moved intraday on September 25; treat as approximate.
1. Business diversity: Akamai 7, Cloudflare 8
Akamai discloses three segments: in Q2 2026, Security did $604M (+10% YoY), Delivery and other cloud applications did $396M (-6%), and Cloud Infrastructure Services did $99M (+39%). Two engines growing, one shrinking, and the shrinking one is still 36% of revenue, which is why consolidated growth sits at 5% while the interesting parts grow at double digits.
Cloudflare reports a single segment and discloses no product-level revenue, a transparency minus. But the growth is broad-based: large customers ($100K+ ARR) now drive 73% of revenue (up from 63% in 2022), remaining performance obligations grew 48%, and new annual contract value grew nearly 50%. When everything grows near 30%, diversification matters less.
2. Product portfolio: Akamai 7, Cloudflare 9
Akamai's catalog is enterprise-grade and deep: App & API Protector, Bot Manager, Account Protector, Guardicore Segmentation ($293M in 2025, up 43%), Prolexic, Enterprise Application Access, Edge DNS; Ion and Adaptive Media Delivery; and a Linode-built compute stack (CIS, Inference Cloud, video processing units, managed containers). It wins RFPs. It does not win developers' hearts.
Cloudflare's catalog is the broadest developer platform at the edge: CDN, DNS, unmetered DDoS, WAF, Bot Management, API Shield; Cloudflare One (full SASE: Access, Gateway, CASB, DLP, Magic WAN, WARP); and the developer cloud: Workers, Pages, R2 with zero egress fees, D1, KV, Queues, Vectorize, AI Gateway, Workers AI, and Pay Per Crawl, the June 2026 HTTP 402 mechanism letting publishers charge AI crawlers (1 billion payment-required responses a day). Self-serve sign-up and a generous free tier make the top of the funnel enormous.
3. Edge compute and AI: Akamai 9, Cloudflare 7
Cloudflare is uniform: every one of its 330+ PoPs runs the full stack. Workers run on V8 isolates with sub-5-millisecond cold starts; Workers AI puts NVIDIA H100 GPUs in 180+ cities with serverless per-request pricing across 89+ open models. The most elegant inference fabric on earth for a developer. The limitation is physics: you are not running a 70B-parameter model at the far edge.
Akamai is tiered: GPU clusters of RTX PRO 6000 Blackwell cards at metro edge sites, dedicated clusters for heavy multimodal reasoning, regional and core capacity behind them, stitched by the NVIDIA AI Grid orchestration layer. Less elegant, more proven at enterprise scale: $14.4B of signed multi-year commitments against Cloudflare's largest disclosed deal at $42.5M per year. Signed enterprise backlog is the market's verdict.
4. Network scale: Akamai 9, Cloudflare 8
Akamai: 4,300+ points of presence, 130+ countries, 700+ cities, 345,000+ edge servers, over a petabit per second of throughput. Twenty-five years of embedding inside ISP networks is not replicable on a budget or a timeline. Cloudflare: 330-337 cities, 120-125+ countries, ~300-500 terabits per second (sources vary), 13,000+ interconnections, ~20-23% of all websites, 230 billion threats blocked daily.
Honest nuance: Akamai's count includes thousands of small embedded clusters, often a rack or two inside someone else's facility, that cannot run heavy compute. Cloudflare's fewer locations are each a real data center presence. Raw reach goes to Akamai; uniformity goes to Cloudflare.
5. Growth: Akamai 5, Cloudflare 10
| 2023 | 2024 | 2025 | Latest quarter | 2026 guide | |
|---|---|---|---|---|---|
| Akamai revenue growth | n/a | n/a | +5% | +5% (Q2 2026) | ~+6 to 8%; low teens in 2027 (mgmt) |
| Cloudflare revenue growth | +33% | +29% | +30% | +36% (Q2 2026) | ~+29% |
Cloudflare is growing roughly six times as fast and accelerating: Q2 2026 printed 35.9% with RPO up 48% and new ACV up nearly 50%. Akamai's consolidated 5% hides a split screen: Security +10%, CIS +39%, Delivery -6%. Management's line of sight to low-teens growth in 2027 depends on the Anthropic backlog converting to revenue starting in the second half of 2027.
6. Profitability and cash flow: Akamai 8, Cloudflare 5
Akamai is a real business by GAAP: $452M of net income in 2025, non-GAAP operating margin guided at 25-26% for 2026, $699M of free cash flow (~17% of revenue), mid-60s gross margin. The near-term wrinkle: Q2 2026 GAAP operating margin fell to 7% as depreciation and capex ramped ahead of contracted revenue.
Cloudflare is profitable only after adjustments: GAAP net loss of $102M in 2025 with GAAP operating margin near negative 10%. The bridge is stock-based compensation: $451.5M in 2025, over 20% of revenue, excluded from the non-GAAP operating income of $89.6M (14.6% margin). Free cash flow was $261M (12%, improving); gross margin 73.6%, compressed from 76.4%. When a fifth of revenue goes to employees in stock, shareholders should read the GAAP line first.
7. Balance sheet and capital allocation: Akamai 8, Cloudflare 6
Akamai: $4.62B of cash and securities against ~$2.95B of net debt, investment-grade rated, converts that never exceeded a 1.125% coupon, every issue paired with buybacks or note hedges. Share count down 18.5% from 2015 to 2025 on ~$6.1B of repurchases. The 2026 buyback pause is explicitly to fund contracted CIS demand, not distress. Cloudflare: $4.1B of cash and securities, converts outstanding (net debt ~$2.4B per secondary data), no buyback record, ongoing SBC dilution. Both can fund their AI expansions internally. Only one has a decade of shrinking the share count.
8. Valuation: Akamai 9, Cloudflare 3
Akamai at ~$115 is about 4x FY2025 revenue and about 17x forward non-GAAP earnings ($6.40 to $7.05 guidance midpoint), even after the Anthropic rip. Trailing GAAP P/E near 41x is depressed buildout earnings on a cheap enterprise value: the classic value-setup shape.
Cloudflare at ~$355 is about 50x trailing revenue (46.6x EV/revenue LTM) with negative GAAP EPS and roughly 280x forward non-GAAP earnings (analyst FY2026 non-GAAP EPS near $1.25). Up ~78% year to date, near its 52-week high. Priced for 30% growth compounding for a very long time, with no room for a stumble. One of these is priced as an option on the future. The other is priced as a business.
9. Leadership: Akamai 9, Cloudflare 9
Tom Leighton co-founded Akamai in 1998, has run it as CEO since 2013, takes a $1 salary since 2013 with nearly all pay at risk, and bought 50,000 shares in the open market in August 2025. Under-promise, over-deliver, engineer. Matthew Prince co-founded Cloudflare in 2009 and built it from a TechCrunch Disrupt launch to a $126B company: the industry's clearest communicator of the "agentic internet" thesis and a relentless product shipper. More promotional than Leighton, and the stock-compensation culture starts at the top, but founder-led intensity is not in doubt.
10. Moat: Akamai 8, Cloudflare 9
Akamai's moat is physical and contractual: 4,300 embedded locations that cannot be rebuilt cheaply or quickly, two decades of enterprise trust, and $14.4B of contracted backlog on infrastructure it already owns. Wide but static; it deepens as footprint converts into contracts. Cloudflare's moat is the flywheel: 23% of websites behind it, a developer ecosystem where each product cross-sells the next, and telemetry from a fifth of web traffic feeding its security products. It compounds with usage. Cloudflare's moat compounds faster; Akamai's is harder to replicate with money alone.
Final scorecard
| Section | Akamai | Cloudflare |
|---|---|---|
| 1. Business diversity | 7 | 8 |
| 2. Product portfolio | 7 | 9 |
| 3. Edge compute and AI | 9 | 7 |
| 4. Network scale | 9 | 8 |
| 5. Growth | 5 | 10 |
| 6. Profitability and cash flow | 8 | 5 |
| 7. Balance sheet and capital allocation | 8 | 6 |
| 8. Valuation | 9 | 3 |
| 9. Leadership | 9 | 9 |
| 10. Moat | 8 | 9 |
| Total | 79 | 74 |
The verdict, in full
Buy the business per dollar: Akamai. Real profits, real cash flow, contracted backlog at 4x sales, a management team with a decade of shrinking the share count.
Buy the growth: Cloudflare. The best developer platform at the edge, 30%+ growth with accelerating bookings, and a flywheel moat, at 47x sales.
The risk is symmetric and opposite. Cloudflare at 47x sales has priced in a decade of flawless execution; a single growth stumble reprices violently. Akamai at 4x sales has a margin of safety on price but real execution risk: $5.5B of capex ahead of revenue, customer concentration (Anthropic alone is ~2.75x FY2025 revenue), and a legacy Delivery business still eroding.
In five-pillar language: Akamai is the evidence-gated compounder at a reasonable price. Cloudflare is the wonderful business at a price that assumes the wonder never stops.
Open gaps and honest caveats
Cloudflare discloses no product-level revenue; the diversity and portfolio scores rest on company commentary, customer metrics, and bookings data. Network capacity figures (terabits per second) come from mixed primary and secondary sources and are not audited. Neither company discloses per-location compute or GPU counts; the technical section rests on announced deployments and architecture descriptions. Akamai's Anthropic agreement terms await the Q3 2026 filing; "take-or-pay" wording appears in press coverage but was not found in Akamai's own disclosures. Prices and market caps are intraday September 25, 2026 and moved during the session.