Hub
Institutional Equity ResearchCybersecurity · Edge, SASE & Developer Platform

NET
A Great Company, Full Price

A valuation-led note on NYSE: Cloudflare, the premium-multiple mirror image of Zscaler. The programmable edge (global CDN + security + Cloudflare One SASE + the Workers / Workers AI developer platform) is positioned as where agentic-AI workloads run. The quality is elite: revenue +36% in Q2, dollar-based net retention up to 120%, remaining performance obligations +35% YoY, and free cash flow still inflecting. But it is priced for all of it: after a +52% YTD run to ~$300, NET trades at ~28x forward sales, and the Street's average target (~$250) now sits below spot. The 20% "agentic-AI-first" workforce cut is a margin tailwind and a real execution risk. Fairly valued. We rate NET a Hold.

Naina Garg · Data as of August 7, 2026
Read the Report ↓ Try the Scenario Model
$300
12-mo Price Target
Hold
Rating
0%
Base Upside
~28x
FY27E EV/Sales
120%
DBNRR
NETCloudflare, Inc. · NYSEHold
Analysis: Aug 7, 2026
Last$300.27
YTD+52%
52w$158.83–$305.20
Mkt Cap$109.6B
Fwd P/E238x
PT$300

Snapshot: Executive Summary

One-page summary · institutional view

Cloudflare is one of the highest-quality growth franchises in software, and the market knows it. The "connectivity cloud" runs on a global network in 330+ cities that delivers CDN, security, Cloudflare One (SASE / Zero Trust), and the Workers serverless developer platform as one integrated system. In Q2 2026 revenue grew +36% to $696.1M, dollar-based net retention rose to 120%, remaining performance obligations grew +35% YoY, and non-GAAP operating income reached $96.1M (13.8% margin) with free cash flow still inflecting1. This is a business firing on every cylinder.

So why only a Hold? Because you are paying for all of it, and then some:

  • The multiple prices perfection. After a +52% YTD re-rate to ~$300, NET trades at ~43x trailing sales, ~28x FY27E sales, and ~240x forward non-GAAP earnings, the priciest name in our cyber set (PANW ~14x, CRWD ~21x, ZS ~7x). The Street's average target (~$250) now sits below the ~$300 spot, still catching up to the move6.
  • The AI story is real but early. Workers AI (edge inference, an AI gateway, vector DB, the agentic push) is the differentiator, but Cloudflare does not yet break out AI monetization as a discrete line. The premium is paying for optionality, not yet proven ARR.
  • The 20% cut cuts both ways. The "agentic-AI-first" restructuring (a GAAP operating loss of ($205.7M) ran through Q2) is a deliberate step-up in margin and FCF, but layoffs of that size can dent go-to-market execution just as growth is decelerating from the low-30s3.

Our 12-month target is $300 (Hold), which equals spot and sits above the lagging Street average (~$250). It is derived from the same EV/Sales math: FY27E revenue of ~$3.75B at ~28x sales, plus net cash, rolled forward. The risk/reward from ~$300 is roughly balanced: our base equals spot, a two-sided distribution around it. We would move to a Buy on evidence the agentic edge is a durable second engine (re-accelerating revenue with the multiple holding) or a durable step-up in FCF margin. Absent that proof, the better-paid entry is lower, toward the mid-$250s support. This completes our cybersecurity coverage alongside PANW (Buy), CRWD (Hold), and ZS (Hold).

Rating
Hold
12-mo PT $300
Q2 Revenue
$696.1M
+36% YoY
DBNRR
120%
+8pp YoY
FCF margin
~13%
inflecting
Mkt Cap
$109.6B
$300.27 × ~365M sh
FY27E EV/Sales
~28x
priciest in cyber

Tactical: NET at ~$300 sits right on our $300 target after a +52% YTD re-rate; the Street average (~$250) is below spot and lagging the move. Elite quality at a full multiple; the consensus is still Overweight, so this is a "great company, full price" Hold, not a bearish call. Rating Hold; the swing variables are Workers-AI monetization and whether ~30% growth holds through the restructuring.

Investment Thesis

Bull · Base · Bear · Rating

Bull Case

$360
+20% vs spot · re-accel + re-rate
  • Workers AI and the agentic edge inflect into a disclosed, growing second engine the market underwrites on its own merits
  • Revenue re-accelerates toward the mid-30s; DBNRR pushes further into the 120s
  • The restructuring lifts FCF margin sharply (Rule-of-40 comfortably in the 50s); the multiple holds or expands toward ~33x
  • FY27 revenue ~$3.95B at ~33x sales, plus net cash, supports ~$360

Base Case

$300
0% upside · roughly fair value
  • Revenue decelerates gently from the low-30s to ~31% in FY27; Cloudflare executes as a premium ~30% grower
  • FCF margin expands with the restructuring; the multiple holds ~flat at ~28x EV/Sales
  • The AI optionality is real but stays priced-in; SASE and enterprise adoption keep compounding
  • FY27 revenue ~$3.75B at ~28x sales, rolled forward, lands ~$300 (most likely, 45% weight)

Bear Case

$210
-30% vs spot · premium compresses
  • Growth normalizes toward the low-20s as enterprise budgets tighten and hyperscaler edge/serverless competition bites
  • The premium compresses toward ~21x EV/Sales as the market de-rates a decelerating SaaS
  • The 20% cut disrupts go-to-market; GAAP profitability stays elusive
  • FY27 revenue ~$3.55B at ~21x sales re-tests toward $210

Rating: Hold. The probability-weighted blend (30% bull / 45% base / 25% bear) is ~$296, versus the ~$300 spot. We rate NET a Hold: the quality is not in question, but the multiple already discounts it, so upside to a conservative base is thin and the distribution is two-sided. Upgrade trigger: a discrete Workers-AI / agentic-edge ARR disclosure, or a durable FCF-margin step to ~20%. Downgrade trigger: revenue decelerating below the mid-20s, or the restructuring visibly denting bookings.

The Programmable Edge: One Integrated Network

CDN · security · Cloudflare One · Workers · sold as the connectivity cloud

Cloudflare sells a network, not a point tool. Every service runs on the same global fabric (330+ cities), so a customer that starts with CDN or DDoS protection can adopt security, Zero Trust, and serverless compute without adding vendors or hardware. That architectural integration is the moat: the marginal product ships to the entire network at near-zero incremental cost, and the switching cost compounds as customers consolidate onto it. Three pillars carry the revenue.

Revenue mix by pillar (% of revenue, estimated)

Application services (CDN + security) remain the anchor; Network / Zero Trust (Cloudflare One) and the Developer platform (Workers) are the growth surface. Mix is an author estimate: Cloudflare does not report revenue by pillar.

Large customers (>$100K annualized)

The enterprise engine keeps compounding: 4,698 customers now spend >$100K per year (+27% YoY), and remaining performance obligations grew +35% YoY4. The move upmarket is the story behind the DBNRR expansion.

The three pillars

  • Application services, the origin and still the anchor: CDN, WAF, DDoS mitigation, bot management, and DNS. The land motion that puts Cloudflare in front of a customer's traffic.
  • Network services / Cloudflare One, the SASE and Zero Trust suite (secure web gateway, ZTNA, CASB, DLP, magic transit). This is where Cloudflare overlaps the pure-play SSE names (ZS, PANW) and where enterprise consolidation deals land.
  • Developer platform / Workers, the serverless compute layer (Workers, Workers AI, R2 storage, D1, KV, Durable Objects, an AI gateway and vector DB). The "build anything on our network" surface and the agentic-AI substrate covered next.

Workers & Workers AI: The Agentic Edge

The differentiator · inference at the edge · the optionality the multiple is paying for

The bull case lives here. Cloudflare's pitch is that agentic AI does not want to run in a distant hyperscaler region; it wants to run close to the user, on a programmable network, with governed access to data and tools. Workers is the serverless runtime; Workers AI adds edge inference, an AI gateway (observability and routing across model providers), a vector database for retrieval, and the primitives for autonomous agents. The Q1 restructuring was explicitly framed as re-tooling the whole company to an "agentic-AI-first operating model."

The honest counter is that this is optionality, not yet proven revenue. Cloudflare does not break out Workers AI monetization as a discrete, growing ARR line, so the market is paying a ~28x sales premium partly for a story it cannot yet size. The single most important catalyst in the name is a disclosure that turns the agentic edge from a narrative into a number. Until then, the developer platform is a fast-growing but still-minority slice of revenue, and the premium is an act of faith in the roadmap.

Workers platform
Serverless
edge compute + storage
Workers AI
Inference
at the edge, 330+ cities
AI gateway
Routing
observability + governance
Disclosure
no discrete AI ARR yet

Note: Cloudflare also sits at the center of the "AI scraping" debate, offering tools that let publishers control and charge for AI crawler access. Strategically resonant, but not yet a material revenue line. Treat the agentic-edge thesis as high-optionality, low-proof, which is exactly why we hold rather than chase.

Growth & Retention

Revenue · DBNRR recovery · the deceleration debate

The reason a premium Cloudflare is defensible and not a bubble is the durability of the growth. Revenue has compounded above 30% for years, and the key retention metric keeps climbing: dollar-based net retention troughed near 110% and is now up to 120%. The debate is not whether Cloudflare grows; it is the pace of the inevitable deceleration off a larger base.

DBNRR recovery (%)

Dollar-based net retention rose to 120% in Q2 2026, up 8 points YoY, off a ~110% trough. Existing customers are expanding again, the cleanest sign the franchise is healthy.

Revenue growth (% YoY)

Growth is steady in the low-30s (Q2 +36%, FY26 guide raised to +32%). A gentle deceleration off scale, not a break. The FY27E bar is an author estimate.

Q2 revenue
$696.1M
+36% YoY
>$100K customers
4,698
+27% YoY
RPO
+35%
YoY backlog build
FY26 guide
~$2.87B
+32% (raised)

Competitive Landscape

Edge, CDN, serverless, and the SASE overlap

Cloudflare competes on multiple fronts, and its edge is the single integrated network that spans all of them. The table frames the field.

RivalVectorThreat levelCloudflare's defense
Hyperscalers (AWS / Azure / GCP)Edge / serverless (Lambda@Edge, CloudFront, Front Door) and AI inferenceHighA purpose-built global edge network and developer experience, not a bolt-on to a region-first cloud
AkamaiLegacy CDN + security incumbent moving to edge computeMediumCloud-native architecture, faster product velocity, integrated security
FastlyDeveloper-first CDN / edge computeMedium-lowFar larger scale, broader security and Zero Trust attach, enterprise depth
Zscaler / Palo AltoSASE / Zero Trust overlap with Cloudflare OneMediumSingle-network delivery and price; bundled with CDN/security the SSE pure-plays lack
Microsoft (Entra / E5)Bundled Zero Trust into E5 licensesMediumSuperior network breadth and developer platform; not tied to the Microsoft estate

The honest read: Cloudflare rarely loses on architecture. Its risk is that the hyperscalers, which own the AI-compute relationship, extend "good enough" edge and inference to their own customers, and that the SASE overlap with ZS and PANW turns Cloudflare One into a price fight. Neither is fatal, but both cap how far the premium multiple can stretch.

Financial Health & Trends

Elite gross margin, an inflecting FCF profile, and heavy SBC

Annual revenue ($B)

Revenue has more than doubled in three years ($0.98B FY22 to $2.17B FY25) and is guided to ~$2.87B in FY26 after the Q2 raise2. FY27E and FY28E are author estimates.

Margins (%): non-GAAP op margin & FCF margin

The profitability inflection: non-GAAP op margin climbed from 4% (FY22) to ~14% (13.8% in Q2); FCF margin went from negative to ~13-14%. The 20% cut is designed to accelerate this.

Q2 2026 print highlights

MetricQ2 2026YoYNote
Revenue$696.1M+36%above guide
Non-GAAP gross margin73.1%elite, network scale
Non-GAAP operating income$96.1M13.8% margin
Non-GAAP EPS$0.29FY26 est $1.25-1.26
GAAP operating loss($205.7M)restructuring charge
RPO+35%YoYbacklog / durability

Note: GAAP swung to an operating loss of ($205.7M) on the restructuring and heavy stock-based compensation, while the non-GAAP margins carry an SBC add-back. The cash economics are inflecting nicely, but a quality-aware investor should haircut for SBC dilution and treat the layoff-driven margin step as needing a couple of quarters of proof.

Capital & Cash

Modest net cash · convertible notes · network capex

Cloudflare carries $4.17B of cash and investments against ~$3.52B of total debt, most of it convertible senior notes (with a ~$1.29B current portion approaching maturity), for roughly $0.65B net cash5. It pays no dividend and has no buyback of note; capital goes to network capex (the global edge is capital-intensive) and R&D. Net cash is small relative to the ~$110B market cap, so the balance sheet is a footnote to the valuation rather than a driver of it. The convertible structure means a chunk of the "debt" could convert to equity rather than consume cash if the stock stays elevated.

Cash & investments
$4.17B
Q1 2026
Total debt
~$3.52B
mostly converts
Net cash
~$0.65B
cash less total debt
Dividend
none

The network is a real asset (330+ cities of owned infrastructure) but also a real cash cost: capex runs roughly $75M a quarter. The FCF inflection is impressive precisely because it comes despite that ongoing network build.

Valuation

The premium is the whole debate

Unlike Zscaler (cheap and decelerating), Cloudflare is expensive and compounding. After a +52% YTD re-rate it trades at ~43x trailing sales, ~28x FY27E sales, and ~240x forward non-GAAP earnings6. That is a multiple you only pay for a franchise the market believes will grow 25-30% for a very long time with expanding margins. The question is not whether Cloudflare is great; it is whether ~28x forward sales leaves any margin of safety.

NTM EV/Sales: the premium

NET has re-rated up to ~28x on the AI narrative and FCF inflection, the opposite of the SaaS de-rating that hit most peers. That leaves little cushion if growth or the story wobbles.

EV / NTM sales: cyber & edge peers

At ~28x, NET is the priciest name in the set, well above CRWD (~21x), PANW (~14x), and ZS (~7x), and a multiple of the CDN/edge incumbents. The premium is the story.

The framing

Our base target of $300 does not assume the multiple compresses; it assumes ~28x holds on a still-compounding, FCF-inflecting base. That is roughly fair from spot. The Street average (~$250) still sits below spot, lagging the re-rate, with the consensus rating Overweight7; the upside case lives above if the agentic edge inflects. Our internal bull ($360) and bear ($210) bracket that debate. The cleaner way to see it is to build it, which the interactive model does next.

EV/Sales Scenario Model

FY27E revenue × multiple + net cash, rolled forward · scenario tabs · sensitivity grid

An EV/Sales model is the right tool for a high-growth software name with modest GAAP earnings: project FY27E revenue × an EV/Sales multiple, add net cash, divide by shares, and roll the spot fair value forward twelve months. Toggle Base / Bull / Bear / Reset and drag the revenue and multiple sliders; the per-share value recomputes live. Every scenario reconciles to the published targets ($210 / $300 / $360) and to the blended fair value of ~$296.

Scenario Inputs

$3.75B
28.0x
$300
12-month price target
0% vs $300

Reset (reverse) holds the base anchors so you can drag the sliders to see what revenue and multiple the current price implies.

The Build

FY27E revenue × EV/Sales → enterprise value → equity → per share → rolled forward (reloads with the scenario tabs)
FY27E revenue$3.75B
× EV/Sales28.0x
= Enterprise value$105.0B
+ Net cash+$0.65B
= Equity value$105.7B
÷ Shares365M
= Spot fair value$289
× 12-mo rollroughly fair (x1.036)

For a Hold the spot fair value (~$289) sits close to the 12-month target ($300), so the forward roll is small. Bull and bear flex both the revenue and the exit multiple. All forecast figures are author estimates.

Sensitivity: PT / share ($) vs FY27E revenue × EV/Sales

Base-case share count, net cash and forward roll across the grid (FY27E revenue × EV/Sales multiple). The highlighted cell tracks the active scenario's anchors.

PT calculator (quick cross-check)

Implied PT
$300
0% vs $300

A simple revenue × multiple + net cash cross-check on the base scenario. ~$3.75B at ~28x, plus ~$0.65B net cash over ~365M shares, lands at the $300 base.

Risk / Reward calculator

R/R
checking…
0 / 2000

Note: The assistant reasons from the dashboard's data snapshot and thesis sections. It does not browse the web or access real-time fundamentals beyond what is in data.js. Treat its responses as scenario-modeling support, not primary research. Author judgments on rating, PT, and probabilities remain with the analyst.

Upcoming Catalysts

Next 12 months
CatalystWindowWhy it matters
Q3 2026 earnings~Nov 5, 2026The read on whether revenue holds ~30% and DBNRR keeps climbing, plus more of the restructuring flowing through to margins and FCF.
Workers AI / agentic-edge disclosureOngoingThe single most important catalyst: any discrete, growing AI-monetization number would turn the premium from optionality into underwritable ARR and could re-rate the stock.
FCF-margin step-up2H 2026Evidence the 20% cut is a durable margin/FCF inflection (toward ~20%) rather than a one-off, which is what justifies holding the ~28x multiple.
Large-customer & RPO momentumEach quarterThe move upmarket is the engine behind DBNRR; continued large-customer adds and RPO growth keep the durability case intact.
Restructuring execution2H 2026Whether a 20% headcount cut disrupts go-to-market and bookings, the key risk to the growth trajectory.
Software multiple regimeOngoingAs the most expensive name in the set, NET is the most exposed to any broad software de-rating or rate move.

Risk Factors (Including Upside Cases)

What could break, or re-rate, the premium
  • Multiple compression. The pre-eminent risk: at ~28x forward sales, even a modest growth disappointment or a broad software de-rating hits NET disproportionately. There is little margin of safety in the multiple.
  • Deceleration below the mid-20s. If growth normalizes faster than expected as the base scales and enterprise budgets tighten, the premium is hard to defend and the bear path toward $210 opens.
  • Restructuring disruption. A 20% workforce cut is a large, simultaneous shock to sales, product, and support. If it dents bookings, the margin win is offset by slower growth, the worst of both worlds.
  • Hyperscaler competition. AWS, Azure, and Google own the AI-compute relationship and can extend "good enough" edge and inference. If they close the developer-experience gap, Cloudflare's agentic-edge premium erodes.
  • AI optionality stays optionality. If Workers AI never becomes a disclosed, material revenue line, the market eventually stops paying a premium for a story it cannot size.
  • Upside case: the second engine is real. The symmetric risk to a Hold is that Cloudflare discloses a fast-growing agentic-edge line and the FCF margin steps to ~20%. That would justify the premium and move us to a Buy toward $330+. This is why we hold rather than sell a generational franchise.

Bull vs Bear Debate

The hardest questions, both sides
IssueBull viewBear view
Valuation ~28x sales is fair for a 30% grower with an FCF inflection and a generational platform; quality compounders never look cheap. ~28x forward sales for a decelerating ~30% grower with GAAP losses leaves no margin of safety.
Workers AI The agentic edge is where AI inference is going; Cloudflare owns the network and the developer mindshare. It is undisclosed optionality; the premium pays for a story the company cannot yet size or monetize.
Growth durability DBNRR up to 120%, RPO +35% YoY: enterprise adoption is accelerating, not fading. Growth is decelerating off scale; the law of large numbers and budget scrutiny catch every SaaS.
The 20% cut A deliberate step-up in margin and FCF that proves capital discipline and funds the AI pivot. A large go-to-market shock that risks bookings just as growth needs defending.
Competition The single integrated network is a durable architectural edge over hyperscalers and CDNs. Hyperscalers own AI compute and can bundle "good enough" edge; SASE overlap invites a price fight.

Our resolution: the bull and bear are close enough that the stock is roughly fairly valued from ~$300. We rate NET a Hold and would need a discrete agentic-edge disclosure or a durable FCF-margin step to underwrite the Buy case.

Technical Analysis

Price path · overlays · momentum · relative strength

Price path (12 months) with overlays

Spot $300.27, 52-week range $158.83-$305.20. -1.6% vs $305.20 high. The stock is trading near its highs after a +52% YTD run off the February low; the more constructive add zone sits toward the mid-$250s support, where the distance to the bull case widens.

RSI (daily / weekly / monthly)

Momentum is firm but not stretched: RSI in the 55-60 zone across timeframes, consistent with an uptrend that is consolidating rather than breaking down.

2026 YTD relative strength (%)

NET (+52% YTD) now leads the cyber set after its re-rate, edging out PANW and well ahead of CRWD and the software index. A strong tape running near the highs.

Sources & Footnotes

The quarterly print, guidance, and market data
  1. Cloudflare Q2 2026 results (quarter ended June 30, 2026), reported Aug 6, 2026: revenue $696.1M (+36%), non-GAAP gross margin 73.1%, non-GAAP operating income $96.1M (13.8% margin), non-GAAP EPS $0.29, GAAP operating loss ($205.7M), dollar-based net retention 120%, RPO +35% YoY. Cloudflare Investor Relations →
  2. FY2026 guidance raised (Q2 2026 call): revenue $2.864-2.870B (+32%), non-GAAP EPS $1.25-1.26; Q3 2026 guide revenue $736-737M (+30%), non-GAAP EPS $0.34. Next report ~Nov 5, 2026.
  3. The 20% workforce reduction (~$140-150M charge), framed as an "agentic-AI-first" restructuring, announced with the Q1 2026 results; the GAAP operating loss of ($205.7M) in Q2 reflects the related charges.
  4. Enterprise momentum: remaining performance obligations grew +35% YoY; 4,698 customers paying >$100K per year (+27% YoY). Per the Q2 2026 print.
  5. Balance sheet (last reported, Q1 2026, March 31, 2026): cash & short-term investments $4.17B; total debt ~$3.52B, largely convertible senior notes (with a ~$1.29B current portion); net cash ~$0.65B. Cloudflare Investor Relations →
  6. Valuation: ~43x trailing sales (P/S), ~28x FY27E EV/Sales, and ~240x forward non-GAAP EPS ($300.27 / $1.26).
  7. Sell-side average target ~$250 per WSJ (Aug 2026), below the ~$300 spot after the +52% YTD re-rate; consensus Overweight. Our internal bull $360 / bear $210 bracket the debate. WSJ market data →

All figures are drawn from Cloudflare's public filings and press releases, the Q2 2026 earnings materials, and WSJ market data as of August 7, 2026. Forward estimates (FY27E, FY28E) and the revenue-mix breakdown are author estimates used for the model. Conclusions are the author's view and are illustrative, not investment advice.

1 / 15 Snapshot