Refresh — what changed since the 2026-04-10 deep-dive

The prior deep-dive (2026-04-10-net-deep-dive.md) caught NET in the depths of the spring SaaS capitulation: $169.96, RSI 33.9, a confirmed death cross, −18.7% below SMA20. The note's call was that the fundamental thesis (edge compute, Workers AI, R2) was intact but "the market doesn't care about fundamentals during sector capitulation."

That bottom held. NET has since recovered to $237.24 (+~40% off the April low), flipped the death cross back to a golden cross, and re-established a strong-up trend. The fundamental story did roughly what the April note hoped — the AI-revenue narrative re-accelerated (Piper Sandler upgraded to Overweight with a $222 target on accelerating AI revenue, since exceeded), Q1 FY2026 beat with improving free-cash-flow margin, and management made a decisive efficiency move. The capitulation-bottom thesis paid; the question now is the re-entry on the pullback, not the bottom.

The Story Right Now

Cloudflare runs the connectivity-cloud layer — CDN, edge compute (Workers), object storage (R2), and Zero Trust security — and is positioning that edge network as the place AI inference and agents run close to users. The bull driver in the recent flow is exactly that: AI-related product adoption (Workers AI, agent infrastructure) lifting growth and large-customer expansion.

The financials are the SaaS-compounder shape: GAAP-unprofitable but strongly cash-generative. FY2025 revenue was $2.17B (+33.5% +29.8% YoY — correction 2026-08-03: the filed year grew +29.8% over FY2024's $1.67B; +33.5% is Q1 FY2026's year-over-year rate, mislabeled onto the annual figure) with an operating loss of −$207.2M and a net loss of −$102.3M (stock-comp-driven), but operating cash flow of +$603.1M — the tell that the business funds itself. Q1 FY2026 (period 2026-03-31) posted $639.8M revenue (+34% YoY), large-customer revenue up ~38%, operating cash flow +$158.3M, and a beat (non-GAAP EPS −$0.05 vs −$0.06 expected; GAAP diluted was −$0.07). Cash sits at $932.2M.

The most important recent corporate action: alongside Q1 (reported 2026-05-07), Cloudflare announced a ~20% workforce reduction (~1,100 roles), framed as a re-org to accelerate an "agent-era" AI strategy. That is double-edged — margin- and FCF-accretive (FCF margin already ticked to 13% from 11%), but a large headcount cut at a +33% grower is also a tell that management is prioritizing efficiency over land-grab growth, which fits the deceleration the leading indicator is flagging.

On the tape (settled Fri 2026-06-26): $237.24, RSI 54.6, strong-up, regime "pullback." Price is −0.7% below its rising SMA20 ($238.86), +6.8% above SMA50 ($222.11), and +14.2% above SMA200 ($207.80), golden cross active, on elevated volume (relative volume 1.53). It is +13.4% over 30 days and +16.9% over 3 months, only −14.3% off its 52-week high. The 2-year return is +186% but the last 1-year is +22% — i.e., a big run, a year of basing/consolidation, now re-accelerating. This is a confirmed uptrend resetting to its SMA20, not a chase.

Setup

• Entry zone: the pullback is the zone. Buy a hold/reclaim of the rising SMA20 (~$239) with RSI reset to ~54; add on a flag→breakout over the recent swing high. You are buying a confirmed uptrend on a shallow dip, not guessing a bottom.
• Stop: below the ~$220–222 shelf — SMA50 ($222.11) and the nearest support shelf ($220.28) stack there; a clean break voids the pullback structure. Deeper line: AVWAP-from-low ($210.61) and the rising SMA200 ($207.80) at ~$208–211 — a break under that would mean the recovery uptrend itself has failed.
• Target: retest the 52-week high, then trend continuation. Trail it — NET is high-beta, anchor to structure.
• Conviction: MEDIUM. Best-shaped chart on the board (reset, not extended), but decelerating growth + heavy insider selling + premium multiple cap it below high.
• Sizing note: normal trend-hold size on confirmation of the SMA20 hold; don't oversize a high-beta name into a decelerating top line. The April reversal (−19.7% in one week) is the volatility lesson — keep the stop honored.

Bull case

• Confirmed uptrend that has reset, not blown off. Golden cross, +14% above a rising SMA200, RSI 54.6, only −14% off the high — the rare combination of "in an uptrend" and "not extended." The pullback to the rising SMA20 is a clean trend-hold entry.
• AI-accelerated growth on a real platform. FY2025 +29.8% YoY (Q1 FY26 +33.5%), large-customer revenue +~38%, AI product adoption (Workers AI, edge inference, agent infrastructure) the cited driver; FY guidance above estimates.
• Self-funding economics. +$603M FY25 operating cash flow and a rising FCF margin (13% from 11%) despite GAAP losses — the SaaS rule-of-40 shape. $932M cash, no solvency question.
• Efficiency turn. The ~20% workforce cut is FCF-accretive and signals a margin-discipline pivot — if growth holds even at a decelerated rate, the FCF inflection is the re-rating lever.
• Capitulation already happened. The spring SaaS washout (the April deep-dive's $170/RSI-34 bottom) flushed weak hands; the recovery to $237 on a golden cross is the proof the panic was sector beta, not a broken business.

Bear case

• Revenue growth is decelerating — the one real flag. Leading-indicator revenue read is bearish: QoQ growth rate halved to +4.1% (from +9.3%), a −5.23pp deceleration even as YoY stays +33.5%. A premium multiple on a decelerating grower is the classic SaaS de-rate setup, and it's what bit NET in the spring.
• Heavy insider distribution. 0 buys / 100 sells, net −$738M — management and insiders selling aggressively into the recovery, not adding.
• The 20% layoff cuts both ways. Margin-positive, yes — but a large headcount reduction at a +33% grower can read as a growth-confidence tell, not just efficiency.
• Premium valuation / high beta. NET carries a rich multiple and a violent beta (the April note logged a −19.7% single-week reversal). It moves fast in both directions; a sector risk-off (Apollo's rotation out of high-beta growth) hits it first.
• Composite Neutral-negative (−0.2). The leading blend is not a clean buy — strong earnings/cash offset by the bearish revenue-deceleration and insider signals.

Catalysts

• SMA20 hold / reclaim (~$239) — the mechanical trigger that confirms the trend-hold add; a loss of the ~$220–222 shelf is the invalidation.
• Next quarterly report (~late July). The single most important read: does revenue growth stabilize or decelerate further, and does the FCF margin keep climbing post-layoff? This is the bull/bear referee.
• AI / agent-platform traction — Workers AI and agent-infrastructure adoption metrics; large-customer net-expansion. Sustained AI-revenue acceleration re-rates it; a stall confirms the deceleration bear.
• Post-restructuring margin print — whether the ~20% headcount cut shows up as the FCF-margin step-up management is implying.
• Software-rotation beta — per the Apollo work, software trades at its cheapest relative valuation in years; if money rotates back toward profitable/cash-generative software, NET (FCF-positive) is a beneficiary — but as high-beta growth it also leads any risk-off down.

Financials

Leading-indicator read (leading NET): composite Neutral (−0.2); revenue growth bearish (QoQ rate Δ −5.23pp, latest +4.11% vs prior +9.34%, YoY +33.5%); earnings strong, 100% beat rate; insider trading bearish (0 buys / 100 sells, net −$738.3M).

Runway / balance sheet: $932.2M cash and strongly positive operating cash flow (+$603M FY25, +$158M Q1 FY26) — GAAP losses are stock-comp-driven, not cash burn. No solvency question; the debate is growth durability and multiple, not the balance sheet.

Cross-references

• Prior deep-dive: the April 10 deep dive caught the spring SaaS-capitulation bottom ($170, RSI 34, death cross). This update marks the recovery to a golden-cross uptrend and the re-entry-on-pullback setup.
• Watchlists: focus (core), cybersec, web-cloud, ai-scan (also in cloud-etf-holdings / holdings-stocks / insider-universe / tech-insider-buys / bargain-bin).
• Software-valuation context: the June 29 Apollo Daily Spark flags software at its cheapest relative valuation in years — NET is FCF-positive software, a rotation beneficiary, but premium-multiple and high-beta (not the cheap-aggregate trade).
• No perspective opened off this deep dive; NET is a focus single-name, not a lane.

Sources

• Price truth (settled Fri 2026-06-26 close): summaries — $237.24, RSI 54.6, strong-up/pullback, SMA20 $238.86 / SMA50 $222.11 / SMA200 $207.80, golden cross, −14.3% off 52wk high, 30d +13.39%, 3m +16.86%, 1y +22.26%, 2y +186.42%, VWAP $204.42, AVWAP-from-low $210.61, POC $195.70, support shelf $220.28, relative volume 1.53.
• Financials: the desk's own tooling (SEC EDGAR XBRL, CIK 0001477333). Filings: 10-K 2026-04-29 (cloud-20251231); 10-Q 2026-05-08 (cloud-20260331); 8-K 2026-05-07.
• Leading indicators: leading NET (composite Neutral −0.2; revenue bearish/decelerating; earnings strong, 100% beat; insiders net −$738.3M).
• News (self-synthesized from the cached news NET corpus, fresh through 2026-06-12; the NVIDIA synthesis step was 429-throttled, so the raw articles were read directly): Q1 FY26 reported 2026-05-07 — $639.8M revenue +34% YoY, large-customer revenue +38%, FCF margin to 13% from 11%, EPS beat; a ~20% workforce reduction (~1,100 roles) announced as an AI-agent-era efficiency pivot; Piper Sandler upgrade to Overweight ($222 target, since exceeded) on accelerating AI revenue; FY guidance above estimates. Next report due ~late July 2026.