SMWB · Similarweb Ltd. · Stock research

Current Thesis

The 2026-08-12 print resolved the pivot question the July frame left open: Q2 revenue $77.2M (+9% YoY vs $71.0M), the first positive GAAP operating profit in company history ($0.7M, 1% margin), three seven-figure multi-year AI contracts worth ~$60M in total contract value, and FY2026 guidance raised to $314–318M from $307–315M. The eight sessions since answered a different question — who was on the other side of the spike. Form 4 filings dated 2026-08-17 report CEO Or Offer selling 709,597 ordinary shares on 2026-08-13 and 2026-08-14 for roughly $6,253,467, at weighted-average prices between $8.75 and $8.90, leaving 4,598,291 shares held directly (as aggregated by MarketBeat and TipRanks from the filings); as broker with sales beginning 2026-08-13. Price made a 52-week high of $10.40 and closed 2026-08-21 at $8.24, 20.8% below it. The narrative leg on offer is unchanged — a single-digit-growth web-analytics vendor repricing as a supplier of crawl-derived data to LLM builders — but the fast money in that leg has been paid, and nothing dated tests it until the Q3 print.

The narrative is maturing as of 2026-08-21. Dating it: the sell-side repricing ran from Oppenheimer's $8.50 (2026-08-05) through Northland's upgrade to Outperform at $10 (2026-08-12) to Citigroup $9 and Barclays $11 (both 2026-08-13), and no further analyst action has appeared in the eight sessions since; RSI(14) has cooled from 83.2 at the 2026-08-14 close of $8.89 to 64.5 at 2026-08-21; the 52-week high was made and not recovered. Well known, still working on the fundamentals, moderating flow.

Bullish and bearish views on Similarweb Ltd.

The model's bull view on Similarweb Ltd. (SMWB), in brief: Q2 2026 (2026-08-12): revenue $77.2M beat the $74.5–76.5M guide; non-GAAP diluted EPS $0.06 vs $0.03 consensus; non-GAAP operating profit $6.5M at an 8% margin against $2.4M and 3% in Q1 2026. The bear view: Insider supply arrived one day after the guidance raise. Both cases follow in full.

Bull Case

• Q2 2026 (2026-08-12): revenue $77.2M beat the $74.5–76.5M guide; non-GAAP diluted EPS $0.06 vs $0.03 consensus; non-GAAP operating profit $6.5M at an 8% margin against $2.4M and 3% in Q1 2026.
• Q3 2026 guidance of $80.5–82.5M sits above $78.8M consensus, an implied step up from the +9% just reported. FY non-GAAP operating profit guidance moved to $24–26M from $17–19M.
• RPO $345.3M at 2026-06-30, +26% YoY — contracted backlog compounding roughly three times faster than reported revenue, which is where multi-year licensing shows up before the income statement does.
• Retention: NRR 107% in the >$100K ARR cohort vs 103% in Q1 2026; 473 customers above $100K ARR (+9% YoY) at 69% of ARR; multi-year contracts 66% of ARR. ARR crossed $300M in June 2026.
• Product line extended into AI ad measurement on 2026-08-17 with a release covering ad placements across ChatGPT, Google AI Overviews and AI Mode — the same crawl panel sold as a second AI-native SKU. No revenue contribution has been disclosed for it.
• Every live sell-side target published after the print — $8.50, $9, $10, $11 — sits above the 2026-08-21 close of $8.24. In July the bull and neutral houses both stood at $7 (Citigroup, 2026-07-09).

Bear Case

• Insider supply arrived one day after the guidance raise.
• The 52-week high of $10.40 was made within days of the print and 20.8% of it was handed back by the 2026-08-21 close of $8.24. A high made and lost that fast is distribution, not accumulation, until price proves otherwise.
• Overall NRR is exactly 100%. The installed base outside the >$100K cohort is flat, and reported growth is still single-digit while the three-month price change reads +99.5%.
• The ~$60M is total contract value across multi-year terms. Annual recognition from those three deals is not broken out, so an outside modeller cannot separate recurring licensing from a lumpy signing quarter.
• Citigroup's raised target of $9 is a Neutral rating — after the raise, the neutral house sees fair value roughly 9% above the last close, which is not a re-rating cushion.
• Three contracts carried the quarter's AI headline. One LLM customer deferring or renegotiating would be visible in the AI-revenue percentage immediately.
• A CEO succession plan was disclosed on 2026-05-18 and remains unresolved; the executive associated with the licensing strategy has now sold into the strength the strategy produced.

Setup & Price Structure

• Reference close 2026-08-21: $8.24. 52-week high $10.40, 20.8% above. RSI(14) 64.5, down from 83.2 at the 2026-08-14 close of $8.89 — the overbought condition has cleared without a base having formed.
• The three-month price change reads +99.5% as of 2026-08-21 against +181.3% as of 2026-08-14. Most of that gap is the rolling window's start date advancing into the May–June rise rather than the price decline itself; both readings describe the same vertical move.
• $8.00 is the level with the most attached to it: below the lowest live target ($8.50, Oppenheimer 2026-08-05) and below the $8.75–$8.90 band in which the reported CEO sales printed. Price has not produced a weekly close under it since the print.
• Shares jumped roughly 17% in premarket trade on 2026-08-12 on the results, so the post-print gap is the structural feature of the chart; the question the next four to six weeks answer is whether the gap zone holds as support or fills.
• Crowding observables, stated as observables: four target revisions inside eight sessions and none since 2026-08-13; disclosed insider selling of ~$6.25M into the first two post-print sessions; RSI mid-range after an extreme reading; no earnings date inside 30 days to force positioning either way.

Catalyst Calendar (next 30 days)

• 2026-08-23 → 2026-09-22 — nothing company-scheduled. The IR press-release page as of 2026-08-23 shows the most recent item dated 2026-08-17 (AI ad-placement release) and no announced event, conference or earnings date in the window.
• 2026-08-13 → ~2026-11-11 — any executed sale generates a Form 4 within two business days, so further insider prints are observable in near-real time inside this window.
• ~2026-11-11 (est.) — Q3 2026 results. Outside the 30-day window and the first dated event capable of confirming or breaking the licensing line. Similarweb has not yet announced the date.

What Would Change Our Mind

The structure that breaks first is the post-print gap. Losing it would mean the 2026-08-12 guidance raise and the ~$60M of contract value bought roughly one week of price and nothing durable: specifically, a weekly close below $8.00 puts price under the lowest live sell-side target and under the band where the CEO's reported sales cleared, with no dated event before ~mid-November to reverse it. Three secondary conditions would confirm the break rather than start it: additional Form 4 sales executed into weakness under the open Form 144 window; a Q3 print with revenue beneath the $80.5M guide floor or AI-related revenue still stalled at 13% of sales; or a target cut back toward the July $7 level. The opposite case — a disclosed new seven-figure AI contract before the Q3 print, or a weekly close reclaiming $9.00 with the sell-side moving again — would restore the accelerating read and reset the level higher.

Correlation Notes

• The name trades with the AI-data-licensing complex rather than with analytics software peers: its bid is a function of LLM builders' willingness to pay for grounding and training data, so headlines on model-builder data budgets and publisher licensing deals move it before any company-specific news does.
• Second-order exposure to the same crawl-rights question that governs publisher litigation; the panel is the input to both the analytics product and the licensing line, so a forced sourcing change would hit both legs at once.
• Ordinary shares list on the NYSE and the company reports in USD, so there is no depositary structure or FX translation layer between the reported numbers and the quote. Israel domicile leaves standard geopolitical headline exposure.
• Small-cap enterprise software beta: the +99.5% three-month move means index or factor drawdowns in unprofitable-to-barely-profitable SaaS transmit with amplification, and the float has just absorbed ~$6.25M of disclosed insider supply.