First published

Similarweb’s AI deals are growing. How much cash is left?

Explore this signal

Long Term at a glance

Short reader summary

Similarweb says AI-related customers accounted for roughly two-thirds of newly added annual recurring revenue in its latest quarter. That is not two-thirds of total sales. The question is how much cash those contracts eventually leave after costs.

Exact asset or item

Similarweb Ltd. ordinary shares

Asset type

Listed investment

Reference price

8.84

Currency code

USD

Price observed at

2026-11-17T16:00:00-05:00

What this number represents

Tracking price

What remains unproven

Historical tracking reference only. No executable entry after publication, position, sale, profit or price target. Both cash measures are non-GAAP. The higher measure adds back real cash payments.

Exact asset or item

Exact asset or item

Similarweb Ltd. ordinary shares

Asset type

Listed investment

Ticker

SMWB

Exchange

NYSE

How to identify this exact asset

Similarweb Ltd. is an Israeli company with ordinary shares listed on the New York Stock Exchange (SMWB). US dollars. This is a conditional research watch, not an instruction to buy.

Price and what it means

Entry or reference price

8.84

Currency code

USD

Price observed at

2026-11-17T16:00:00-05:00

What this number represents

Tracking price

Price evidence status

Indicative

Price source

Similarweb rises 13.04% in Tuesday’s session

Publisher: Global Banking & Finance Review

Original language: English

Reading access: partial

Underlying origin: SMWB-MARKET

published date precision: day

retrieved date precision: day

Source publication date 2026-11-17

Source retrieval date 2026-11-18

Matching price source ID

S8

Purchase, sale and realised profit

None established. No purchase, sale, dividend or broker fill was recorded. The US$8.84 market observation predates the original issue.

The developments we connected

Separate development

  1. What changed

    Contracts signed, not cash collected

    Date of the development

    Evidence status

    Observed

    What the evidence shows

    Q2 revenue was US$77.2m; three multi-year deals exceeded US$60m combined. That contract value is not revenue already earned.

    Source IDs

    • S2
  2. What changed

    Retention and cash tested

    Date of the development

    Evidence status

    Observed

    What the evidence shows

    Q3 revenue reached US$83.0m; net revenue retention was 104%. Free cash flow was US$4.509m before the company’s normalization.

    Source IDs

    • S1
  3. What changed

    Historical market reaction

    Date of the development

    Evidence status

    Observed

    What the evidence shows

    The regular NYSE session closed at US$8.84, up 13.04%; not a current offer.

    Source IDs

    • S8

Emerging insight

Similarweb measures online traffic and searches. A company building an AI tool might pay for those measurements rather than collect its own data. Signing a contract is only the first step: Similarweb must deliver the service, earn revenue and collect payment. The latest quarter gives readers two very different cash figures to track.

How this exact asset could benefit

If customers renew and spend more, selling additional data to them may cost less than winning new customers. More earned revenue could leave cash for the business after staff, technology and other bills. That does not establish a dividend or a gain on anyone’s shares.

Who could benefit

  1. Who

    Similarweb

    How they could benefit

    Renewals and collections could leave more cash to reinvest after costs.

    Evidence status

    Still unproven

  2. Who

    Shareholders

    How they could benefit

    Lasting growth in cash per share could support shareholder value, but market prices and investment returns may diverge.

    Evidence status

    Still unproven

What we observed

For Q3 2026 Similarweb reported US$83.0m revenue, up 16% year on year, and 104% net revenue retention versus 98% a year earlier. This measures customer spending retained, not the percentage of customers who stayed. Multi-year contracts accounted for 66% of total annual recurring revenue. Separately, AI-related customers contributed roughly two-thirds of new annual recurring revenue. These percentages measure different things.

What changed in one year?
View comparison values
What changed in one year?
MeasureQ3 2025Q3 2026
Spending retained from existing customers98%104%
Cash from operations (US$m)3.4504.927
Free cash flow before add-backs (US$m)2.9504.509
104% describes spending, not the share of customers who renewed. Neither cash measure is a bank balance.

What remains unproven

The company reported US$4.509m free cash flow and US$10.544m normalized free cash flow for Q3. Both are non-GAAP measures; neither is a bank balance. The US$6.035m difference reflects adding back cash payments for a legal settlement and an acquisition. It is not additional cash collected. Nine-month GAAP net loss was US$6.487m. AI-partner cash receipts and lasting incremental profit are not disclosed.

How $4.509m becomes $10.544m
How $4.509m becomes $10.544m
Quarter ended 30 September 2026US$ million
Free cash flow before add-backs4.509
Add back legal settlement payment+4.000
Add back deferred acquisition payments+2.035
Company’s normalized free cash flow10.544
The $6.035m difference adds payments back on paper; the money did leave the company. Neither free-cash-flow measure is a GAAP subtotal.

What to check before acting

Entry conditions

Entry price unavailable. US$8.84 was the historical 17 November 2026 regular-session close, not a price available to a reader after publication. A current broker quote, eligibility and actual fees would be needed before assessing any transaction.

Possible upside

Customers could renew larger contracts while the company keeps more cash per share. Check multiple quarters: one earnings release or a price rise cannot establish that result.

Possible downside

Subscriptions could grow while staff and technology absorb the gains. Customers could gather data themselves, reduce purchases or gain bargaining power. Weak cash conversion or share dilution could leave investors worse off.

Price or return range method

No probability, fair value or price target is supported. The dated tests examine customer retention, cash and dilution, not a promised future market price.

Costs and fees

Consider commissions, bid–ask spread, currency conversion, taxes and dilution. Purely illustrative: 100 shares at US$8.84 cost US$884 before fees. With US$2 each to buy and sell plus US$5 currency costs, break-even is US$8.93 per share before tax. A hypothetical US$10 sale would leave US$107 before tax, not an observed return.

How hard it may be to sell

NYSE listing does not guarantee a particular price or fill. No broker spread, live executable quote, Singapore eligibility, purchase or sale was verified.

Key risks

Platforms may restrict access to data; large customers may demand better prices. Mid-tier customer counts may weaken even as larger customers spend more. Retention and cash can reverse. Currency moves, valuation and share dilution matter; the shares could lose their entire value.

What an ordinary reader can do next

Compare the company’s cash reconciliation with the headline. Read the next two quarterly filings: are customers spending more in total, is unadjusted free cash flow positive, and are shares being issued fast enough to dilute those gains? Mark absent disclosures as unknown.

Original forecast record

Original forecast in full

If Similarweb’s larger AI-related and multi-year data contracts continue to lift customer retention and actual unadjusted free cash flow, the business may gain lasting economic value by 2028-11-18. This fails if contracts grow but real cash and customer retention do not improve on sustained reporting.

Original forecast date and time

2026-11-18T12:24:00+08:00

Forecasts and review dates

6 months

Review date

2027-05-18

Original forecast

Compare the latest retention and both cash measures; check if planned AI-shopping measurement actually launched.

What to measure

Compare the latest retention and both cash measures; check if planned AI-shopping measurement actually launched.

What must happen

Later issuer filings must document the change on a comparable basis. A signed contract or a share-price move is not evidence of cash received.

What would challenge the forecast

Profit turns negative or unadjusted free cash flow falls despite contracts.

Current outcome

Not yet due

1 year

Review date

2027-11-18

Original forecast

Look for two more quarters above 100% overall retention and positive unadjusted free cash flow.

What to measure

Look for two more quarters above 100% overall retention and positive unadjusted free cash flow.

What must happen

Later issuer filings must document the change on a comparable basis. A signed contract or a share-price move is not evidence of cash received.

What would challenge the forecast

Signed deals fail to renew or cash repeatedly trails the adjusted headline.

Current outcome

Not yet due

2 years

Review date

2028-11-18

Original forecast

Compare large-customer count, remaining contracted work, cumulative unadjusted free cash flow and share count.

What to measure

Compare large-customer count, remaining contracted work, cumulative unadjusted free cash flow and share count.

What must happen

Later issuer filings must document the change on a comparable basis. A signed contract or a share-price move is not evidence of cash received.

What would challenge the forecast

Revenue rises but operating cash or cash per share stagnates as share count grows.

Current outcome

Not yet due

3 years

Review date

2029-11-18

Original forecast

Check whether service revenue and unadjusted free cash flow rose for multiple years with stable renewals.

What to measure

Check whether service revenue and unadjusted free cash flow rose for multiple years with stable renewals.

What must happen

Later issuer filings must document the change on a comparable basis. A signed contract or a share-price move is not evidence of cash received.

What would challenge the forecast

Competition or platform restrictions cause repeated material declines in retention.

Current outcome

Not yet due

4 years

Review date

2030-11-18

Original forecast

Check AI customers’ paid business and evidence that Similarweb can sustain prices.

What to measure

Check AI customers’ paid business and evidence that Similarweb can sustain prices.

What must happen

Later issuer filings must document the change on a comparable basis. A signed contract or a share-price move is not evidence of cash received.

What would challenge the forecast

Larger contract announcements coincide with flat or negative annual operating cash.

Current outcome

Not yet due

5 years

Review date

2031-11-18

Original forecast

Compare cumulative cash, revenue quality and after-cost price observations with the frozen claim.

What to measure

Compare cumulative cash, revenue quality and after-cost price observations with the frozen claim.

What must happen

Later issuer filings must document the change on a comparable basis. A signed contract or a share-price move is not evidence of cash received.

What would challenge the forecast

No durable improvement in cash conversion and retention, even if the share price rose.

Current outcome

Not yet due

What would make the thesis fail?
  1. 18 May 2027Check retention and cash without add-backs in the latest filing.
  2. 18 Nov 2027Look for two more quarters with retention above 100% and positive free cash flow before add-backs.
  3. 18 Nov 2028Compare cumulative cash generation and cash per share against revenue growth and dilution.
  4. 18 Nov 2029Test whether revenue and cash rose over several years with stable renewals.
  5. 18 Nov 2030Do paying AI customers and pricing power show up in operating cash?
  6. 18 Nov 2031Compare actual outcomes and price after costs against the frozen 2026 claim.
These are OTR review dates, not company guidance, forecast returns or scheduled earnings releases. Source: OnTheRice’s original 18 November 2026, 12:24 SGT dossier, pp. 6–8.

Sources

Dated source

  1. Source ID

    S1

    What this source supports

    Revenue US$83.0m, GAAP profit, retention, cash-flow bridge, 2026 guidance. Company-reported unaudited quarter; normalized cash adds back items that cost real cash.

  2. Source ID

    S2

    What this source supports

    Q2 revenue US$77.2m, GAAP operating profit US$0.7m and US$60m of disclosed multi-year contract value. Contract value does not equal cash collected.

  3. Source ID

    S3

    What this source supports

    2025 revenue US$282.6m, net retention 98%, full-year GAAP loss. Prior baseline, not a forward guarantee. Issuer release: February 2026, exact day not checked in dossier. Publisher's dated footer subsequently verified: 2026-02-17.

  4. Source ID

    S4

    What this source supports

    Risk of source/platform dependence, historical losses and historic free cash flow; management estimates of market opportunity are unverified.

  5. Source ID

    S6

    What this source supports

    Initial AI-shopping measurement solution described as planned for Q4 2026, not evidence of delivered profitable product or paid adoption. September 2026, day not checked in dossier. Publisher's dated footer subsequently verified: 2026-09-02.

  6. Source ID

    S7

    What this source supports

    Notes falling mid-tier customer count and difference between actual/normalized cash. The numerical base overlaps the SEC report, so this is not independent confirmation of the issuer figures.

  7. Source ID

    S8

    What this source supports

    Dated automated market-data report, 17 November 2026. Reports last traded price US$8.84, not exchange-certified official close. S5 separately reports the same regular-session close.

Source limits

The quarter is unaudited. Multiple articles repeating issuer figures are not independent confirmations. A planned AI-shopping measurement service is not paid adoption. The share-price sources record historical market observations, not executable orders. Dates with no verified day or publication date retain their original uncertainty. The historical 17 November close was separately checked against the dated price row on Stock Analysis (https://stockanalysis.com/stocks/smwb/history/). The rolling page has no verified publication date; its observation date is not its publication date.

Back to Signals

AI powered research