Part of Lusha’s ongoing measurement of B2B buying signals. Every finding in one place: Five checks before you trust a buying signal.
All data in this report is drawn from Lusha’s live signal database, tracking 10 signal categories, funding events, IT spend change, hiring surge, headcount change, website traffic change, commercial activity, product activity, corporate strategy news, market intelligence and executive moves, across 25 named AI native companies between January 1 and June 22, 2026. This is a focused sample of named companies, not a population census. Source: Lusha live signal database, pulled June 22, 2026.
A single signal tells you something changed. Several firing on the same account in the same window tell you something larger is underway, and in this sample that pattern is unevenly spread. Five of the 25 companies fired all 10 categories. Three fired four or fewer. The rest sit between.
This report maps what stacking looks like in practice, what it does and does not predict, and one operational finding about company naming that affects anyone building a target list from signal data.
The numbers at a glance
Part one: what a full stack looks like
Five companies in this sample, Anthropic, ElevenLabs, Sierra, Harvey and Cursor, fired a signal in every one of the 10 categories inside the same six month window. Each of them showed, at the same time: a funding or investment event, rising IT spend, a hiring surge, headcount growth, rising website traffic, and commercial activity, product activity, corporate strategy, market intelligence and executive move news.
Read this as a map of detectable activity, not a ranking of companies. A lower position here means our signal layer picked up fewer kinds of public activity in this window, which can reflect company size, a quieter communications posture, or how a company is represented in the underlying sources. It says nothing about how the business is doing.
Part two: stacking measures breadth, not speed
It would be a cleaner story if the five full stack companies were also the five fastest growing. They are not, and forcing the numbers into a tidier shape would make the rest of this page worth less.
Group sizes are five, thirteen and four companies. None of these averages is stable at that size, and a single company moving would shift any of them by several points. Treat the direction as a hint and the individual figures below as the more informative numbers.
The full stack group and the 8 to 9 group land at effectively the same average, 13.8% against 14.4%, and with five companies in one group the gap is not worth interpreting. The visible drop is in the 5 to 7 group at 9.8%, and with four companies there, that is a hint rather than a finding.
The individual numbers are more useful than the group averages. The two fastest growers in the sample both sit in the 8 to 9 tier, at 26% and 22%. The 5 to 7 tier contains both the slowest grower in the sample, at 1%, and one growing at 24%. Within a single tier, companies move at completely different speeds.
The takeaway is not that stacking is meaningless. It is that stack depth counts how many kinds of activity were detected, and headcount growth measures how fast a company is adding people. A company can be visible across ten categories while growing steadily, and another can grow sharply while tripping five. Use stacking as a breadth indicator and hiring or headcount change as a speed indicator. They answer different questions, and a complete account view uses both.
Part three: a full stack account up close
Sierra is a useful example because every category fired inside the same window: a late stage funding round, rising IT spend, an active sales hiring surge, headcount growth, rising website traffic, and commercial activity, product activity, corporate strategy, market intelligence and executive move news, all within H1 2026.
No single one of those would necessarily move the account up a list. Together they describe a company where several functions are expanding at once, which is a different account to walk into than one where hiring is up and everything else is flat.
Part four: what this revealed about company naming
While building this report, one company returned under a name we had not searched for. We searched for Writer, the AI writing platform at writer.com, and the signal database returned results under Qordoba, the company’s original registered name before it rebranded. The data was not wrong. It was filed under the legal entity rather than the current public brand.
The practical takeaway. If you build a target list from brand names alone, you can miss or misattribute signal data for any company that has rebranded, spun out or been acquired since incorporation. Search by domain rather than by name, and check unfamiliar company names in your results before assuming a data error. It may be the legal entity behind a brand you already know.
This is also why the stack depth for that company should be read with care: a name based search returns what is filed under that name, and part of the record may sit under another.
Part five: what this means for account priority
1. Use stack depth to find accounts worth a closer look, not to rank them. A full stack account is broadly active and worth investigating. It is not automatically a better opportunity than an account with fewer categories firing and a sharper hiring surge.
2. Pair breadth with speed before deciding anything. The two measure different things. An account that is both broadly active and growing quickly is a stronger case than either measure alone.
3. Verify identity by domain before building a list. The Writer and Qordoba case shows how a name only search can miss data filed under a legal entity. Domain matching avoids it.
What this cannot tell you
Sample size. Twenty five companies, split into groups of five, thirteen, four and three. Group averages at that size are indicative at best, and we have shown the group sizes so you can weigh them accordingly.
Signal magnitude. Stack depth counts categories with at least one detected signal. It does not weight by size or frequency, so a company with one modest hiring surge scores the same in that category as one hiring at scale.
Detection, not activity. A category that did not fire means nothing was detected, which is not the same as nothing happening. Public communications posture and source coverage both affect what surfaces.
Causation. Nothing here shows that stacking causes growth or that growth causes stacking. Both are observations on the same 25 companies in the same six months.
Methodology and data notes
Data source: Lusha live signals API, pulled June 22, 2026.
Sample: 25 named AI native companies spanning foundation model labs, AI coding tools, AI customer service platforms and AI native infrastructure vendors. A sample study, not a population census.
Signal window: January 1 to June 22, 2026.
Signal categories tracked: financial events (funding, asset investment, strategic investment), IT spend increase, hiring surge, headcount increase, website traffic increase, commercial activity news, product activity news, corporate strategy news, market intelligence news and executive moves. Stack depth counts how many of these 10 categories had at least one detected signal for a company in the window. It does not weight by magnitude or frequency.
Data quality note: one company in the sample returned results under a different registered name due to legal entity naming in the underlying data. We identified and disclosed it rather than relabelling or dropping it.
Privacy and compliance: all figures are company level signal data drawn from public sources. No individual contact details are included or derivable from the figures presented. Current certifications are listed on the trust center.
Related resources
- Five checks before you trust a buying signal — what we learned running eight signals live
- Find accounts with stacked buying signals — the play that runs this on your own list
- Check buying signals on your target account list
- Find which signals are actually converting to pipeline
- How fast B2B contact data decays, and what it costs
