ZETA · Research
Zeta Global Grew 44 Percent but Stock Pay Still Eats Its Cash and Thins the Founder Majority
A fast growing marketing software platform with negative owner cash, rising share count, and a founder holding a thin 51.4% voting majority.
Spreadsheet & IC Deliverables: Download the formula-linked ZETA 5-tab Excel (.xlsx) workbook and printable 1-page tearsheet.
Download ZETA .xlsx →- Owner cash after capex, capitalized software and all stock pay
- Minus 0.8% of revenue
- David Steinberg total voting power
- 51.4%
- Class A shares added despite buybacks
- 6.9M
1. Our call: Hold off, because growth is real but owners still pay for it in stock
Our call: Hold off, conviction 3 of 5. Zeta Global Holdings Corp. is a marketing software company: its platform combines a large consumer data set, a customer data cloud and software that runs email, search, social, mobile, display and connected TV campaigns for enterprises [1]. It is a Delaware company based in New York, so this is a U.S. pick [1].
Revenue rose 44% in the second quarter of 2026 to $442.8M, and the company raised its 2026 outlook to $1,811M to $1,824M, or 24% to 25% growth excluding acquisitions and political candidate revenue [4]. That is a large, fast growing software name, and it posts results on Nov 9, 2026, after our seven-day window [8].
We chose it for the control setup. Co-founder, chairman and chief executive David Steinberg holds all of the Class B stock, which carries 10 votes a share, and 51.4% of total voting power, so the company is a controlled company under NYSE rules [3][1].
That majority is thin, and every share the company issues for deals or pay narrows it [2][3].
We measure owner cash as operating cash flow, less capital spending, less capitalized software, less all stock pay. For the six months to Jun 30, 2026 that was minus $6.3M, minus 0.8% of revenue, our calculation
2. A data and software platform with a media pass-through
| Metric | Value |
|---|---|
| Direct platform revenue | 73% of revenue in the six months to Jun 30, 2026; the rest is integrated platform revenue run through other ad platforms [2] |
| Super-scaled customers | 197 in the second quarter of 2026, up 17%, average revenue about $1.8M each [4] |
| Stock pay | $105.1M in six months, 12.5% of revenue, our calculation [2] |
| Customer concentration | Top ten customers were more than one-third of 2025 revenue, and one customer more than 10% [1] |
| API quality | 3 of 5, our scoring: the data cloud connects to partner platforms such as Snowflake, but most revenue comes from campaigns Zeta runs [1][4] |
| Toll-gate score | 0 of 4 pass, 3 partial, 1 fail (section 2b) |
2b. Zeta holds consumer data, but marketers can route around it
A gate is valuable when its owner holds a record others must check and prices what passes through. Zeta holds a large consumer identity and intent data set, and marketers use it to decide whom to reach, which is close to a gate but not a required one [1].
| Test | Result | Evidence |
|---|---|---|
| Owns the record | Partial | The 10-K calls its proprietary data a key competitive advantage, but marketers hold their own customer records and buy data from others [1]. |
| Owns the connection | Partial | The platform runs campaigns across email, search, social, display and connected TV, but 27% of revenue runs through other ad platforms [1][2]. |
| Prices close to outcomes | Partial | A substantial portion of revenue is usage based, tied to campaign volume rather than customer results [1]. |
| R&D on the gate vs features | Fail | Capitalized website and software development was $20.1M in 2025, 1.5% of revenue, while acquisitions took $90.3M of cash and $98.7M of stock, our calculation [1]. |
The verdict is no clean pass. Zeta is a strong operator of marketing programs, but the record it holds is one input among several for its customers, and it has bought growth through acquisitions such as LiveIntent and Marigold [1].
Usage-based revenue also swings more than subscription revenue, as the 10-K warns [1].
3. Growth is fast, and part of it is bought
| Period | Revenue | Growth |
|---|---|---|
| 2023 | $728.7M | Base year [1] |
| 2024 | $1,005.8M | Plus 38.0%, our calculation [1] |
| 2025 | $1,304.7M | Plus 29.7%, our calculation [1] |
| Six months to Jun 30, 2026 | $839.1M | Plus 46.5% against $572.9M, our calculation [2] |
| 2026 outlook | $1,811M to $1,824M | Plus 39% to 40%, or 24% to 25% excluding acquisitions and political revenue [4] |
Acquisitions explain a large share of the headline. Zeta bought LiveIntent in October 2024 for $277.0M and the Marigold enterprise business in November 2025 for $302.8M, paying in both cash and Class A stock [1][7].
International revenue rose to $77.1M in the first half of 2026 from $19.7M, mostly from Marigold, our reading [2].
Organic growth is still strong. Management guides to 23% to 24% growth in the third quarter excluding acquisitions and political candidate revenue, and super-scaled customer count grew 17% [4].
That growth is the main reason not to call this an Avoid [4].
4. The quarter on Nov 9
Third quarter guidance is revenue of $469M to $472M, 39% to 40% growth, and adjusted EBITDA of $115.0M to $116.0M, a margin of 24.4% to 24.7% [4]. The company has called its last 20 reports beat and raise quarters [4].
The quarter will show whether organic growth holds above 20% as the Marigold contribution laps in late November [4][7]. It will also show how many Class A shares were issued, which matters for control as section 11 explains [2][3].
5. GAAP profit has just turned positive
The adjusted EBITDA margin rose 170 basis points from a year earlier, and management raised the full-year adjusted EBITDA outlook to $404.1M to $406.3M, a margin of 22.1% to 22.4% [4]. Those are strong figures, but they leave out the largest cost owners bear [4].
Second quarter 2026 GAAP net income was $8M, which the release put in its headline, against a net loss of $31.5M for all of 2025 [4][1]. Adjusted EBITDA was $92M in the quarter, a 20.7% margin [4].
The gap between the two is mostly stock pay and depreciation. In the first half of 2026 stock pay was $105.1M and depreciation and amortization $46.2M, against a net loss of $5.1M [2].
Changes in the fair value of acquisition-related liabilities added another $36.7M of cost in 2025 [1].
Full-year 2026 GAAP earnings guidance was raised, but it remains close to break-even [4]. The quality of profit depends on stock pay falling as a share of revenue, which it has: 33.3% in 2023, 13.6% in 2025, our calculation [1]
6. Owner cash is still slightly negative
Owner cash is our calculation from the cash flow statements [1][2]. The trend is good: from minus 25.8% of revenue in 2023 to minus 1.0% in 2025 and minus 0.8% in the first half of 2026, our calculation [1][2].
But stock pay of $105.1M in six months still consumes nearly all of operating cash of $118.9M [2]. The company's free cash flow guidance of $254.8M to $255.8M for 2026 treats that stock pay as free [4].
Shareholders see the cost in the share count. The company spent $55.6M buying back stock in the first half of 2026 and $121.0M in 2025, yet Class A shares outstanding rose to 227,375,179 on Jul 29, 2026 from 220,485,174 on Feb 13, 2026 [2][1].
7. Capital: a new billion dollar facility
On Jul 24, 2026 Zeta signed a five-year $1.0B senior secured credit facility with Bank of America as agent: a $250.0M term A loan and a $750.0M revolver [5]. It repaid the $200.0M owed under the August 2024 facility [5].
Loans bear SOFR plus 1.75% to 2.50% depending on leverage, and the net leverage cap is 3.25 to 1, rising to 3.75 to 1 for four quarters after any acquisition of at least $100M [5]. That step-up is written for more deals [5].
Cash was $310.0M at Jun 30, 2026 against $319.8M at the end of 2025 [2]. The 2025 repurchase program allows up to $200M through Dec 31, 2027, including shares withheld for employee taxes [2].
8. The main risks are dilution, concentration and usage swings
Dilution. Shares issued in connection with acquisitions and other agreements were $98.7M in 2025 and $77.1M in the first half of 2026 [1][2].
Stock pay adds more each quarter [2].
Concentration. The top ten customers were more than one-third of 2025 revenue, and the 10-K warns that agency holding companies could pull several accounts at once [1].
Usage. A substantial portion of revenue is usage based, which the company says is less stable than subscription pricing [1].
Political candidate revenue also inflates even years [4].
Data rules. Zeta depends on consumer data, so privacy laws and platform changes can limit what it can collect and use, the 10-K warns [1].
About 2,000 of its 3,300 employees are outside the U.S. [1].
9. Management wants to be the enterprise decision layer
Mr. Steinberg describes the data cloud and its Athena intelligence layer as being at the center of enterprise decision making, and cites new partnerships with Snowflake and Palantir [4]. Chief financial officer Christopher Greiner, in the role since 2020, credits adoption across the platform [4][3].
Steven Gerber has been president since 2009 and runs product, business development, customer success and operations [3]. The company was founded in 2007 by Mr. Steinberg and John Sculley, who retired from the board on Jun 9, 2025 [4][3].
10. Superintelligence is the pitch, and stock pay is the bill
Zeta now calls itself an infrastructure company for intelligent enterprise software and introduced a framework to measure adoption of its superintelligence tools in the second quarter [4]. The product is real, but most revenue is still campaign usage [1][4].
The 10-K says it has over 500 data scientists and engineers [1]. Paying them largely in stock is why owner cash is negative while reported free cash flow is positive, which supports the Fail on research spend in section 2b [1][2].
11. A founder majority that buybacks defend
Vote control. Class B stock carries 10 votes a share, Class A one, and they vote together [3].
At Apr 17, 2026 there were 225,609,644 Class A and 23,638,426 Class B shares, and Mr. Steinberg held all of the Class B through entities including ACI Investment Partners, at 19.1% of the vote, and ACI Investment Company XXVII, at 20.7% [3].
His total voting power was 51.4%, against an economic stake of about 10% [3]. Directors and officers as a group held 51.8% [3].
The company relies on controlled company exemptions from some NYSE governance rules [1].
The Sunset Date. When Class B holders no longer hold a majority of voting power, directors can be removed only for cause [3].
On the Jul 29, 2026 count, the Class B votes alone were about 51.0% of the total, and roughly 9M more Class A shares would end that majority, our calculation [2][3]. Every acquisition paid in stock moves the date closer, and every buyback pushes it back
12. The questions we would ask management, and the calls we are tracking
Question one: when will stock pay fall below 10% of revenue? [2]
Question two: will the board stop paying for deals in Class A stock as the founder's majority narrows? [2][3]
Question three: how much of 2026 growth comes from Marigold and political revenue? [4]
Question four: who is the customer above 10% of revenue? [1]
Question five: what happens to board governance after the Sunset Date? That answer sets our conviction.
Prediction one: third quarter 2026 revenue is at or above $469M (80%). [4]
Prediction two: 2026 owner cash as we define it is positive (45%). [2]
Prediction three: Class A shares outstanding on the next 10-K cover are above 227.4M (70%).
Bottom line: a fast grower whose owners still fund it in shares
Zeta is growing above 20% organically and has turned GAAP profitable, but stock pay still absorbs nearly all operating cash and the share count keeps rising. Its founder controls a thin voting majority that buybacks protect.
We would turn positive if owner cash turns clearly positive and the Class A count stops rising.
Stock Profiles & Comparable Companies
ZETA Stock Profile →More Published Research
View All →Mobility Global Owns the CARFAX Record and the Debt That Paid for the Spin
CARFAX grew 8% and no customer tops 10% of revenue, but the spin left $2.0B of notes and S&P Global's finance chief on the board.
Cerence Looks Fixed but a Samsung Patent Check Did Most of the Work
An in-car voice software supplier whose nine month growth falls to about 4% without one settlement, with no activist, a slipping pay vote and an ex-Intel board tie.