Zeta Global: fast growth, but the stock pay takes most of the cash

In the second quarter Zeta Global reported GAAP net income of $8.2 million, adjusted EBITDA of $91.7 million and stock-based pay of $52.1 million. Which of those three numbers you trust decides whether the shares look cheap, fair or expensive. Management guides to free cash flow of $255 million for 2026, and after the roughly $215 million I expect to be paid to employees in shares, the owners are left with about $40 million.

At $33.09 on October 9, 2026 the shares trade 23.7% above my triangulated Base value of $26.74 and 28% below my Bull value of $46.04. The Bear value of $9.72 is 70.6% below the price. The required margin of safety of 50% puts the maximum buy price at $13.37, and the price is about 2.5 times that level. For existing holders this is a hold, with no reason to add. For new money it is a watchlist name. The shares traded around $33.57 on October 12, up 1.5% on a broker quote taken during the session, which had not closed when I wrote this.

Route: SF-07 SaaS and technology growth, adapted: Zeta is an AI marketing platform with a data cloud, not an annual recurring revenue (ARR) company, so the ARR sanity check is only indicative. Economic classification: Quality Growth, 6/12 on the sector scorecard, so the engine is GROWTH_ADJUSTED_MULTIPLE. Primary method: enterprise value (EV) to owner EBITDA (adjusted EBITDA less stock pay) for 2028. Secondary method: a mature-year (2031) owner free cash flow model with an EV to free cash flow multiple. Values are present fair-value estimates, not 12-month price targets. Owner cash flow here is free cash flow minus stock-based compensation.

What Zeta is

Zeta sells a marketing platform for large consumer brands. It combines a database of consumer identities and signals (its data cloud) with software to plan, run and measure campaigns across email, mobile, advertising and web, and in 2026 it added Athena, an AI assistant that customers talk to. Revenue was $1.3 billion in 2025 (my sum of the four reported quarters) and is guided to about $1.82 billion for 2026, up 39%, or 25% excluding acquisitions and political candidate revenue.

Growth is part organic and part bought. The company took over the enterprise business of Marigold under an agreement dated September 27, 2025, which contributed $48.1 million of second-quarter revenue, and it also reports political candidate revenue that is not recurring. Excluding Marigold's enterprise business and political revenue, second-quarter revenue was $395 million, which is about 28% above last year's $308.4 million (my arithmetic from the release). The customer base is concentrated at the top: 197 super-scaled customers (those spending at least a set annual amount, a company definition) averaged $1.8 million of quarterly revenue each, both up 17%. One unnamed customer was more than 10% of revenue in the half. Direct platform revenue was 72% of the total in the quarter and revenue integrated into other companies' platforms 28%.

The latest quarter

The quarter beat guidance by about 5% and the company raised its outlook again. 2026 guidance is revenue of $1,811 million to $1,824 million (the midpoint up $33 million), adjusted EBITDA of $404.1 million to $406.3 million, free cash flow of $254.8 million to $255.8 million (up $20 million at the midpoint) and GAAP EPS of $0.09 to $0.11, excluding a possible one-time tax benefit. Third-quarter guidance is revenue of $469 million to $472 million, up about 40% (23% excluding acquisitions and political revenue), and adjusted EBITDA of $115 million to $116 million. Management said the guidance keeps a cushion of 2% to 5% and assumes little new partnership revenue. This is the 20th consecutive beat and raise by the company's own count.

On the call management said Athena has been available for about 130 days, more than 40% of super-scaled customers use it monthly, the top 20% of customers by AI adoption generate about 70% of revenue, multi-use-case customers rose 90% and the sales pipeline grew more than 60%. The data cloud was integrated with Palantir Foundry as of July 31. The next report is due around November 3, 2026 (an estimated date).

The hard metric: free cash flow after stock pay

Zeta's own free cash flow is operating cash flow less capital spending and capitalized software. It is a good number: $185.1 million in 2025 and $224.4 million over the last twelve months on the simpler operating-cash-flow-minus-capex definition of a data provider, up each year since 2021. The headline figure used by the market is adjusted EBITDA of about $405 million. The hard figure is lower still, because stock-based pay was $194.5 million over the last twelve months and $105.1 million in the first half of 2026, about 12% of revenue. Paying staff in shares is a real cost to the owners: each share issued dilutes everyone else, and the company has to buy shares back with cash to offset it.

Hard metric: owner free cash flow of about $40 million, which is 16% of the company's free cash flow guide and a 0.5% yield on a market value of $8.7 billion. On the same basis adjusted EBITDA less stock pay (what I call owner EBITDA) was $39.6 million in the second quarter and $52.7 million in the half, 8.9% and 6.3% of revenue (my arithmetic). Most of the valuation therefore rests on how fast that 2.2% margin can rise. I anchor value on the hard figure, and I label the full-year stock pay ESTIMATED.

Cash, debt and shares

Cash was $310.0 million and long-term borrowings $197.5 million at June 30, so net cash is about $112.5 million. I also subtract $50.9 million of acquisition-related liabilities (earnout-type payments for deals already done), which leaves net cash of about $62 million. The company closed a new $1 billion credit facility that includes a $250 million term loan and a $750 million undrawn revolver, described as available for acquisitions and share repurchases. I did not read the facility terms and do not know whether the term loan was drawn after June 30, so I test a case where $250 million was spent.

Class A and Class B shares total 251.0 million (227.4 million and 23.6 million) as of July 29, 2026. Diluted weighted shares in the second quarter were 256.6 million against 245.3 million basic, so dilutive securities add about 11 million, which puts my diluted count at about 262 million. The count is up about 13% in a year (weighted basic shares 245.3 million against 217.3 million in the second quarter of 2025), partly because shares were issued for acquisitions, $77.1 million in the first half. Buybacks were $55.6 million in the half (3.2 million shares at an average of about $17.40), and about $89.4 million of the authorization remains. The shares bought back since the start of the year were bought at roughly half of today's price. In the model I assume buybacks offset stock pay dilution in the Base case, and a 2% annual increase in the Bear.

Primary valuation: EV to owner EBITDA

I value Zeta on 2028 owner EBITDA (adjusted EBITDA less stock pay) at a multiple, because 2026 owner earnings are too small to anchor a price and growth is the point. The SF-07 tiers use EV to free cash flow of 20x to 30x for mature software with growth below 15%, 30x to 50x for growth of 20% to 40% and a Rule of 40 above 50, and 50x to 80x for hypergrowth. Zeta's Rule of 40 (organic growth plus free cash flow margin) is about 39 (25% plus 14%), and about 27 on owner free cash flow (25% plus 2.2%), below 30 and so red on the owner basis. I use 14x, 20x and 26x owner EBITDA, which is lower than the growth tier because owner margins are low and a fifth of that is acquired revenue. Gross margin of 59% is also below the 70% the framework treats as standard for software, which is a reason for the discount.

The primary Base value is $28.39. In the Base, enterprise value at the end of 2027 is $8.17 billion (2028 owner EBITDA times 20), plus net cash of $62 million, divided by 262 million shares gives $31.41, discounted at 10% over 1.22 years to $27.96, plus $0.43 of owner cash flow in the meantime. The Base margin of 16.0% is adjusted EBITDA of about 25.5% less stock pay of about 9.5%, a path from 22.3% and about 11.8% today. Revenue growth of 20% in 2027 and 17% in 2028 is my estimate: Marigold annualizes in 2027, organic growth is guided near 25% for 2026, and I let it fade. The Bear assumes growth of 12% and 10%, adjusted EBITDA of 22.5% and stock pay of 11%. The Bull assumes 25% and 22% growth, adjusted EBITDA of 27.5% and stock pay of 8.5%.

Sensitivity: primary value per share

Secondary valuation: a mature year, 2031

The secondary method values a more mature Zeta in 2031, when growth has faded to about 10% to 12% (the framework's mature software tier of 20x to 30x EV to free cash flow). Revenue grows 17%, 15%, 12% and 10% after 2028 in the Base, reaching $3.6 billion. The owner free cash flow margin rises to 11.5%, which is adjusted EBITDA of 28% converting to free cash flow at about 68% (19%), less stock pay of about 7.5%. That is $416 million. At 25x, plus net cash, divided by 262 million shares, it is $39.92 in 2031, or $24.27 today at 10% a year.

The secondary Base value is $24.27. Almost all of it is terminal value, because the model values the business in 2031 and discounts it back, so it is a sensitive estimate. The Bear case also assumes the share count rises 2% a year (no buyback capacity, more stock pay), which makes it 290 million shares by 2031.

Sensitivity: secondary value per share

Growth diagnostics

The EV to revenue sanity check says the shares are not obviously expensive: EV of $8.6 billion is 4.7 times 2026 guided revenue and 3.9 times my Base 2027 revenue, below the 5x to 12x band that the framework calls fair for subscription software. That check is weak here, because Zeta is not an ARR business, its gross margin is 59%, and about 12% of revenue goes to employees in shares. On adjusted EBITDA the shares trade at 21 times 2026 guidance, and on owner EBITDA at about 45 times (EV of $8.6 billion over about $190 million for 2026, my estimate).

Per-share growth quality (SM-19 and SM-20) is a diagnostic only. GAAP EPS was a loss of $0.01 over the last twelve months, so a price-to-earnings ratio and a PEG are not meaningful, and the forward ratio of 31 on the quote source is on an adjusted basis that I could not verify. Free cash flow on the company's definition grew from $34.8 million in 2021 to $185.1 million in 2025 (data provider basis), but over the same years stock pay was $259.2 million, $299.0 million, $242.9 million, $195.0 million and $177.8 million, so owner free cash flow was negative until about 2025, and the improvement is real but recent. The share count rose about 13% over the last year. Growth in cash flow per share is therefore much lower than growth in cash flow, and the acquisition effect is large (about a third of 2026 revenue growth is acquired, and political revenue is not recurring). Forecast reliability is MEDIUM: guidance is one year and built with a 2% to 5% cushion, which supports near-term numbers and not the 2028 and 2031 figures. I did not rebuild a history of the company's own valuation multiple, so the historical cross-check is UNAVAILABLE.

Valuation triangulation

The two methods diverge by 15.6% of their midpoint, below the framework's 20% warning line, though the methods share the same revenue path and the same stock pay assumptions, so they are not independent. The primary gets 60% because it relies on a nearer year. Weighting them 50/50 gives $26.33, and the primary alone $28.39.

What is the market already pricing in?

At $33.09 market value is $8.7 billion and enterprise value about $8.6 billion. Holding my Base revenue path, the price needs a multiple of 23.4 times 2028 owner EBITDA, against the 20 times I use, or 34.1 times 2031 owner free cash flow, against my 25 times. The price therefore asks for either higher margins than my Base (owner EBITDA of about $480 million in 2028 at 20x, instead of $408 million) or a richer multiple for a company still growing 17% in 2028. With stock pay at 9.5% of revenue in 2028 in my Base, the owner margin gain needed is about 3 points.

For the Base case an investor needs organic growth to stay near 20% for two more years, adjusted EBITDA margin to rise from 22% to about 25.5%, and stock pay to fall from about 12% to 9.5% of revenue. Evidence so far: guidance has been raised 20 times, adjusted EBITDA margin rose 170 basis points in the quarter, and customers and ARPU grew 17%. Against that, gross margin fell about 3 points in a year, the stock pay ratio has barely moved (12.5% in the half), and the quarterly numbers include an acquisition and political revenue that will not repeat.

Bear, base and bull scenarios

Bear: growth slows to about 10% once Marigold is absorbed and political revenue disappears, the largest customer or the data supply weakens, stock pay stays near 11% of revenue and the share count rises 2% a year, so owner earnings stay thin and the market pays 14 times for it. Base: organic growth of about 20% fades to 10% by 2031, the adjusted EBITDA margin reaches 25.5% by 2028, stock pay falls to 9.5% of revenue and buybacks offset dilution. Bull: Athena and the data cloud push growth to 25% and 22%, margins reach 27.5% and stock pay falls to 8.5%, and the market pays 26 times owner EBITDA. The Bear is further from the Base than the Bull is, because owner earnings are small enough that a few points of stock pay or margin swing the value by half.

The probability-weighted value, at 30%, 50% and 20% (I weight the Bear at 30% because stock pay and customer concentration make it easy to reach), is $25.49, which is 4.7% below the Base and 23.0% below the price.

How the buy price is set

Confidence C starts the margin of safety at 40%, the middle of the framework's 35% to 50% range. I add 5 points for customer concentration (one customer above 10% of revenue, and the top fifth of customers by AI adoption generating about 70% of revenue) and 5 for terminal value and exit multiple dependence (both methods rest on a multiple of a future year, and the secondary is almost all terminal value). I add nothing for leverage (net cash) or accounting opacity beyond what the owner metrics already remove. That leaves 50%.

A 50% discount to the $26.74 Base gives a maximum buy price of $13.37. At the bottom of the Confidence C range (35%) the price would be $17.38, and with no add-ons (40%) $16.04. Even the lightest of those is 47% below the price. The 52-week low of $14.37 is just above my buy zone, which is a useful reminder of how far the shares have moved: the shares are up about 79% in a year, and near the top of a 52-week range of $14.37 to $34.80.

Reasons to own ZETA

• Consistent beat and raise. Revenue beat the guidance midpoint by about 5%, and full-year revenue, adjusted EBITDA and free cash flow guidance were all raised again, the 20th consecutive time by the company's count.
• Organic growth near 25%. Excluding acquisitions and political revenue, growth is about 23% to 28%, with super-scaled customers and their spend each up 17%.
• Operating leverage. Adjusted EBITDA margin rose 170 basis points to 20.7%, and GAAP income from operations turned positive at $16.9 million in the quarter.
• Net cash and a large credit line. About $62 million of net cash after acquisition liabilities and a new $1 billion facility, $750 million of it undrawn.
• Visible AI traction. More than 40% of super-scaled customers use Athena monthly, and multi-use-case customers rose 90%.

What could go wrong?

• Stock pay. About 12% of revenue ($105.1 million in the half) is paid in shares. It exceeds GAAP net income many times over and consumes about 84% of company free cash flow.
• Bought growth. Marigold and earlier deals fuel the 39% headline growth, paid partly in stock ($77.1 million in the half) and with earnout liabilities, and political candidate revenue ($7 million in the third quarter and $8 million in the fourth) does not recur.
• Customer concentration. One customer is above 10% of revenue and of receivables, and the filing lists dependence on adding and retaining large enterprise customers as a risk.
• Data and platform rules. The filing lists restrictions on online data collection from browser and phone changes, privacy and AI regulation, reliance on third-party AI models and data centers, and email standards set by inbox providers.
• Margin trend and valuation. Gross margin fell about 3 points in a year to 59%, and at $33.09 the shares are 5% below their 52-week high after rising 9% on a single day in late September, so a miss would be costly.

Management execution

Management has delivered on every quarterly outlook for 20 quarters, grew adjusted EBITDA 56% in the quarter, turned GAAP income positive and bought back shares each quarter at prices well below today's. Against that, stock pay has stayed near 12% of revenue, the share count rose 13% in a year with acquisitions paid partly in stock, and the headline growth rate leans on acquisitions. I could not verify insider transactions. Management credibility scores 13/20: strong delivery on guidance, mixed on the cost of the growth.

Stock Analyza scorecard

The moat score is average: the data cloud and customer integrations create switching costs, but I have no retention or net revenue retention data (UNAVAILABLE), and the platform competes with much larger marketing software groups. Economic return is low because GAAP operating margin is under 4% and $710 million of the balance sheet is goodwill and intangibles. Cash flow quality is weak because stock pay takes most of the free cash flow. Dilution and governance gets 2 for a share count up 13% in a year and a dual-class structure (23.6 million Class B shares).

The valuation score is 24/100: discount to Base 2/40, protection against the Bear 1/20, agreement between methods 11/15, historical valuation 4/10, market implied expectations 3/10, data and model penalty 3/5. The score is low because the price is 24% above the Base and the Bear is 71% below it. It does not say the business is poor.

Confidence score

What I would watch from here

The next review point is the third-quarter 2026 report, expected around November 3, 2026. Two dashboard lines are red today (stock pay and free cash flow after stock pay), which is the reason for the 50% margin of safety.

The three most important thesis breakers

1. Organic growth falls below 15% once acquisitions and political revenue are removed.
2. Stock pay stays above 12% of revenue while the share count keeps rising, so owner earnings do not grow.
3. The largest customer (above 10% of revenue) reduces spending, or data collection rules cut the data cloud's reach.

Adversarial review

I rebuilt the key figures from Zeta's second-quarter 2026 earnings release and quarterly filing: revenue, gross profit, income from operations, net income, adjusted EBITDA, stock pay, cash flow lines, cash, borrowings, acquisition liabilities, goodwill, share counts, revenue mix and guidance. The gross margins are my calculation from the quarterly income statement of a data provider. Net cash of $62 million is cash of $310.0 million less borrowings of $197.5 million and acquisition liabilities of $50.9 million. The price is the October 9 close from the history page of the quote source. Annual and last-twelve-months free cash flow and stock pay come from the same data provider and use its simpler definition, which differs from the company's.

Weak points, in order. First, 2026 stock pay of about $215 million is my estimate, and owner free cash flow of $40 million is a difference of two larger numbers, so a $20 million error changes it by half. Second, the 2028 and 2031 margins and multiples are my judgment on the SF-07 tiers, with no peer table and no history of the company's own multiple. Third, the diluted share count of 262 million is my estimate (cover count plus the gap between diluted and basic weighted shares), because the counts of restricted stock units and options were not in the part of the filing I read. Fourth, I did not read the credit facility terms or know whether the $250 million term loan was drawn. Fifth, the filing was read only in part (the acquisition note was cut off), so the Marigold purchase price and the earnout terms are not in this analysis. Sixth, I have no net revenue retention figure or customer cohort data.

The strongest counter-argument is that Zeta is cheap for what it grows. EV is 4.7 times 2026 revenue and 21 times adjusted EBITDA, with 25% organic growth, rising margins, a strong data asset and a raised outlook 20 times in a row. If stock pay falls faster than I assume and growth holds near 25%, owner earnings could double quickly, and the Bull value of $46.04 is 39% above the price. The reply is that the owner margin is only 2.2% today, that the price already asks for 23 times my 2028 owner EBITDA, and that the shares have almost doubled in a year, so a buyer is paying for the margin gain in advance.

Single changes keep the Base between $23.41 and $30.07, and the maximum buy price between $11.71 and $15.04, below $33.09 every time. Stacking all cautious inputs gives $17.12, 48% below the price, and stacking all generous inputs gives $38.70, 17% above it. The price sits above every single-change Base value and inside the generous stack, which is why the verdict is a hold and not a sell. The publication gate status is PASS_WITH_WARNING.

Final verdict: hold at $33.09, do not add

Zeta is a fast-growing marketing platform with a rising margin, a net cash balance and a record of beating its own outlook. At $33.09 the shares already pay for the next two years of margin gains, and the cash left to owners after stock pay is only about $40 million. That leaves little room for a miss, and the stock is up almost 80% in a year.

Verdict: HOLD at $33.09, do not add. Triangulated Base value $26.74, downside 19.2%, Confidence C (14/25). Required margin of safety 50%, maximum buy price $13.37.

For an existing holder the growth, the margin trend and the net cash justify holding. New money waits for $13.37 or below, or for proof in the numbers: stock pay below 10% of revenue and free cash flow after stock pay above $100 million. The main thesis breaker is organic growth falling below 15%. Twenty quarters of beat and raise are in the price, and the next ones need to arrive with less stock pay attached.

Sources

Primary sources are Zeta Global's second-quarter 2026 results release (August 2026) and its quarterly report on Form 10-Q. Call commentary on Athena, guidance, buybacks and the credit facility is from the Nasdaq call summary and the Quartr earnings summary. The headline on the 20th consecutive beat and raise is from the company release carried by CMSWire. Quarterly income statements and multi-year cash flow and stock pay come from stockanalysis.com quarterly financials and cash flow statements, the price history from the history page, and the market value, 52-week range and next earnings date from the quote page. Owner free cash flow, owner EBITDA, 2026 stock pay, the diluted share count, net cash, the multiples and every scenario input are my own calculations or estimates, labeled as such.

The reference market price is $33.09, the October 9, 2026 regular session close. Balance sheet figures are as of June 30, 2026. Share count of 251.0 million is as of July 29, 2026 (cover of the quarterly report).