For most of the last decade, the answer to “how much should a founder pay themselves” was a posture, not a number. Pay yourself as little as you can stand. Eat the ramen, prove your conviction, signal to investors that every dollar goes into the company and not your rent. The number was never stated because the point was that the number should be embarrassing. That era is over, mostly because two accounting firms started publishing the actual payroll data, and the data says something the posture never allowed: the founder who underpays themselves into a corner is frequently the founder who does not make it to the next round.
The most useful thing about the current numbers is that they come from payroll records, not aspiration. Kruze Consulting, a CPA firm that does the books for hundreds of venture-backed startups, builds its Startup CEO Salary Report from actual payroll runs rather than survey responses. That distinction matters, because founders lie to surveys in a predictable direction. They report the salary they think they should be taking, which is lower than the one they are actually taking, because the posture is still in their heads.
What the payroll data actually shows
Kruze’s headline figure is the average startup CEO salary, and its arc tells the funding cycle in one line. Per the 2026 report, the average climbed from $130,000 in 2018 to $150,000, dipped to $141,000 to $142,000 across 2023 and 2024 as fundraising slowed, then resumed climbing to $161,000 in 2025 and $165,000 in 2026. The median in 2026 sits at $159,000, close enough to the average that Kruze reads the distribution as a tight band rather than a few highly paid outliers dragging the mean.
By stage, the 2026 benchmarks are concrete enough to hold a number against. Kruze puts the seed CEO range at $130,000 to $170,000 with a $153,000 median, Series A at $180,000 to $230,000 with a $203,000 median, and Series B at $200,000 to $260,000 with a $216,000 median. The firm’s separate compensation guide, built from over 450 funded startups, breaks the seed founding team apart by role: CEO at $132,000, CTO at $134,000, COO at $135,000, and product or CPO founders at $149,000. The detail worth pausing on is that founding CTOs often out-earn founding CEOs at the earliest stage, which Kruze attributes to how much harder a technical founder is to find than a chief executive.
Then there is the other dataset, and it disagrees in a way that is more instructive than a single source would be. Pilot, another accounting firm serving startups, surveys a broader and smaller-company base. Its 2025 Founder Salary Report, drawn from 1,844 founders, found the median founder salary fell from $132,000 in 2024 to $75,000 in 2025, a 43 percent drop. Nearly twice as many founders, 60 percent against the prior 37 percent, now pay themselves under $100,000. Bootstrapped companies in the sample rose 77 percent year over year to 18 percent of respondents.
The two firms are not contradicting each other so much as describing two different populations. As Jason Lemkin noted at SaaStr, Kruze skews toward mostly B2B, venture-backed companies and tends not to take on the tiniest clients, while Pilot’s base is broader and includes more of the very early and self-funded. The Kruze number is what a funded founder with runway pays. The Pilot number is what the wider field, including a swelling cohort of bootstrappers, can actually afford. A founder reading these should locate which population they belong to before they borrow a benchmark from the other.
The detail buried in the Pilot data
One finding in the Pilot report does more work than the headline, and it is the one most founders skip past. Most founders, 31 percent of the sample, set their salary based primarily on “what the startup can afford.” The founders who instead benchmarked their pay against market rates earned 79 percent more on average. That is not a story about greed. It is a story about anchoring. The founder who asks “what can we afford” anchors on the company’s fear. The founder who asks “what is this role worth” anchors on the market, and then negotiates down from there for runway reasons rather than up from zero out of guilt.
Waseem Daher, Pilot’s co-founder, framed the 2025 drop as discipline rather than distress, telling the wire that founders are “rethinking compensation” and “make smarter calls to extend runway.” That is true at the level of the ecosystem. It is also the kind of sentence that gives an individual founder permission to underpay themselves past the point of discipline and into the point of personal risk, which is a different thing.
The three phases, and where each one breaks
The data sorts founder pay into three rough phases, and each has a distinct failure mode.
The first is the ramen phase, before there is real revenue or a meaningful round. Paul Graham’s definition remains the cleanest articulation of the goal here: ramen profitability means “a startup makes just enough to pay the founders’ living expenses,” and its main value is not that it signals virtue but that “it buys you time” and means “you’re no longer at the mercy of investors.” Graham is precise that this is feasible mainly for software startups, “because they’re now so cheap,” where “the only real cost is the founders’ living expenses.” The phase is real and the discipline is real. The failure mode is treating ramen as a permanent identity rather than a bridge, and wearing the low number long after the company can pay more.
The second is the market-minus phase, after a seed or Series A round closes. This is where Kruze’s stage bands apply, and where the posture does the most damage. A founder who has raised an institutional seed and still pays themselves $40,000 out of pride is not being disciplined. They are introducing a personal-finance crisis into the life of the person whose judgment the company most depends on. Kruze’s own founder self-check flags a salary more than 20 percent below the stage midpoint as a possible underpay that means “you may be taking on unnecessary personal risk, especially if your company has already raised meaningful capital at a healthy valuation.”
The third is the market phase, at Series B and beyond, where the question inverts. Here the risk is overpaying relative to performance, and the check is whether the valuation and traction justify the cash. Kruze’s data shows Series B salaries are the volatile ones, having declined from a 2022 peak to $227,000 in 2024 and sliding to $214,000 in 2025 per the SaaStr breakdown, as boards pushed efficiency. The later-stage founder is no longer fighting their own guilt. They are negotiating with a board that reads salary as a signal of discipline.
The founder who asks “what can we afford” anchors on the company’s fear. The founder who asks “what is this role worth” anchors on the market.
The cost of underpaying yourself
The case against underpaying is not that founders deserve comfort. It is that a founder in personal financial distress makes worse decisions, on a faster clock, with the entire company downstream of them. The founder who cannot make rent starts optimizing for a near-term outcome that solves their personal cash problem rather than the company’s actual strategy. They take the wrong acqui-hire conversation, raise at the wrong terms because they personally need the round to close, or burn out and leave. The Pilot data showing only 5.4 percent of founders now take zero salary, down from 9 percent, is quietly encouraging on this point. Fewer founders are running on fumes, even in a tighter market.
There is a second-order cost that the salary surveys cannot measure but the cofounder-conflict literature can. A founder paying themselves nothing while a cofounder draws a salary, or two founders who never reconciled their pay early, builds a resentment that compounds quietly. Money disputes are one of the documented triggers of cofounder breakups, and pay asymmetry that nobody discussed at the start is one of the cleanest ways to manufacture one.
The cost of overpaying yourself
The inverse error is rarer at the early stage and lethal at the growth stage, and it is mostly a runway problem. Every dollar of founder salary above market is a dollar of runway converted into personal income, and runway is the only thing that buys the time to reach product-market fit. Kruze’s self-check makes the arithmetic literal: it asks how much a $25,000 salary increase changes your runway. If the answer is “several months,” you cannot afford the raise yet, whatever the benchmark says. The founder who indexes their pay to the stage table without checking it against their own burn has confused a benchmark for a budget.
A usable rule
The honest synthesis of the data is not a single number, because the right number depends on which population you are in and how much runway you have. The usable rule is a sequence. Benchmark first, against the market rate for the role and stage, using the Kruze stage bands as the reference point, not against what the company can afford. Then discount deliberately for runway, treating the discount as a temporary financing decision with a date attached, not a permanent identity. Then re-benchmark at every funding event, because the stage changed and the posture should change with it.
The reason the “pay yourself as little as possible” advice survived so long is that it was free to give and impossible to disprove without data. The data exists now. It says the average funded CEO takes $165,000, the median founder in the broader field takes $75,000 in a tight year, and the founders who anchored on the market rather than on fear took 79 percent more than the ones who asked what they could afford. The number that should worry you is not the average. It is the size of the gap between you and it, and whether you can say out loud why the gap is that big.