Where did Stripe's 2025 leavers go?

In December 2025 we pulled every publicly visible role change from Stripe employees going back twelve months. 312 people. The expected story was Ramp, Mercury, and a wave of fintech founders. The actual pattern was nothing like that. Three findings, all measurable, none of them what the VC Twitter consensus claimed at the time.

TL;DR

Of the 312 Stripe leavers tracked from December 2024 through November 2025, roughly 40% landed in non-fintech infrastructure (AI, dev tools, data), more than half of those who founded a company started outside payments, and the much-discussed "competitor defector" path to Ramp or Mercury was real but accounted for under 15% of the cohort. The implications for recruiting and venture sourcing are very different from the conventional fintech-mafia narrative.

How we built the dataset

The cohort is 312 people who left Stripe between December 2024 and November 2025, identified from publicly visible LinkedIn role-change events. We pulled the data through LinkFetch's companies.timeseries and profiles endpoints in two passes: one to enumerate Stripe employees with a role change in the window, one to fetch the destination role for each. Public profiles only; private profiles were excluded, which means the real cohort is likely 15 to 20% larger than 312. The destinations we report are the modal landing role; people who went through a 90-day stealth gap before announcing a new company were re-pulled in February 2026 to catch the eventual disclosure.

This was a one-shot research dataset, not an ongoing tracker. The companion hiring-signal recipe is what we'd use to maintain it as a live workflow.

For context on the denominator: Stripe employed around 8,017 people as of August 2025, following the January 2025 layoff of 300 roles across product, engineering, and operations. So 312 disclosed leavers in twelve months sits at roughly 4% of headcount, slightly above the disclosed turnover rate. Our cohort over-indexes on senior ICs and managers, who are more likely to update LinkedIn promptly.

Pattern 1: 40% went to non-fintech infrastructure

The biggest single destination bucket was AI and infra companies. Anthropic, OpenAI, Linear, Vercel, and a long tail of dev-tools and data-infra companies absorbed roughly 40% of the disclosed leavers. Almost none of these moves are payments-adjacent in the way Ramp or Mercury are; the engineers and PMs going to Linear are not solving fintech problems, they're solving issue-tracker and developer-velocity problems.

This isn't a payments story at all. It's an infra-maturity story. The people who built Stripe's last decade of payments infra are an unusually well-shaped fit for the systems-engineering hiring waves at Anthropic and OpenAI, where the bar is high-throughput, high- reliability, and high-leverage backend work. Three years ago, the same profile would have gone to a Snowflake or a Databricks. The center of gravity moved.

For recruiters, this is the actionable finding. If your sourcing profile says "ex-Stripe", you're competing against AI labs first, and fintechs second. You need the AI-lab story, not the fintech story.

Pattern 2: Founders started outside payments

Of the 312, 41 (about 13%) were founding a new company by November 2025. The conventional wisdom would have most of these in payments, billing, or vertical fintech. The actual mix was: 18 in dev tools or AI infra, 9 in vertical SaaS (legal, healthcare, climate), 8 in fintech-adjacent (compliance, treasury, B2B billing), and 6 still in stealth as of the data pull.

The two highest-profile public examples illustrate the split. Duna, which closed a €30M Series A in February 2026, is built by Stripe alumni and is squarely fintech (B2B compliance and KYB). Seapoint, which closed a €7.5M seed in April 2026, is also Stripe-alumni-founded and serves the European founder banking niche. These two get the press, but they're 2 out of 41. The modal ex-Stripe founder in 2025 was building a dev tool, not a payments product.

There's an interesting second-order effect in the data: every fintech founder we identified had a payments or compliance person in the founding team. The non-fintech founders did not, almost without exception. Stripe's payments-and-compliance bench is what makes Stripe-alumni fintech startups credible; outside that domain, the ex-Stripe brand is a hiring asset but not a thesis asset.

Pattern 3: Competitor defection is real but small

The story that sells the most newsletters is "Stripe alumni jumping ship to Ramp / Mercury / Brex". It's real, but it's small. Of the 312 leavers, 44 (around 14%) went to a direct or adjacent competitor in fintech. Ramp absorbed the most (16), Mercury second (11), Brex and Adyen tied for third (5 each), and the remainder spread across smaller players.

Compared to the 40% who went to AI / infra and the 13% who founded, the competitor-defection pool is the third-biggest bucket but not the dominant one. The narrative inversion matters: when a recruiter says "we're sourcing from Stripe", the highest-yield search is into AI infra, not into Ramp's headcount-delta data.

A finer-grained look at the 44 competitor-defectors shows another asymmetry: most of them are in GTM roles (sales, partnerships, success). The engineering and product talent went to AI infra. Ramp and Mercury hired Stripe's go-to-market bench; Anthropic and OpenAI hired the eng bench. Both make sense given each company's hiring need at the time; the data is just unusually clean here.

Geographic split: SF and Dublin moved differently

The cohort doesn't move uniformly across Stripe's offices. Of the 312 leavers, roughly 58% were based in San Francisco / Bay Area at the time of departure, 14% in Dublin (Stripe's EMEA HQ), 9% in New York, 7% in Seattle, and the remainder spread across Singapore, London, Toronto, Mexico City, and remote roles. The destination patterns differ enough by origin office that lumping them together hides the real story.

San Francisco leavers heavily skewed toward AI and infra (around 52% of SF departures landed in that bucket), with a smaller-than-average share founding companies. Dublin leavers, by contrast, over-indexed on fintech-adjacent destinations: Adyen, Tide, Wise, and the European banking-as-a-service set absorbed close to a third of disclosed Dublin departures. The Dublin office is materially closer to the European fintech operator network than the SF office is to the US fintech operator network, and it shows.

New York leavers split closer to even between fintech and AI infra, which tracks with NYC's general 2025 hiring pattern. Seattle leavers went almost entirely to Microsoft, OpenAI, and a small cluster of Vercel and Cloudflare roles, which is the most concetrated destination pattern in the dataset by a wide margin.

For a recruiter or VC running a Stripe-alumni search in 2026, the practical takeaway is to filter by Stripe office of origin, not just by Stripe-as-employer. The candidate signal is meaningfully different across the cities, and the conversion rate of an outreach message that names the right adjacent destination is materially higher than one that pitches a generic "ex-Stripe is great for our culture" line.

How this cohort compares to the 2023 baseline

We ran the same destination analysis on a smaller 2023 cohort (118 disclosed leavers from December 2022 to November 2023) for comparison. Three things changed.

First, the AI-infra destination bucket roughly tripled in share. In 2023, around 13% of disclosed Stripe leavers went to companies we'd classify as AI-infra-first; in 2025, that share is close to 40%. The absolute count of AI-infra hires grew faster than the cohort itself grew, which is the cleanest read on where the talent gravity actually pulled.

Second, the founder share grew but stayed small. The 2023 cohort had 9% founders (11 of 118); the 2025 cohort has 13% (41 of 312). That matches the broader founder-formation uptick coming out of the 2024-2025 AI-tooling boom, but it's not the dominant story. Most ex-Stripe people in 2025 are still going to take a senior IC or manager role at another company, not start one.

Third, the average seniority of leavers crept up. The 2023 baseline skewed mid-level (L5 or equivalent at the modal point); the 2025 cohort sits closer to senior IC and manager (L6 to M7 equivalent in Stripe's bands). This is partially mechanical (the company itself matured and so did its tenured staff), and partially behavioural (the AI-lab hiring waves of 2025 specifically targeted senior engineers and managers, which is the cohort that got pulled hardest).

These three deltas, together, explain why the 2025 narrative diverges so sharply from the 2023 narrative. The talent market wasn't the same market, and the places that were hiring at the most aggressive comp were not in fintech.

Cross-cutting finding: stealth duration is shrinking

One small but useful methodological note: the median time from a Stripe departure to a public new-role disclosure was 47 days in this cohort, down from a 2023 baseline of around 80 days for the same function-mix. Stealth gaps are getting shorter, partly because LinkedIn's own "Open to Work" features now nudge users to update sooner, partly because the AI-lab hiring wave is moving fast enough that companies announce hires within a week of accept.

For talent-watching workflows, this means a weekly cadence is no longer fast enough to be at the front of the line. A daily scan, like the VC investor radar prompt, is the appropriate cadence in 2026.

The shrinking-stealth pattern has a second-order consequence worth calling out. In the 2023 cohort, roughly a quarter of founders disclosed their new company before formally announcing a fundraise, giving investors a clean 4 to 6-week window to introduce themselves. In the 2025 cohort, that gap has compressed: closer to half of the founders who eventually raised a seed round in 2025 had already closed the round (or held a verbal commitment) by the time their LinkedIn went live. Funds that wait for the public profile update are increasingly arriving at a closed party. The right play is to track the departure, not the announcement, and the right cadence to do that on is daily.

We also saw a small but real pattern of "soft stealth": founders who updated their LinkedIn to "Founder, Stealth" without removing their prior Stripe role. About 18% of the 2025 founders used this pattern, versus around 6% in the 2023 baseline. The signal is easy to miss unless your prompt looks for headline changes specifically, not employer changes; the VC prompt library has a dedicated trigger-word scan for exactly this case.

Mid-2026 update: the alumni-fund signal

Between the original publication and this June refresh, a second-order signal showed up in the dataset that we did not anticipate. The Stripe alumni network has formalized into a recognizable fund-and-list ecosystem: third-party trackers like StartupHub now publish a curated Stripe Alumni-Founded AI Startups list, and several alumni-led seed funds have started backing each other's portfolios.

What does that mean for our cohort? It means the "stealth" window we flagged at 47 days is partially gameable: a founder who joins a Stripe-alumni fund's portfolio is publicly identifiable through the fund's announcements, often weeks before their own LinkedIn updates. The practical consequence is that a sourcing or VC workflow that watches both LinkedIn role changes AND alumni-fund portfolio disclosures catches roughly 22% more new-company formations than the LinkedIn-only signal we used in the original analysis. For founders who left Stripe in the first half of 2026, we'd expect to see this gap widen further by Q4.

The other thing this confirms is the founder share trend. The 13% founder rate in the 2025 cohort was tracking toward a higher number through the spring, and the alumni-fund formation supports that. We plan to re-run the destination analysis in December 2026 and expect the founder share to land in the 16 to 19% range for the 2026 calendar-year cohort. The single biggest source of new founders will likely still be the AI-infra adjacency, not payments. For founders in that group, the design partner recruitment routine is the right second step after the LinkedIn headline goes live.

What this means for sourcing

Three concrete takeaways.

First, if you're recruiting out of Stripe, design the sourcing search around AI-infra fit, not fintech fit. The candidate pool is more likely to be evaluating Anthropic than Ramp.

Second, if you're a fund, the cofounder-detection signal at Stripe is strongest in dev tools and AI infra, not payments. The next Duna exists, but for every one of those there are roughly four ex-Stripe founders building something with no payments component at all.

Third, monitor the stealth window. The median is now 47 days. A quarterly portfolio scan misses most of the action; a weekly one catches some; a daily one catches almost everything. We use the hiring-signal recipe internally for this.

Frequently asked questions

Why 312 and not the full leaver count?

312 is the count of disclosed, public-profile role changes that we could verify against a destination employer in the window. Private profiles, profiles where the new role wasn't yet listed, and profiles where the leaver hadn't updated LinkedIn at all are excluded. Our internal estimate of the true 12-month leaver count is closer to 380 to 420; the missing 70 to 110 are mostly people who left without updating their profile, weighted toward functions where LinkedIn matters less (some support and ops roles).

How is this different from a Crunchbase or Specter cohort study?

Crunchbase and Specter index on funding events and company creation, which under-counts the people who took a senior IC role at an established company. LinkFetch's cohort here is people-first: every person who left Stripe, regardless of whether their next employer is in any startup database. That difference is exactly why the AI-infra landing pattern shows up here and doesn't show up in funding-data studies.

Can I get the underlying dataset?

We share the deidentified destination breakdown with recruiting and venture teams on request. The individual-level data isn't shared externally, since it's sourced from public LinkedIn profiles under a passive-observer model that constrains how the data can be reused. Email info@linkfetch.io with the cohort question you'd run next.

Is this kind of analysis legal?

Yes. Every data point is from a public LinkedIn profile, fetched under the user-principal model that powers LinkFetch's Chrome extension. The legal and compliance details are walked through in the GDPR + CCPA guide for LinkedIn data APIs.

---

Originally published April 2026. Last updated 2026-06-22 with the Stripe alumni-fund formation signal, an updated founder-share projection for the 2026 cohort, and the practical consequences for sourcing workflows.