PostHog vs Amplitude vs Mixpanel: Real Product Analytics Cost for an AI SaaS at 100 Million Events in 2026

The short answer: these three meter usage on genuinely different units, and two of the three stop publishing a predictable rate well before 100 million events a month. PostHog bills per event with a steeply volume-discounted tiered rate and a generous 1-million-event free allowance; its own calculator puts 51 million events at roughly $3,934/month for analytics alone, and extrapolating that tier's rate upward (an ILLUSTRATIVE upper bound, since further volume discounts likely apply) puts 100 million events somewhere in the $6,000–$7,700/month range for analytics, before session replay or other add-on products. Mixpanel also bills per event, but at a materially steeper rate — roughly $1,120/month at just 5 million events — and does not publish a rate card that extends cleanly to 100 million events; at that scale, Mixpanel's cost is realistically QUOTE-ONLY / UNKNOWN, requiring an Enterprise conversation. Amplitude doesn't bill per event at all — it bills per Monthly Tracked User (MTU) — and its own pricing page discloses no per-MTU dollar figure past its free ceiling, making Amplitude's cost at any meaningful scale, including the 100,000-MAU comparison point this article also checks, QUOTE-ONLY / UNKNOWN as well.

The billing-unit mismatch, up front

Events (PostHog, Mixpanel) and Monthly Tracked Users (Amplitude) are not the same thing and do not convert cleanly. An event is one logged action (a page view, a button click, an API call); an MTU is one unique user tracked at least once in a month, regardless of how many events that user generates. A product where each user fires hundreds of events a month will hit PostHog's or Mixpanel's event ceiling far faster, relative to its user base, than it hits Amplitude's MTU ceiling — meaning a direct "100 million events equals X MTUs" conversion does not exist without knowing a specific product's actual events-per-user ratio. This article treats the 100-million-event figure and the 100,000-MAU figure as two separate reference points, not one normalized workload, exactly because the units don't translate.

What each vendor bills

PostHog (OFFICIAL, posthog.com/pricing, restructured February 4, 2026). Bills per event, with separate, independently-resetting free allowances per product — not one shared pool: Product Analytics gets 1 million events/month free, Session Replay gets 5,000 recordings/month free (2,500 for mobile), and Feature Flags get 1 million requests/month free, each tracked and billed on its own meter, so exhausting one product's allowance has no effect on another's. Beyond the free tier, the per-event rate for analytics declines steeply by volume tier rather than staying flat — PostHog's own calculator reportedly puts 51 million analytics events at roughly $3,934/month, implying an effective blended rate near $0.000077/event at that tier, itself already well below the headline per-event figures (as low as $0.00005, or as high as $0.00031 depending on which tier a given source was quoting) that circulate across comparison sites. Session replay's own free allowance was cut from 15,000 to 5,000 recordings a month in a 2026 repricing, a real reduction worth noting for any team relying on the old, more generous allowance. PostHog's core platform remains open source and self-hostable, which eliminates the per-event bill itself, but PostHog's own current guidance is clear that this is not equivalent to free operation: running the platform at real production scale carries substantial infrastructure and operations work of its own, a cost this article does not attempt to price since it varies enormously by team.

Mixpanel (OFFICIAL for the mechanism; REPORTED for exact rates). Bills per event with a 1-million-event-per-month free tier, then a tiered Growth plan. One sourced reference point: 5 million events a month costs roughly $1,120 on the Growth plan — a materially steeper effective per-event rate than PostHog's at a comparable volume. Mixpanel's own pricing documentation does not publish a rate card extending to 100 million events a month; costs are widely reported to "escalate significantly" beyond the low-single-digit-million-event range, which is consistent with Mixpanel steering high-volume customers toward a custom Enterprise conversation rather than a published per-event rate at that scale. Mixpanel now bundles session replay, experiments, and feature flags directly into both its Free and Growth plans — a real 2026 product expansion that moved these from Enterprise-only add-ons: Free includes 10,000 session replays/month and up to 10 active feature flags; Growth includes 20,000 to 500,000 session replays (customizable) and up to 50 active feature flags, with experiments sharing the feature-flag infrastructure rather than metering separately.

Amplitude (OFFICIAL for the mechanism; pricing itself undisclosed past the free tier). Bills per Monthly Tracked User, not per event — a structurally different meter from the other two. Free: up to 50,000 MTU. Amplitude's own pricing page states that its paid "Plus" tier starts at $0 and lists an event ceiling around 70 million events a month, but discloses no per-MTU or per-event dollar figure anywhere on the page, not even in a calculator. Exceeding the Plus tier's limits — including the 100,000-MAU reference point this article checks — requires the custom-quoted Growth plan. Session Replay is a separate add-on product on Amplitude, not bundled into core analytics plans at any tier, unlike PostHog's bundled-by-default approach.

Cost at 100 million events/month

Formula: PostHog ILLUSTRATIVE estimate = extrapolated from the sourced 51-million-event data point, treated as an upper bound since PostHog's declining tiered structure means the true 100-million rate is likely somewhat better than a flat doubling.

PostHog's self-hosted option is the one genuine alternative to all of this at real scale: a team with the operational capacity to run its own infrastructure can eliminate the per-event bill entirely, replacing it with infrastructure and operational cost this article does not attempt to price, since it varies enormously by team.

Cost at 100,000 MAU

This is Amplitude's native billing dimension, but even here its published pricing stops well short. Formula: Amplitude free tier covers up to 50,000 MTU; a product at 100,000 MAU is double that ceiling.

Even on its own native unit, Amplitude does not offer a self-serve, published price at the exact scale this article is checking — a 100,000-MAU AI SaaS product evaluating Amplitude should expect a sales conversation, not a calculator result.

Session replay: a separately-metered cost on every vendor that offers it

Session replay is never folded into the base per-event or per-MTU rate on any of these three, though Mixpanel has recently moved it closer to being a genuinely bundled feature. PostHog meters recordings on its own separate free allowance (5,000/month), materially smaller than its analytics event allowance (1,000,000/month) — a high-traffic consumer application generating many sessions can exhaust its replay allowance long before its event allowance, and replay cost "can dwarf analytics costs" for such a product. Amplitude sells session replay as a wholly separate add-on product, not included in any core analytics tier at any price point. Mixpanel, as of its 2026 product expansion, now includes session replay directly in both Free (10,000/month) and Growth (20,000–500,000/month) — the most generous included replay allowance of the three at the free tier specifically, though still a separately-tracked allowance from event volume rather than drawn from the same pool.

Why two of these three can't be cleanly priced at 100 million events

This is not an oversight — it reflects how these two vendors actually sell at scale. Mixpanel's and Amplitude's pricing pages are both structured to surface a clean number only at low-to-moderate volume; past a certain point, both funnel the conversation toward sales rather than a self-serve calculator. PostHog is the outlier specifically because its pricing philosophy is built around staying usage-transparent even at high volume, which is why it is the only one of the three with a sourced reference point anywhere near this article's 100-million-event target.

Sensitivity

1. Events per user. The variable that would let a reader convert between PostHog's/Mixpanel's event-based pricing and Amplitude's MTU-based pricing; without it, the three platforms' costs for "the same product" cannot be meaningfully compared.
2. Session replay volume relative to event volume. A product with few events per session but many sessions can be dominated by replay cost rather than analytics cost, especially on PostHog's much smaller replay allowance.
3. Self-hosting capacity, for PostHog specifically. Eliminates the per-event bill entirely in exchange for infrastructure and operational cost this article does not price.
4. Whether a given volume crosses into Enterprise-quote territory. This happens earlier on Mixpanel and Amplitude than on PostHog, changing a calculable number into a sales conversation.

Budgeting traps

• Extrapolating PostHog's lower-volume per-event rate linearly to 100 million events. The real rate at that volume is almost certainly lower per event than this article's upper-bound estimate, due to continued volume-tier discounting PostHog's calculator applies but this article cannot precisely replicate without live access to it.
• Assuming Amplitude's "$0" Plus tier headline means free at any realistic AI SaaS scale. The free ceiling is 50,000 MTU, and the tier above it has no published price at all.
• Treating PostHog's 1-million-event free allowance and its session-replay free allowance as the same size. The replay allowance is 5,000 recordings, a much smaller number that a high-session-volume product can exhaust far sooner.
• Comparing Mixpanel's low-volume $1,120-at-5M reference point to a 100-million-event workload as if the rate stays flat. Mixpanel's own pricing structure does not support that extrapolation, and the real number at that scale is a sales quote.

What to ask before you buy

Ask Amplitude and Mixpanel directly for a quote at your actual expected event or MTU volume once it approaches or exceeds the ranges this article could price from public information, since both vendors route high-volume pricing through sales rather than a published calculator. Measure your product's actual events-per-user ratio before trying to compare PostHog's or Mixpanel's event-based cost to Amplitude's MTU-based cost, since no universal conversion between the two exists.