# Clay Pricing in 2026: What Actions and Data Credits Actually Cost

> Canonical: https://www.yalc.ai/blog/clay-pricing-2026-actions-vs-data-credits/

The March 2026 overhaul split Clay's meter in two. Marketplace data got 50 to 90 percent cheaper. Orchestration got taxed. Here is what changed, who wins, and when the numbers stop making sense.

Clay pricing in 2026 splits into actions and data credits, two separate meters introduced on March 11 2026. Data credits pay for marketplace data from Clay's 150 plus providers, and got 50 to 90 percent cheaper. Actions pay for workflow steps, AI calls, and CRM pushes, and are now what taxes daily reruns.

The pricing page reads like an accountancy change. It is not. The split is a strategic statement about where Clay thinks its moat actually sits, and it quietly moves the cost of iteration from free to per run. If your outbound cadence reruns a play every morning, your bill goes up. If your motion is a one shot pull, your bill goes down. This is the operator read of what changed, what it costs, and when the meter stops making sense for the shape of [GTM stack](/blog/gtm-stack/) your team actually runs.

## What changed in Clay's March 2026 pricing update

On March 11 2026, Clay retired its flat credit model and replaced it with two meters. Under the old system, every unit of work in a Clay table, whether pulling an email from a provider or running an AI formula, drew from one shared credit pool. Under the new system, the two are separated. Data credits cover marketplace enrichment. Actions cover platform work. Legacy customers had until April 10 2026 to switch plans or stay on the old contract ([Salesforge](https://www.salesforge.ai/blog/clay-pricing)).

Three headline shifts came with the split. Marketplace data prices dropped 50 to 90 percent, since Clay renegotiated with providers and passed the savings through. The top up premium above your plan allowance dropped from 50 to 30 percent, so overages hurt less. And every workflow step, AI call, and CRM sync now consumes an Action, priced under one cent each ([Clay pricing](https://www.clay.com/pricing)). Clay's own internal memo projected a roughly 10 percent revenue hit from the change ([Clay pricing memo](https://www.clay.com/blog/clay-pricing-memo-internal)).

The operator reading is simple. Marketplace data used to be Clay's most visible line item, and the vendor voluntarily cut its own most visible line item by up to 90 percent. Vendors do that when they are conceding a category is commoditizing. What replaced it as the visible cost is orchestration, which tells you where Clay now believes the moat sits.

## Data credits vs actions, plainly

The two meters answer two different questions, and the confusion in Clay's community is entirely about which question a given feature belongs to.

**Data credits** answer the question, is Clay reaching out to a marketplace provider to buy a fact. Every email lookup, phone number pull, technographic append, or company data fetch consumes data credits. A typical waterfall enrichment for one contact burns 2 to 5 data credits for the email, 3 to 5 for the phone, and 13 to 24 for a full contact record ([Cleanlist](https://www.cleanlist.ai/blog/2026-03-12-clay-pricing-changes-2026)). Data credits are around 5 cents each at plan rate. Unused data credits roll over up to 2x the monthly allowance, which sounds generous until you notice it is also a ceiling. Bank two months of underuse and the meter resets to zero on month three ([Clay pricing](https://www.clay.com/pricing)).

**Actions** answer the question, is Clay doing work on your behalf inside the platform. Running a formula, executing an AI prompt, exporting a row, syncing to your CRM, pinging an HTTP endpoint. Each of those is one Action. Actions are priced at less than one cent each and do not roll over. They reset monthly.

> Figure: Comparison of Clay data credits and actions across five criteria including cost per unit, rollover, and 2026 change

The rule to memorize, data credits pay for facts Clay had to buy, actions pay for work Clay had to do. A row that runs three enrichments, applies two AI prompts, and pushes to HubSpot consumes both meters at once, and reruns the same day burn both meters again. This is the part most operators discover in month two, not on the pricing page.

## The 2026 Clay pricing lineup at a glance

Clay publishes four tiers as of September 2026. All monthly prices below are the billed monthly rate. Annual commits are meaningfully cheaper, so verify against the [live Clay pricing page](https://www.clay.com/pricing) before you sign.

| Plan | Monthly price | Data credits | Actions | Best for |
|---|---|---|---|---|
| Free | 0 dollars | 100 | 500 | Kicking the tires |
| Launch | 185 dollars | 2,500 | 15,000 | Solo operator, one motion |
| Growth | 495 dollars | 6,000 | 40,000 | Team with a live outbound cadence |
| Enterprise | Custom, median around 30,400 dollars a year | Custom | Custom | Multi motion GTM team ([Salesmotion](https://salesmotion.io/blog/clay-pricing)) |

Two details the plan grid hides. The Free plan will not survive a real workflow. 500 Actions runs out inside a single afternoon of experimentation, since even a small table with three enrichment columns and an AI prompt burns Actions in the hundreds. And the Growth to Enterprise jump is the wall most growing teams hit around the 800 to 1,200 enriched contacts a month mark, where Launch credits are exhausted and Growth headroom is uncomfortable.

If Clay lands above your budget for the shape of workflow you actually run, the [best Clay alternatives](/blog/clay-alternatives/) breakdown walks through the tools that fit the specific failure mode, since "too expensive" and "wrong architecture" both show up as "Clay does not fit" in the abstract.

## Where the meter hits your workflow, worked with real numbers

The pricing page numbers only tell you what a subscription costs. The number that matters is what one enriched record costs end to end, and how many times you rerun the play.

Take a standard outbound record. It hits a Clay table, runs a waterfall email and phone enrichment across two providers, applies a fit score with an AI prompt, drafts a first line with a second AI prompt, and pushes the result to your CRM. That single record consumes roughly one Action to run the enrichment step, 3 to 5 data credits for the email waterfall, 3 to 5 more for the phone, two Actions for the AI writes, and one Action to sync to your CRM. Call it 4 Actions and 6 to 10 data credits per record.

> Figure: Four step flow showing a single Clay record consuming actions and data credits across enrichment, AI, and CRM push, with a return loop

At Growth plan rates, 1,000 fully enriched contacts a month land somewhere between 1,295 and 2,395 dollars in combined meter cost ([Cleanlist](https://www.cleanlist.ai/blog/2026-03-12-clay-pricing-changes-2026)) on top of the plan fee. That is the honest floor. The number climbs sharply the moment you rerun that play daily to catch new prospects, because Actions do not roll over and each rerun bills at full rate. Under the old flat pricing, iteration was free. Under the new pricing, iteration is a per run tax. This is the single biggest behavioral shift in the March update, and it is not on the pricing page.

## Who wins and who pays more under the new Clay pricing

The split creates clear winners and clear losers. The winners are operators who consume Clay's marketplace data heavily and rerun the workflow rarely. A recruiter pulling one large enriched list a quarter, or a founder running a market map every few months, sees a real price cut on the data half without paying much on the Actions half. Their bill goes down.

The losers are operators who bring their own API keys, or who rerun a daily outbound cadence against a moving list. Bring your own key workflows used to be effectively free of the marketplace tax. Now every step in the workflow still burns Actions, so the total bill can climb even though the data half is free ([Salesforge](https://www.salesforge.ai/blog/clay-pricing)). And daily rerunners get hit twice, once in Actions for every re executed step, and again in data credits because Clay reruns the waterfall on records that changed even a little.

For a head to head against another platform many operators shortlist alongside Clay, our [Clay vs Apollo](/blog/clay-vs-apollo/) breakdown compares both meters against Apollo's flat seat pricing on the same worked example. And if you are already inside Clay and finding the meter uncomfortable, [Clay migration mistakes](/blog/clay-migration-mistakes/) walks through the mistakes that make an exit expensive.

## The strategic read, what Clay just admitted about its own moat

The pricing memo is not just a price change. It is a statement about where Clay thinks the defensible layer sits. Marketplace data got cheaper because the marketplace data is not the moat anymore. Actions got expensive because orchestration is.

The reason is architectural. Clay's marketplace routes through the same providers your competitors can access directly. Crustdata, Apollo, FullEnrich, and a dozen others sell the same emails and phones through APIs to anyone with a credit card. What Clay had, and what nobody else had at the same fluency, was the spreadsheet UI that let a non engineer compose those APIs into a workflow. That composition, not the data underneath, is what Clay is now charging for.

That reading has one operator consequence. If orchestration is the moat, and orchestration is the taxed meter, the buyer's job is to ask whether Clay's orchestration is worth 40,000 Actions a month. Sometimes yes. Sometimes the same composition sits in a folder of markdown files that a real operator can read and edit without a spreadsheet UI. [Yalc vs Clay](/blog/yalc-vs-clay/) is the direct comparison on that exact question. The [Clay learning curve piece](/blog/clay-learning-curve-too-steep/) covers why the UI itself is often the reason teams stall inside Clay.

## How to cut your Clay bill by 30 percent this week

If you are staying on Clay for now, the actionable move is an Actions audit. Almost every Clay table has three to five leaks that burn Actions with no proportional payoff, and closing them is a one afternoon job.

- **Consolidate multi step formulas** into a single Action. Two chained cells that each apply a light transformation cost two Actions per row. One cell doing both costs one.
- **Cache expensive lookups.** If a company enrichment feeds five downstream columns, run it once into a hidden column and reference the cached value, instead of re running the enrichment inside every dependent formula.
- **Gate AI writes behind a fit score.** Most tables run the first line AI on every row, then discard 60 percent because they do not pass qualification. Score first, write second, and you skip AI Actions on the 60 percent that never mattered. The [lead qualification skill](/skills/qualify-leads/) does this outside Clay for the same reason.
- **Kill duplicate CRM syncs.** Two enrichment paths that both push to HubSpot for the same record are two Actions where you needed one. Route through a single sync step.
- **Push signal data to a real store, not back into Clay.** If [Crustdata](/tools/crustdata/) or a hiring signal feed is refreshing a table daily just so Clay can rerun the same waterfall, you are paying Actions for a job a cheap Postgres row would do for nothing.

A team we watch do this in one afternoon typically strips out 20 to 35 percent of monthly Actions with no change in outbound output. The Actions are burning on plumbing, not on prospecting, and the plumbing is the cheapest half to fix.

## When the Clay meter stops making sense

Every operator using Clay eventually hits the point where the meter costs more than the value it produces. The tipping point is different for different motions, but the math is the same.

The tipping point is roughly where your monthly Actions plus data credit top ups cross the flat rate you would pay a data provider directly plus a lightweight operator OS underneath. A team spending 800 dollars a month on Clay Growth plus 400 in top ups is at 14,400 dollars a year. That number is close to what a Crustdata API contract plus a flat rate orchestration layer would cost the same team, with the difference that the orchestration layer is markdown files the operator can read and edit rather than a spreadsheet UI whose logic dies with a subscription. [How to migrate from Clay to Yalc](/blog/how-to-migrate-from-clay-to-yalc/) walks the mechanical steps of the move, and [AI outbound without Clay](/blog/ai-outbound-without-clay/) covers the shape of the daily motion once the Clay meter is off.

The migration is not for everyone. Clay's UI is still the best on ramp for an operator who is not comfortable in a terminal, and the [Claygent alternative for non technical teams](/blog/claygent-alternative-for-non-technical-teams/) piece is the honest read on when to stay. But the meter itself now forces the question. Under old pricing, "just stay on Clay" was the free default. Under new pricing, staying is a monthly line item that grows every time you rerun a play.

## What to do this week

Pull your Clay usage dashboard and split it two ways. How many Actions did the workflow consume last month, and how many data credits. Then divide by the number of enriched contacts your team actually messaged. Anything over 25 Actions per contact means your workflow has plumbing leaks worth an audit. Anything over 15 data credits per contact means you are paying for waterfall redundancy that a single provider could cover.

If the numbers hold up, stay on Clay and run the audit above. If they do not, price the migration. A Crustdata API contract plus [FullEnrich](/tools/fullenrich/) for the waterfall and a markdown configured operator OS is meaningfully cheaper than Growth plus top ups for a rerun heavy motion, and the composition is auditable in a way a Clay table never is. Whichever way it goes, the one wrong move is to keep paying the meter without knowing which side of it is bleeding.

## FAQ

### How much does Clay cost in 2026?

Clay charges 185 dollars a month for Launch with 2,500 data credits and 15,000 actions, 495 dollars a month for Growth with 6,000 data credits and 40,000 actions, and custom Enterprise pricing that lands around 30,400 dollars a year at the median. There is also a free plan with 100 data credits and 500 actions that runs out fast on any real workflow. Annual commits are meaningfully cheaper than monthly billing, so verify on the live pricing page before you sign.

### What is the difference between Clay data credits and actions?

Data credits pay for marketplace data Clay had to buy from a third party provider on your behalf, such as emails, phone numbers, and technographic appends. Actions pay for platform work Clay had to do itself, such as running a formula, executing an AI prompt, exporting a row, or syncing to your CRM. A single enriched record usually burns both meters at once, and rerunning the same workflow tomorrow burns them again.

### When did Clay change its pricing?

Clay split its pricing into data credits and actions on March 11 2026. Legacy customers had until April 10 2026 to migrate to the new plan structure or hold their old contract. Marketplace data costs dropped 50 to 90 percent as part of the change, and the top up premium above plan allowance dropped from 50 to 30 percent.

### Does Clay charge per seat?

No, Clay does not charge per seat. Every paid plan includes unlimited seats, so you can add teammates without a per user line item. The meter that constrains a team is the combination of data credits and actions, both of which are consumed by workflow execution rather than by seat count.

### What happens when you run out of Clay credits?

Actions and data credits behave differently at the limit. If you exhaust actions, workflow steps queue or fail depending on your plan, and you can top up at a 30 percent premium above the plan rate. Data credits stop enrichment lookups at zero, but you can top up the same way, and unused data credits roll over up to 2x your monthly allowance on paid plans.

### Does Clay have a free plan?

Yes, Clay offers a free plan with 100 data credits and 500 actions a month, unlimited seats, and up to 200 rows per table. It is enough to test the interface but not enough to run a real workflow, since a single table with a few enrichments and an AI prompt will exhaust 500 actions inside an afternoon. Treat the free plan as a demo, not as a production tier.

### Is Clay worth it for a small sales team?

For a team of one to three with a one shot list pull motion, the March 2026 price cut on marketplace data makes Clay meaningfully cheaper than it was, and the Free or Launch plan can carry the workload. For a team running a rerun heavy daily cadence, the Actions meter often makes a flat rate operator OS with direct data provider APIs the cheaper choice inside a quarter. The deciding variable is how often you rerun the same workflow.