{
  "$type": "site.standard.document",
  "bskyPostRef": {
    "cid": "bafyreiagu464leytp35kid52qzrmcncjruh4y4nogyqdjblzfxcslqci3m",
    "uri": "at://did:plc:uwj5fyuv3lbhhoybn5hnrqx4/app.bsky.feed.post/3mmrtx4mwzdl2"
  },
  "coverImage": {
    "$type": "blob",
    "ref": {
      "$link": "bafkreicz7xylc4k2lsrx322ydj62bd7bx4zn63xofuow22ui2a2aaqotxe"
    },
    "mimeType": "image/png",
    "size": 1274012
  },
  "description": "AI credit expiry is becoming a margin control, not just a billing footnote. The safer pattern is short-lived included credits, explicit paid-credit expiry, and admin controls that turn heavy usage into visible expansion.",
  "path": "/do-ai-credits-need-expiration-dates-pricing-rules-that-protect-nrr/",
  "publishedAt": "2026-05-26T20:38:24.000Z",
  "site": "https://blog.tuguidragos.com",
  "tags": [
    "unused HubSpot Credits expire at the end of each usage period and do not roll over",
    "AI credits expire monthly and do not roll over",
    "net revenue retention of 125%",
    "median NRR of 103%"
  ],
  "textContent": "AI credits have become a pricing control point because AI usage is not a fixed-cost entitlement. In B2B SaaS, inference and agentic workloads create variable cost exposure, so the question is not only whether credits expire. The sharper question is which credits expire, how visibly, and whether the policy supports renewal expansion rather than customer frustration.\n\nThe 2026 pattern across several major vendors is time-boxed entitlement plus metered expansion. Included AI credits often reset quickly with no rollover, while purchased API credits can carry longer but still explicit expiration windows. That structure protects margin, but only if customers can monitor usage and control what happens when they hit the limit.\n\n## Included AI credits are being treated like monthly capacity\n\nHubSpot is explicit: unused HubSpot Credits expire at the end of each usage period and do not roll over. The credits reset every month and apply to usage-based features including Customer Agent, Prospecting Agent, Data Agent, and Data Studio syncs. This is not rollover generosity. It is a recurring capacity allowance.\n\nFigma uses a similar rule for AI credits included with seats. Its help center states that AI credits expire monthly and do not roll over. Figma also says seat credits cannot be shared or transferred, and users who run out cannot complete more AI actions unless credits reset or an admin adds more credits.\n\nFor pricing teams, the logic is clear. Monthly non-rollover credits cap unused entitlement from accumulating into a future cost spike. They also separate normal usage from sustained heavy usage, which can then be monetized through added credits, higher capacity, or usage-based charges.\n\n## Purchased credits get longer lives, but not unlimited lives\n\nAPI credit models show a different balance. OpenAI’s service credit terms state that service credits are not refundable except where required by law and expire one year after purchase or issuance unless otherwise specified. OpenAI also says customers can review the available service credit balance in the applicable account or workspace.\n\nAnthropic’s Claude API support materials describe prepaid usage credits for API and Workbench usage. They state that failed requests are not charged and customers are billed only for successful API calls and completed tasks. Purchased credits expire one year after purchase, the expiration date cannot be extended, purchases are non-refundable, and customers can configure auto-reload when balances fall below a defined threshold.\n\nThis distinction matters. Included credits are part of the subscription entitlement and can reasonably reset monthly. Purchased credits feel closer to prepaid value, so the expiry period needs to be explicit and operationally manageable. The customer should know the balance, the expiry rule, and the refill path before credits run out.\n\n## The NRR case depends on controlled expansion\n\nThe strongest revenue-retention case appears where usage naturally expands. Snowflake reported that product revenue increased by $1.0 billion in FY2026, primarily due to increased consumption by existing customers, and disclosed net revenue retention of 125% as of January 31, 2026. The same filing also says cost of product revenue increased, including a $248.1 million increase in third-party cloud infrastructure expenses, including AI inference-related costs.\n\nThat combination is the pricing problem in miniature. Consumption expansion can support high NRR, but it can also increase infrastructure cost. Credit expiry and metered expansion rules are tools for aligning those two forces, not proof that usage-based pricing automatically improves retention.\n\nThe benchmark context is important. SaaS Capital’s 2026 survey of more than 1,000 private B2B SaaS companies reported median NRR of 103% for bootstrapped SaaS companies with $3M to $20M ARR, with 90th percentile NRR at 117.9%. Against that backdrop, controlled expansion is a premium outcome, not a default assumption.\n\nThe practical rule is to let included AI credits expire quickly, while giving paid add-ons and top-ups explicit expiry, balance visibility, alerts, and admin controls. HubSpot’s model is notable here because it combines monthly expiry with auto-upgrades, pay-as-you-go overages, maximum monthly limits, and usage notifications at 75%, 85%, 90%, and exceeded thresholds.\n\nExpiration dates are not the enemy of NRR. Opaque expiration is. A renewal-safe AI credit policy makes usage visible, makes overages opt-in or bounded, and gives customers a clear path from recurring heavy usage to a higher-capacity plan. That is how pricing teams can protect gross margin without making buyers feel that paid value is disappearing.",
  "title": "Do AI Credits Need Expiration Dates? Pricing Rules That Protect NRR",
  "updatedAt": "2026-05-26T20:38:25.043Z"
}