What are the downsides of using Airtable?

The record counter on your Airtable dashboard tells only half the story. The constraints that actually break customer-facing applications live in the architecture itself — hard limits that no plan upgrade removes, no matter how much you pay.

Airtable enforces a 5 requests-per-second API rate limit per base on every tier, including Enterprise Scale. That ceiling is absolute: paying more raises volume quotas, but it does not lift the per-second throttle, the 50-automation cap per base, or the 25-action limit per automation. As one infrastructure team put it, this is "the one you cannot buy your way out of" — and the limit worth checking before committing to an integration design.

The failure mode is not theoretical. A marketing team that built their entire backend on Airtable hit 429 "Too Many Requests" errors on every page load once they reached roughly 500 concurrent users, forcing a frantic weekend migration to Postgres while the app was effectively down. The practitioner who documented the incident now recommends staying under ~25,000 records for any system that serves real traffic.

These structural limits compound in ways that surprise growing teams:

• Record limits are per base, not per table — two tables with 25,000 records each consume the entire 50,000-record Team allowance
• Free-plan collaborator caps (5 editors) often bite before record limits do
• Failed automation runs consume monthly allowance exactly like successful ones
• Attachment storage and API call quotas scale independently, creating multiple choke points

For businesses evaluating whether a tool can support a customer-facing website, booking system, or lead-capture flow, the API rate limit is the go/no-go checkpoint. At AI Business Sites, we see this pattern repeatedly: teams start in Airtable for speed, then hit the architectural wall when the site goes live. The alternative isn't a different database — it's a done-for-you platform where the infrastructure, content engine, CRM, and automations are built to scale from day one, with no per-seat pricing and no hidden rate limits.

Most teams discover Airtable's record limits the hard way: a base that worked fine for months suddenly stops accepting new rows, and nobody saw it coming. The reason is a constraint that trips up even experienced users — records count across all tables in a base, not per table.

According to documented Airtable limits, each base caps out at:

• Free: 1,000 records and 1 GB of attachment storage
• Team ($20/editor/month): 50,000 records and 20 GB
• Business ($45/editor/month): 125,000 records and 100 GB
• Enterprise Scale: roughly 500,000 records and 1,000 GB — a sales-led figure that isn't published on the public plans page

Here's the part that surprises people: splitting your data into more tables buys you nothing. Two tables holding 25,000 records each have already consumed the entire Team allowance of 50,000. A base is a single bucket, and every table drains from it.

Attachments follow the same logic. A team storing project photos, signed PDFs, or product images can hit the Free plan's 1 GB ceiling long before anyone notices the record count — and when a base exceeds its limit, it can become read-only until records are deleted or the account upgrades, per pricing history tracked by PricePulse.

Counterintuitively, most Free-plan teams hit a different wall first. The Free tier allows only 5 editor/creator collaborators, and as InfoLobby's analysis notes, "for most teams the collaborator cap arrives before the record cap." The moment a sixth person needs to edit — a new hire, a contractor, a second salesperson — you're upgrading.

That upgrade math deserves scrutiny. Moving from Team to Business raises your record ceiling 2.5x (50,000 to 125,000) but more than doubles the per-user price, from $20 to $45 per editor per month. For a 10-person team, that's an extra $3,000 a year for headroom you may not need yet — and per-seat pricing means every new hire compounds the bill, a pattern competitor comparisons consistently flag as Airtable's first pressure point for growing teams.

The deeper issue isn't any single ceiling — it's that these limits arrive quietly, mid-growth, exactly when a business is busiest. A base that started as a simple tracker becomes operational infrastructure, and by then the upgrade decisions are forced rather than planned.

This is the pattern we see repeatedly at AI Business Sites: growing businesses stitching together tools whose pricing and caps were designed for smaller teams, then absorbing surprise migrations and seat-fee escalations. The alternative is a system where someone else owns the scaling question entirely — one flat monthly price, no per-seat math, no record counters to watch — so the business grows without the software bill growing faster.

Software pricing rarely moves in your favor — and Airtable's recent history is a case study in how quickly a tool's economics can change under your feet.

In February 2026, Airtable cut its free-plan record limit from 1,200 to 1,000 per base — a 17% reduction delivered with no advance notice. According to pricing tracker PricePulse, bases that suddenly exceeded the new limit became read-only until the owner deleted records or upgraded. Imagine waking up to find your business data frozen.

That wasn't an isolated move. The pattern over the past two years tells a consistent story:

• Automation runs halved — free-plan automation runs dropped from 100 to 50 per month in Q1 2025
• Entry price doubled — the $10/editor Plus plan was discontinued in Q3 2025, replaced by Team at $20/editor/month with no grandfathering for existing customers
• Free records cut — 1,200 down to 1,000 per base in February 2026, without notice
• Steep upgrade steps — moving from Team to Business more than doubles the per-user price for a 2.5x record increase

The Plus plan discontinuation is especially telling. A 100% price increase on the entry tier, applied to loyal customers with no legacy rate protection, effectively ended Airtable's run as the affordable spreadsheet-database hybrid — and pushed many small teams toward alternatives like NocoDB, Notion, and Google Sheets, as PricePulse's pricing history documents.

Zoom out, and Airtable is riding a broader SaaS inflation wave rather than bucking it. The same tracker shows Airtable's overall pricing up roughly 7%, alongside ClickUp Unlimited at +43%, Monday.com Pro at +19%, and Typeform at +67%. The lesson for any growing business: the price you sign up at is rarely the price you pay in year three.

On September 4, 2026, Bending Spoons completed its acquisition of Airtable. As No Code MBA's migration guide notes, acquisitions of this kind introduce real uncertainty around future pricing, support quality, and product direction — none of which customers control.

That uncertainty matters more because of Airtable's lock-in problem. Leaving means untangling broken automations, messy linked records, permissions, and team retraining — a migration cost that grows with every month you stay. When the vendor can change pricing without notice and ownership has just changed hands, that exit cost becomes a genuine business risk, not a theoretical one.

This is the deeper issue with building operations on rented SaaS: you don't own the roadmap, the pricing, or increasingly the company itself. It's one reason we at AI Business Sites built our platform around ownership — clients own their code, content, data, and domain outright, on one flat monthly price, month-to-month. If we ever stopped earning the business, everything leaves with the client. That's a structural guarantee no acquisition can rewrite.

The practical takeaway: budget for the pricing curve you're likely to face, not the one on today's landing page — and weigh exit costs as heavily as features before committing your operations to any platform.

The hardest part of leaving Airtable isn't exporting a CSV — it's everything that breaks on the way out. Automations stop working, linked records tangle, and permissions need rebuilding from scratch, according to a detailed migration guide that lists broken automations, messy linked records, permission mistakes, and retraining among the core costs of switching.

Lock-in is a structural feature, not an accident. Airtable is cloud-only with no self-hosting option, as Softr's comparison notes — your data lives in Airtable's infrastructure, on Airtable's uptime, and on Airtable's roadmap. A practitioner account adds that you can't index columns for performance or co-locate data with your application servers, which is why senior engineers increasingly shift toward self-hosted alternatives like Baserow or NocoDB when performance and full data ownership start to matter.

The migration checklist is longer than most teams expect:

• Keeping linked-record relationships intact while cleaning duplicates
• Moving attachments — storage that ranges from 1 GB on Free to 20 GB on Team plans per base
• Rebuilding automations from scratch, since the 50 automations per base and 25 actions per automation don't translate to other tools
• Recreating permission roles and retraining the team on a new interface

Then there's the app-layer gap. Airtable is database-first, not app-first — if you want to put an authenticated, user-facing front end on your data, you need a separate layer like Softr, and even then, complex client portals may need dedicated configuration. Airtable's own Portals offering is a paid add-on. Each layer adds cost, complexity, and another vendor relationship to manage.

The acquisition by Bending Spoons, announced completed September 4, 2026, adds product-direction uncertainty on top — pricing, support, and roadmap decisions now rest with new ownership. For a growing business, the practical takeaway is to plan the exit before you need it: keep data structures clean and document your automations early.

This is also why ownership-first models are gaining traction. At AI Business Sites, we treat exit as a feature — clients own their code, content, and data, and can leave with a full export. The easier it is to walk away, the harder a vendor has to work to keep you.

Airtable isn't a bad tool — it's a tool with a well-defined comfort zone. The question isn't whether to use it, but whether your use case still fits inside that zone.

The consensus across practitioners is clear: Airtable works beautifully for focused teams, internal operations, rapid prototyping, and non-critical workflows. As one lead cloud architect puts it, "Airtable wasn't the villain; our use case had just outgrown its happy path." His rule of thumb: stay with it under roughly 25,000 records.

The problems start when the base stops being a convenience and becomes load-bearing. InfoLobby's analysis frames the trigger precisely: "Hitting a number is not the reason to move. The reason to move is when the base has quietly become operational infrastructure."

Here's a practical decision framework. Airtable is likely still right for you if:

• Your bases stay well under the 50,000-record Team cap, and the work is internal — tracking, planning, lightweight ops
• Your API usage is modest (remember: 5 requests per second per base, on every plan including Enterprise)
• Your team is small enough that per-editor pricing hasn't become a line item you flinch at
• Nothing customer-facing depends on it being up and fast

Start planning your exit when any of these become true: the base powers a customer-facing app, you need high-throughput API access, you need full data ownership (Airtable is cloud-only with no self-hosting option), or your seat count keeps climbing. On that last point, the math matters — Team runs $20/editor/month and Business $45, and as Softr notes, "once more people across operations, sales, or client-facing teams need to work inside Airtable, seat-based pricing becomes cost-prohibitive."

There's also a strategic reason to plan early rather than late: migration is genuinely painful. No Code MBA's migration guide catalogs the real costs — broken automations, messy linked records, permission mistakes, and retraining. The teams that exit cleanly are the ones that kept clean data structures and documented their automations before they needed to leave.

A useful self-test: if Airtable went down tomorrow, would you be inconvenienced — or would your business stop? If it's the second, you're not using a productivity tool anymore. You're running infrastructure on a platform with a hard 5-requests-per-second ceiling and a pricing history that includes doubling its entry tier from $10 to $20 with no grandfathering.

This is exactly the inflection point where many growing businesses reconsider the DIY-stack approach altogether. Rather than stitching together a database, a CRM, a website, and follow-up tools — each with its own caps and per-seat fees — some owners hand the whole system to a managed provider. At AI Business Sites, for example, the website, CRM, automations, and lead follow-up come as one managed system with one flat monthly price, and the client owns all the data and code outright. That model isn't for everyone, but it's worth considering the moment you realize your spreadsheet has quietly become your operations department.