The Moat Wasn't Breached. The Water Evaporated.

On 4 August 2026, Bending Spoons agreed to acquire Airtable at an enterprise value of $1.285B — around $2.25B in equity value once Airtable's net cash is counted.

At its 2021 peak, Airtable was valued at more than $11B. Earlier this year its shares were trading on secondary markets at roughly $4B.

And it is not a broken company. Roughly $480M ARR as of June 2026, growing north of 20% year over year. By any operating measure you would have used in 2019, that is a good business having a good year. The multiple tells the story better than the headline: about 2.7× revenue, against something closer to 71× at the peak.

So what erased close to $9B of valuation?

Not churn. Not a competitor. Nothing that shows up in an operating review.

Capability inflation

Airtable's premise was that building software is hard, so don't — configure a flexible database with a friendly interface instead, and get 80% of a custom application without an engineering team.

That premise was worth $11B when building software was hard.

Then it got cheap. Not incrementally cheaper — cheap enough that the thing Airtable existed to help you avoid became something you could just do. The gap between "configure a no-code tool" and "generate the actual application" narrowed until the tool's core value proposition was competing against a genuinely viable alternative that did not exist when the valuation was set.

The moat was never breached. Nobody built a better Airtable and took the market — the revenue is still growing 20% a year. The water evaporated. The difficulty that made the product valuable stopped being difficult, and the moat became a ditch with a very good product sitting in it.

This is worth separating from the usual failure narrative, because the numbers do not support one. Investors are not walking away empty-handed; $2.25B of equity value against $1.4B raised is a return, not a wipeout. What collapsed was not the company. It was the multiple — the market's belief about how much future the business had.

This was not a bad bet

It matters to say this clearly, because the narrative will harden into "overvalued no-code company gets its comeuppance," and that reading is both wrong and useless.

In 2021 the Airtable thesis was obviously correct. Software was expensive to build, engineers were scarce and expensive to hire, and the demand for internal tools vastly exceeded the supply of people who could write them. A product that let a non-engineer close that gap was solving a real, large, durable-looking problem. Investors were not being stupid. They were pricing the world as it existed, and the world as it existed made that price defensible.

The asset did not fail. The ground repriced.

That distinction is the whole lesson. If you read this as a company failure, you will look for the operational mistake and find nothing, because there wasn't one. If you read it as a repricing, you start asking the question that actually generalizes: what else is priced on a difficulty that is about to stop existing?

Every SaaS valuation is a wager

Strip a software business down and there is a bet at the bottom of it: that the problem it solves stays expensive to solve.

That bet has been reliable for twenty years. Problems stayed hard. Building stayed slow. The abstraction layer you sold kept its value because writing the thing underneath it remained out of reach for your customer. The whole category was built on the durability of difficulty.

AI is now solving problems more cheaply, faster than companies can reprice themselves. And repricing is slow — it requires repositioning a product, retraining a sales motion, and convincing a board that the thing that worked needs to change while the revenue still looks fine. Airtable's revenue still looks fine. That is precisely why the repricing happened in the acquisition rather than in the strategy.

Airtable will not be the last. It is the cleanest paper trail — a public number, a clear premise, and an unambiguous gap between what the business was once believed to be worth and what someone will now pay for it, with healthy operating metrics on both sides of that gap to rule out the easy explanations.

The question worth sitting with

The instinct is to ask which companies are exposed. That is the wrong question, because the answer is a list you cannot act on.

The useful version points inward. What are you paying for right now that exists purely because something used to be hard?

Go through the line items. For each one, ask what it is actually abstracting away, and whether that thing is still expensive. Some will hold — a lot of software is valuable because it holds state, enforces process, or carries compliance weight, none of which gets cheap when generation gets cheap. Others are pure difficulty arbitrage, and their price is a bet you are now on the wrong side of.

Do that exercise before your renewal, not after someone else's acquisition makes it obvious.

What abstraction layer are you paying for today that an agent replaces tomorrow?