A month ago, I wrote about Airtable’s sale to Bending Spoons. It seemed to me that Airtable was the first of the good businesses to take the hit from the 2021 ZIRP era. I ended the article by asking who was next, and whether Bending Spoons was already on the phone.
Five weeks later they’d agreed to buy Miro for $1.355bn, a 92% discount to its 2021 valuation. Sounds brutal, but public SaaS on the BVP cloud index currently trades at a median of around 4-5x revenues, with median growth rate ~18%. Miro was growing 5%, roughly a quarter of that, 2.3x ARR feels about right to me.
So, Bending Spoons were on the phone and it’s working its way through the maze of overpriced 2021 unicorns like a very patient game of Pac-Man.
Who is next then…
Half the universe is frozen
According to PitchBook’s latest Global Unicorn Tracker, almost 40% of the unicorn universe has not been revalued in over 3 years, and roughly half has not been revalued in 2 years. So, there are a lot of stale marks out there.
The universe looks healthy overall, but that number is carried by a dozen AI names being repriced constantly. Strip those out and most of the market is dormant, with companies valued at a 2021 or 2022 valuation.
An interesting data point is down rounds. There were six in the first half of 2026, versus 31 in all of 2025 and 34 in 2024. You could mistake that for a recovery on the surface, but a company actually has to raise a down round, and these companies just aren’t raising at all. PitchBook suggest the collapse in down rounds is more of a reflection of the market’s silence than its health.
Chart 1 – Down round activity
The correction
The repricing of the 2021 vintage was obviously always coming. This might have come as a down round where you were marked to reality. But it seems that companies are avoiding doing that, so for the companies that eventually have to face the market, the correction might now be an exit.
The same data shows exits reopening. If you strip out the xAI deal, strategic buyers still took roughly $50bn of unicorn value in the first half of 2026, against $18.9bn in all of 2025, across fewer but far larger deals. If you are a founder who raised at 2021 prices and has been waiting on the IPO window, a sale might be available in a way it probably wasn’t in 2023. The IPO route barely exists for these companies anyway, with one listing, SpaceX, 95% of all unicorn exit value in Q2.
Miro, though, is in a slightly different position. All the above explains the cornered companies running low on cash or needing to grow, staring down a down round and taking a sale instead. But Miro had $435m of cash, was profitable, didn’t need to raise and so didn’t face a forced down round. It could have sat on its $17.5bn mark and coasted around the maze.
Instead, Miro let itself be caught by selling, probably realising that coasting on is just slow death rather than a stable state.
The Miro problem
Miro was pretty solid. The business had cash and was profitable. If I am honest, I never really loved the product, but I didn’t use it often. It’s used by 99% of the Fortune 100, with 750+ customers paying over $100k a year, so what do I know. It responded to AI, rebuilding around the Intelligent Canvas, shipping agents, launching an MCP server with Anthropic and Google and buying Reforge.
However, growth had fallen to around 5% YoY, against a $17.5bn valuation set at ~30x ARR in January 2022 that it was never going to grow back into. Miro’s own sales reps describe the forward view on RepVue claiming renewals hard, new logos close to impossible, the product increasingly seen as “just a whiteboard” against free alternatives and the org focused on churn mitigation rather than expansion.
This puts Miro in an interesting dilemma. The business was stuck, which in some ways is worse than failing, because failing companies have obvious moves and stuck ones don’t. A stuck company probably needs ruthlessness the people running it usually don’t have.
Reinvent
The first is to reinvent (aka doing an Intercom). Cannibalise the old model, change what the business is, and climb back above the growth line where the market pays for trajectory. Intercom is the one of few companies that did this properly, sacrificing around $60m of legacy seat revenue and shifting most of R&D onto its AI product while it was still small. It worked, with Salesforce agreeing to buy Fin for about $3.6bn. Plenty of companies repriced their AI features but few have gone to the extent of Intercom. But by reinventing, you are betting the company, which is tough in a founder-led company.
Miro did respond to AI, but it didn’t move the growth rate. The people who need a whiteboard have one, and the ones who don’t are drifting to free tools that are good enough. AI features are probably not going to create demand that isn’t there. Miro hasn’t been disrupted out of business (at least not yet) it was just a mature business that had stopped being able to win new customers.
This is where the Thoma Bravo framework I leaned on writing about Airtable comes in handy. A product in critical workflows, sitting on proprietary data or protected by regulation is durable. But a broad, general-purpose tool the customer could replace without much blowing up is not. I put Airtable on the exposed end and Miro sits in the same spot. There’s no regulatory moat, no proprietary data the customer can’t take with them, no mission-critical workflow that breaks on exit. Run out of growth on the durable end and you still end up with a defensible business. But if you run out on the exposed end, you get Miro.
Cut
The second way out is to cut. You stop chasing the growth, cut hard, and run the good business lean for cash. It’s just as ruthless as reinventing, and it’s something a founder probably wants to avoid. It means taking your category-leading business from 1,600 people to a few hundred and running your life’s work as a cash machine. I think these companies can be fixed, but not from the inside.
Miro did neither, so it stayed stuck. This is where Bending Spoons comes in. Bending Spoons is ruthless on the cutting side and is happy to take bigger risks with its companies because it’s one of fifty across a portfolio. Evernote is a good example. It was a stalled product taken from 340 odd people to around 60 in a year and reportedly down to 20 since. Phil Libin built Evernote and arguably would never have done that to it. It took an owner with no attachment to take this level of risk and ruthlessness.
The people who could block a sale at a massive haircut are the late-stage investors, and at a 2.3x ARR exit they probably don’t care about the headline, because their liquidation preference gets their money back regardless. ICONIQ led Miro’s Series C at $17.5bn and likely gets its capital back through the pref stack. Some holders are rolling proceeds into new Bending Spoons equity which will likely offer more upside than Miro can.
So, who’s next
Think of Bending Spoons as playing a patient game of Pac-Man, working through the maze one company at a time. The 2021 cohort has spent years dodging reality, but Airtable and Miro are the two that stopped dodging and let themselves get caught. And as I said about Airtable, that’s probably the healthy outcome, the mark was never coming back, and the talent can go and build something else.
So, who follows? Airtable and Miro were both good businesses, stuck below the growth line. They both added AI capabilities but couldn’t shake the whiff of danger. Both companies were never getting back to their old marks, on a base that wasn’t going to hold and run by founders who weren’t going to reinvent and couldn’t bring themselves to gut the place.
A company such as ClickUp would be a candidate. It last raised at a $4bn mark in 2021, untested since. It sits on a similar exposed, replaceable base to Miro, no moat, nothing an agent or a free rival can’t chip away at.
Growth looks like it’s still running at maybe 20%, so higher than Miro, but the data is patchy and 20% isn’t 2021 growth. If growth holds, ClickUp maybe keeps escaping Pac-Man, but Airtable was growing at a similar rate.
Bending Spoons only does three to five deals a year, but says it’s identified over a 1,000 potential targets, worth something like $400bn of revenue between them. With nearly half the unicorn universe frozen on marks nobody’s tested in years, I don’t think it’ll be long until we find out who is next!
Sources: