AIWA "The One Thing" #31: Bending Spoons Bought Airtable. Ended the SaaSpocalypse. Hyperagent Didn't Come With It.

While the acquisition of Airtable by Bending Spoons for $1.285 billion feels like the end of the SaaSpocalypse, there’s a second dimension to the story that I find way more interesting.

Before signing, Airtable reorganized.

The Hyperagent business line, an AI-native autonomous agent platform, was transferred out to a separate Delaware company. Hyperagent Inc. is a named party to the purchase agreement.

*link to Alex Konrad's tweet

Bending Spoons bought the database, 500,000 organizations and 80% of the Fortune 100. The $480 million of ARR.

But it didn’t buy the AI agents.

Nobody has said who runs Hyperagent. The filing doesn't need to. Follow the structure, not the statements.

The operating model is the asset.

Airtable was not slow on AI. Refounding declared in June 2025. An OpenAI engineering leader hired as CTO in October. Superagent in January. Hyperagent in February. AI in every plan, including free.

Then it sold.

Notion was founded the same year. More than half its revenue now comes from AI customers. Still independent.

Bending Spoons has been building one thing since 2013. People. Proprietary technology. Proprietary data. AI inside their internal tools since 2019.

Pull requests authored or coauthored by AI went from under 10% to over 90% in twelve months.

Automate. Augment. Rewire.

When it comes to AI, most companies stop at automate. Some reach augment. Bending Spoons only does the third one. Rewire is their core business.

There are two operating models in this story.

A legacy business at scale needs institutional machinery and a fifteen-year horizon. A frontier agent platform needs founder conviction, speed and the ability to pull the future forward.

No company runs both well. So the deal separated them.

But most companies don't get that option. They have to wire the plane while they're flying it.

Hence the enthusiasm around all things Forward Deployed.

That's the market Anthropic and OpenAI funded in May.

*link to AIWA TOT #19

Read those announcements closely and they aren't about frontier labs. Anthropic says the firm will work with mid-sized companies across sectors and that the consortium's network of hundreds of companies is where the deployments start.

*link to Anthropic announcement

Blackstone President Jon Gray was more direct. The plan is to deploy the technology across businesses in their portfolio and beyond. The bottleneck he named wasn't the model. It was the shortage of highly skilled implementation partners.

*link to Blackstone announcement

Ordinary operating businesses. Same problem Airtable had. No way to sell their way out of it.

Getting AI right is not a model problem.

The operating model is the asset and the operating model is human.

On April 22 at Raleigh Durham Startup Week, I asked whether Marc Benioff going headless was the bottom for traditional SaaS.

*link to AIWA TOT #16 featuring RDSW Talk including a link to the RDSW talk deck, slide 56

Four months later, this is the answer. The bottom wasn't a product announcement. It was a price on an operating model.

Whether you’re an AI native startup or an existing enterprise approaching $500 million in ARR, the moat in 2026 is the human operating system around the AI.

*link to AIWA TOT #20

Humans + Machines. Never Humans vs. Machines.