- The Midas Report
- Posts
- Airtable Rewires Itself to Think Like an AI
Airtable Rewires Itself to Think Like an AI
Is this the business model of the future?
Midas AI
September 01, 2025
Think of us as your AI radar.
Here’s the blips that matter, freshly served and delivered daily, welcome back to The Midas Report.
Did you ever wonder what happens when you let AI takeover an already established big company?
Us too.
Today, you’re in luck because Airtable's founder reshaped the company for AI and here's EXACTLY what changed
When a viral tweet accused Airtable of being “dead,” CEO Howie Liu didn’t ‘clap back’. He wrote code.
Then he rewired the entire company.
Airtable’s full scale reorg wasn’t just about shipping AI features faster. It was a philosophical shift, from classic corporate flowchart to something more like a brain.
So what did this new model do?
Its started by splitting the company into “fast thinking” and “slow thinking” teams, short cycle execution vs. long term investment.
The fast thinkers live on the edge this means things like weekly AI releases, inference experiments, tight feedback loops.
The slow thinkers though build durable systems and infra to support that speed. Both matter but they’re built, staffed, and run very differently.
To lead by example, Liu demoted himself to a builder. He became an “IC CEO,” writing code, testing models, racking up inference bills (he’s reportedly the top inference cost user at Airtable).
At one point, he told teams to cancel meetings for a week just to tinker with AI tools.
Why?
Because moving at AI speed means operating like an AI, fast, decentralized, and sensory. It’s not just about prompt engineering. It’s org engineering.
Many companies are now layering AI onto their products. Fewer are changing their operating systems to allow true AI native behavior. That’s what this is. And it may be the difference between launching one flashy feature and becoming an AI first company.
The results speak for themselves, Airtable is now shipping weekly AI upgrades, gaining product velocity, and reportedly generating over $100 million in free cash flow.
If you’re leading a team right now, the takeaway isn’t to restructure your whole company tomorrow. But carving out a “fast lane” team, even just 2 people, to move fast on AI native workflows is a low cost way to find your own velocity.
Think of it as giving your org a second brain. Just make sure it’s allowed to think fast.
🧠 The Download
Salesforce hits $1B in AI revenue with agentic systems and is no longer a case study, it's a cash engine fuelling the company's fastest growing product line ever.
MIT calls out the AI execution gap as 95% of projects yield zero ROI despite massive hype and investment, only 5% of companies are getting real payoff from AI and most deployments still don’t touch the actual work.
Aviatrix automates 80% of marketing using AI and the numbers are wild. They re architected the workflow using 4 to 5 AI models per employee and slashed execution time across content and video by more than half.
Alibaba stunned markets with a 15% stock surge after revealing booming revenues from its AI cloud services, particularly “custom LLMs” tailored for China’s regulated sectors.
Auspex AI emerged from stealth with a $35M Series A led by a16z and backed by Palantir’s Foundry fund, already securing deals with three U.S. government agencies, firmly planting RegTech in AI’s core terrain.
FAITH token exploded with a 620% spike in 24 hour trading volume after landing on three CEXs and touting its dual AI infrastructure roadmap.
That’s your AI edge for September 1st 2025.
See you tomorrow!
Midas AI