If you’re running a startup right now and your headcount is growing in lockstep with your revenue, you might want to sit down for this.
On May 7, Cloudflare cut 20% of its workforce. One thousand one hundred people. The same earnings call where they announced that cut? Revenue was up 34% year over year. Record quarter. And CEO Matthew Prince said something that should make every founder reconsider their hiring plans: internal AI usage at Cloudflare increased by more than 600% in the last three months alone.
This wasn’t a struggling company trimming fat. This was a company growing faster than ever deciding that a fifth of its workforce was doing work that AI could now handle. And they said it out loud, on the earnings call, to Wall Street.
Two days earlier, Coinbase did something similar. Fourteen percent of the company, roughly 700 people, gone. CEO Brian Armstrong sent the memo at 6:55 in the morning. The subject wasn’t “regrettable reductions.” It was a restructuring around what he called “AI-native pods,” including teams of one person directing AI agents. He killed the role of “pure manager” entirely. From now on, Coinbase only wants “player-coaches” who produce work and lead.
The same day as Cloudflare, Upwork cut a quarter of its workforce. CEO Hayden Brown said AI means “smaller, differently resourced teams can make a bigger impact than ever.” This was Upwork’s third major layoff in three years.
Three companies. Three different industries. The same conclusion: AI didn’t just automate tasks. It made entire organizational layers unnecessary.
The Signal Nobody Wants to Name
Here’s what makes this week different from the AI anxiety headlines we’ve been reading for two years.
These aren’t predictions anymore. They’re earnings reports.
Cloudflare didn’t say “we think AI might eventually change how we work.” They said our AI usage went up 600%, so we’re cutting 1,100 roles that are now redundant. The company explicitly stated this was “not a cost-cutting exercise” and not about individual performance. It was about the structure of the company itself becoming obsolete in certain areas.
The market’s response tells you everything. Cloudflare’s stock dropped on the news because investors were surprised by the scale, but the company is trading at a higher forward multiple than it was a year ago. Coinbase stock actually went up after the announcement. Wall Street doesn’t punish companies for cutting headcount when revenue is growing. It rewards them.
And the venture capital world is saying the same thing. VCs are now explicitly telling portfolio companies that the era of the 100-person tech giant is here, and companies that aren’t operating with a lean, AI-augmented model are having a harder time raising their next round. The old playbook of “raise money, hire people, grow headcount alongside revenue” is being actively penalized.
In 2026 alone, 128,270 tech workers have been laid off across 286 companies. That’s over a thousand people per day. And the companies doing the cutting aren’t failing. They’re growing.
What the New Org Chart Actually Looks Like
So what replaces the traditional structure? Based on what Cloudflare, Coinbase, and the VC ecosystem are describing, the pattern looks like this.
First, the management layer collapses. Coinbase didn’t just cut people. It eliminated an entire category of employee: the person whose job is solely to manage other people. Armstrong’s memo was blunt about it. If you manage but don’t produce, you’re out. The replacement is the player-coach: someone who does the work and coordinates a small team simultaneously. That’s a fundamentally different org chart than what most startups are building.
Second, teams shrink to pods. Coinbase is building “AI-native pods” that can include a single person directing multiple AI agents. Think about what that means. A product feature that used to require a PM, two engineers, a designer, and a QA person can now be shipped by one person with the right AI tooling. Your next competitor really is one person with a laptop, and now even public companies are restructuring around that reality.
Third, AI usage becomes an operating metric. Cloudflare’s 600% increase in internal AI usage isn’t just a fun stat for the earnings call. It’s the metric that justified the restructuring. Companies are starting to measure how much work is being done by AI versus humans, and when that ratio tips past a certain point, the human headcount gets adjusted. This will spread to every industry, not just tech.
Fourth, the restructuring is continuous, not one-time. Upwork has done three major cuts in three years. Each one cited AI as a factor. This isn’t a single “rip the band-aid off” moment. It’s a rolling transformation where the org chart gets redrawn every 12 to 18 months as AI capabilities improve. The companies that treat restructuring as a recurring operating rhythm, rather than a crisis, will outperform the ones that try to do it all at once.
Why Most Founders Are Building the Wrong Company
Here’s the part that hits close to home for anyone building a startup right now.
If you raised a seed round in 2024 and hired the way your investors expected, with a traditional engineering team, a product manager, a head of marketing, and a customer success team, you might already be overstaffed for 2026.
The math has changed. AI-native startups are now hitting $100 million in annual recurring revenue with fewer than 100 employees. A few years ago, that milestone required 500 or more people. Revenue per employee has become the metric VCs actually care about, and the companies clearing $1 million ARR per employee are the ones getting funded.
Meta told 16,000 employees that AI was worth more than they were, and the stock went up. Now Cloudflare is doing the same thing with 1,100 people. The pattern is undeniable: the market values companies that can grow revenue while shrinking headcount.
And I want to be honest about why this is uncomfortable. Because it means the traditional startup advice, the stuff about hiring great people being the most important thing, the “your team is your moat” platitudes, all of it needs a massive asterisk in 2026. Your team is still important. But the definition of “team” now includes AI agents, and the optimal team size for a given revenue target just dropped by 3x to 5x.
What a Founder Should Actually Do Right Now
This isn’t about panicking or firing everyone tomorrow. It’s about being intentional.
Audit your org chart against the pod model. Look at every team of four or more people and ask: could this be two people with AI tooling? Not hypothetically, but right now, with tools that exist today. If the answer is yes and you’re not making the change, you’re choosing to be slower and more expensive than your competitors.
Kill the pure manager role before it kills your burn rate. If you have anyone on your team whose primary job is managing other people but not producing work themselves, you have a 2023 org chart. The player-coach model isn’t just a Coinbase thing. It’s the structural answer to how you maintain coordination with fewer humans.
Make AI usage a tracked metric. Cloudflare didn’t wake up one morning and decide to cut 20%. They measured AI usage, saw it grow 600%, and concluded that the work was being done differently. Start measuring how your team uses AI tools now. Code generation, content creation, customer support, data analysis. You need a baseline so you can see when the ratio shifts.
Hire for AI fluency, not just domain expertise. The player-coach model requires people who can direct AI agents, not just do their own jobs well. That’s a different skill set. When you’re hiring, test for it.
Plan for continuous restructuring. If Upwork is on its third AI-driven restructuring in three years, and they’re a public company with HR processes and board oversight, what makes you think your startup will only need to do it once? Build the expectation of rolling structural changes into your operating rhythm.
The Uncomfortable Bottom Line
Cloudflare grew revenue 34% while cutting a fifth of its workforce. Coinbase eliminated the role of pure manager and built one-person AI pods. Upwork restructured for the third time in three years. The market didn’t punish any of them.
This is not a prediction about the future. This is the present. The AI-native operating model has moved from theory to earnings reports.
For founders, the takeaway isn’t that you should fire people. It’s that you should be building a company where the default is small, AI-augmented teams from day one, not hiring up and hoping AI catches up later. The companies that built traditional org charts and are now having to dismantle them are doing it painfully, publicly, and at scale.
You have the advantage of not having a 1,100-person workforce to restructure. Use it.
