Years from now, this letter may be studied the way we study Bezos's early shareholder letters — a founding document of how AI-native companies think, operate, and win.
This week, Brian Armstrong published an inflection point statement for the AI age. Coinbase is cutting 700 roles — 14% of its workforce. The official framing points to a down market. But the more honest reading is different: Coinbase had built an organizational structure that was simply too large for what modern technology demands. AI didn't create that problem. It just made it impossible to ignore any longer.
This isn't a layoff announcement. It's a doctrine.
"No pure managers: Every leader at Coinbase must also be a strong and active individual contributor. Managers should be like player-coaches, getting their hands dirty alongside their teams."
This is not a new idea — it's the oldest idea in business, finally coming back. We already see this operating model at the very top: Jensen Huang runs NVIDIA with no direct HR reports and 60 people reporting to him directly. Elon Musk codes, engineers, and ships alongside his teams. The pattern is consistent: deep domain expertise combined with laser-focused value creation. The era of the professional manager — the person who coordinates without creating — is closing. AI is the forcing function that makes that closing inevitable and rapid.
"Fewer layers, faster decisions: We are flattening our org structure to 5 layers max below CEO/COO. Layers slow things down and create coordination tax. The future is small, high context teams that can move quickly."
This is first-principles design applied to the corporation itself. Every layer in an org chart was originally justified by information scarcity — managers existed to relay context that couldn't travel faster. AI eliminates that scarcity. When everyone has full context instantly, hierarchy becomes pure friction.
The most disruptive force in business today isn't a product or a technology — it's the combination of first-principles organizational redesign with the guardrails of existing law and regulation. The companies that get there first, within compliance, will define the next decade.
"Why now? Two forces are converging at the same time."
Armstrong names the down market and the AI acceleration. But there is a third force he places in the background — and it is the one the broader blockchain and crypto industry has been building toward for years: the convergence of AI with decentralized infrastructure. This isn't Coinbase's story alone. It is the story of an entire industry that quietly built the rails the AI economy will run on.
AI agents are not just running on servers. They are settling on blockchains. They are transacting in stablecoins. The reason is structural: agents need payment rails that are programmable, permissionless, and instant — traditional banking infrastructure cannot serve non-human principals at machine speed.
Stablecoins like USX, Solstice's Solana-native synthetic dollar, are purpose-built for exactly this: optimized speed, cryptographic security, and on-chain transparency for agentic transactions.
The two tsunamis Armstrong describes — AI capability and market cycles — are actually three. The third wave is the settlement layer for the AI economy. And it runs on-chain.
A critical footnote
Armstrong's letter is a compelling doctrine. But the honest counterpoint is this: the benchmark has already moved further than Coinbase is admitting.
Tether, the largest stablecoin issuer in the world with over $140B in circulation, operates with roughly 100 employees and generates more profit per head than almost any company on earth. Hyperliquid — a fully on-chain perpetuals exchange — runs at a ~$10.5B market cap with 11 to 12 people. Coinbase sits at ~$51B with over 4,000 employees even after the cuts. The ratio is not comparable. It is not even in the same conversation.
This is not an outlier. Across crypto-native infrastructure, the most capital-efficient companies were never built with the legacy headcount model in the first place. They didn't need to flatten five layers of management because they never had them. They didn't need to rediscover the player-coach model because their founders never stopped being engineers.
Armstrong's letter is a meaningful step. But it is also, in part, an admission: that Coinbase spent years scaling a corporate structure that crypto-native challengers never adopted. The lean, AI-native, high-context team isn't where Coinbase is going. For some of its competitors, it's where they started.
The question for every established company reading this letter isn't just whether to flatten the org chart. It's whether you can move fast enough to close the gap on teams that were born flat.
The question isn't whether AI will reshape your company. It's whether you're doing the reshaping — or waiting to be reshaped. The wave doesn't negotiate.
→ Full letter: https://x.com/brian_armstrong/status/2051616759145185723