Software companies have only two options left
Andreessen Horowitz (a16z) growth investing partner David George wrote an open letter to software CEOs. Posted on March 23, the piece boils down to a single sentence.
"The comfortable middle ground is over."
The stock market is repricing the software sector. We're no longer in an era where growth rate alone earned a premium valuation. But profitability isn't earning a stable multiple either. His diagnosis: most public software companies are stuck in an awkward in-between.
Why the middle ground is disappearing
Public software companies have already lived through the first half of this transition. Growth rates have bent downward and valuations have compressed. Free cash flow (FCF) has improved and GAAP margins have ticked up slightly.
The problem is they never reached 'true profitability.' The crux George points to is stock-based compensation (SBC). For the past decade, software companies have touted their free-cash-flow margins while leaving stock-based comp out of the math — ignoring the fact that dilution is a real cost borne by shareholders.
Once you include SBC, most companies land in a difficult middle. Too slow to earn a high-growth premium, too dilutive to earn a stable profit multiple.
By the end of next year, he predicts, companies stranded in this middle will look like no-man's land. Growth pressure, ongoing dilution, and multiple compression all arrive at the same time.
Path one: Add 10 points to your revenue growth
Lift annual revenue growth by 10-plus percentage points with AI-native products — within 12 to 18 months.
What matters here is the definition of 'truly AI-native.' It is not bolting a chatbot or copilot interface onto an existing product. It is not slipping AI features into the existing SKU lineup. You have to build entirely new products that can move the whole company's growth rate by 10 points inside 12 months.
The pricing model has to change too. The first cost customers try to cut when they adopt AI is headcount. Seat-based pricing becomes the direct target. The new growth lives in tokens, consumption, automation, outcomes, and machine-driven workflows. If you're not standing on the token path, you're not standing on the fastest-growing line in the customer's budget.
Path two: Reach a true 40% operating margin
If accelerating growth is out of reach, redesign the company to hit a true operating margin — SBC included — of 40% or more. Ideally 50%.
Getting to 40% takes far more than a 10–20% headcount cut. You have to flatten management layers, standardize implementation, minimize custom services, and abolish committees. Where you own the workflow or switching costs are high, raise prices; for long-tail, low-spend accounts, raise the minimum price or let them churn. Every single share you issue has to be recognized as a transfer from shareholders to employees.
George points to Broadcom as a case study. Broadcom's approach — aggressively rebuilding the cost structure of the companies it acquires — is extreme, but a model worth studying.
The 8–10% layoff headline is dead
This is where he is bluntest. The "8% or 10% layoff" headline no longer works, he says. That's a 'weak form' of restructuring. It only tidies the edges of the org chart, and the market reads it as a signal of weakness.
The 'strong form' is redesigning the machine itself. Whether you choose path one or path two, the requirement is identical: become an AI-native company within 12 months — an organization built to maximize the productivity and efficiency of its engineers. AI has to reshape the company, and the cost structure has to change to match.
Find the five people inside your organization
If you've chosen path one, there's a first move to make: find the leader who can actually pull it off.
It's going to be a 12-month 'death march.' You need someone willing to go through the pain alongside you. The good news, he says, is that somewhere in your organization are five people who can create 100 times more value than you ever imagined — and it doesn't matter how junior they are.
Identifying who those five are, explaining the urgency of the moment, and convincing them that this is the once-in-a-lifetime chance to rebuild the company — that is the CEO's first job.
This isn't a market correction
The reason George's letter spread so quickly across the industry is that it isn't simple investment advice — it confronts a structural shift in the software business head-on.
AI is simultaneously changing how software is built, how it's priced, what its cost structure looks like, and how organizations are shaped. The market is setting a new closing price on the seat-based SaaS model. From the customer's side, the first effect of adopting AI is workforce efficiency — which means fewer seats. For a software company to keep the same revenue from the same customer, it has to change the very unit it charges for.
It's hard to read this message as applying to software companies alone. In every industry where AI is fundamentally rewiring the business model, the "comfortable middle" keeps getting narrower.
