Jason’s Takes on This Week’s 20VC x SaaStr: Pro Rata Fights, the First Real AI Price Drop, and Why $500M ARR Isn’t Safe Anymore

We are watching the structural decay of the old software-as-a-service model in real time. If you have been paying attention to the recent shifts in how venture capital is behaving and how AI pricing is cratering, you know the script has flipped. The assumptions we built companies on five years ago are not just outdated; they are liabilities.

The Pro Rata War and the Death of Handshakes

Lately, there has been a lot of noise regarding pro rata rights and investors getting cut out of follow-on rounds. For the uninitiated, pro rata is the right to maintain your percentage of ownership in a company by participating in future funding rounds. It used to be a standard, gentlemanly agreement. If you supported a founder when nobody else would, they remembered it.

But the market is colder now. We are seeing early investors who bridged companies through the dark times getting shoved aside when a hot lead investor wants the whole pie. My take for founders is simple: if you are going to cut someone out, realize you are burning a bridge that took years to build. For investors, if you get cut out, say your piece once, then move on. Whining about it publicly just makes you look like you are not winning enough elsewhere. The reality is that loyalty in this environment is becoming a luxury most founders feel they cannot afford when a Tier-1 firm demands 20% of the cap table.

The AI Price War Has Arrived

For the last eighteen months, everyone has been high on the potential of AI. But we are finally seeing the first real price drop. Look at models like Claude 3.5 Haiku. The cost of intelligence is falling faster than the cost of storage did in the early 2000s. We are witnessing a race to the bottom that should terrify anyone building a wrapper and excite anyone building a platform.

When the underlying utility—the LLM—becomes a commodity, the premium people are willing to pay for your 'AI features' evaporates. You cannot charge a 50% markup on something that costs your competitor pennies. Builders need to stop selling 'AI' and start selling 'outcome.' If your business model depends on arbitrage between API costs and user subscriptions, your margin is currently being hunted by the providers themselves.

Why $500M ARR is No Longer a Fortress

There was a time when hitting $100 million in Annual Recurring Revenue (ARR) meant you were untouchable. Then that number moved to $500 million. The theory was that at half a billion dollars in revenue, your churn is manageable, your brand is a default choice, and your sales engine is a permanent moat.

AI has changed the math. The incumbents—the companies doing $500M to $1B ARR—are now the most vulnerable targets. Why? Because they are bloated. They have thousands of employees maintaining legacy codebases and managing manual processes that AI can now automate for a fraction of the cost. A lean team of ten engineers with the right agents can now build a functional competitor to a specific vertical of a Salesforce or an Airtable in months, not years.

The moat is not the revenue; it is the agility. A company doing $500M ARR is like a tanker ship. It takes miles to turn. A startup is a speedboat. When the tech stack shifts as violently as it just did, the tanker is stuck on the old course while the speedboat pivots to the new reality of low-cost, high-velocity software.

The Decacorn Trap

We are also seeing a shift in how we value these 'Decacorns'—companies valued at $10 billion or more. In the zero-interest-rate environment, we valued them on growth at all costs. Today, the market is asking a harder question: if AI makes your core product 10x cheaper to produce, why should I pay you a 20x multiple on your current revenue?

Builders need to realize that the 'valuation' is a vanity metric that can become a prison. If you raised at a $10B valuation but your growth slows because a cheaper AI alternative is eating your mid-market customers, you are in a down-round trap. You cannot hire, you cannot exit, and you cannot pivot easily.

What This Means for Builders

If you are starting something today, do not aim to be the next 'safe' $500M ARR incumbent. The goal should be to build a high-margin, low-headcount machine that solves a problem so specific that a generic LLM can't replace it.

• Focus on Proprietary Data: If the model is cheap, the data you feed it is the only thing with value.
• Ignore the Hype Cycles: Pricing will continue to drop. Build your margins around the assumption that AI calls will eventually be free.
• Protect Your Reputation: Even in a cutthroat market, how you treat your early backers matters. The market cycles, but people remember who stayed in the foxhole with them.

The era of comfortable software is over. We are entering the era of hyper-efficient, high-utility tools. If you aren't terrified by how fast prices are dropping, you aren't paying attention. But if you are building with a lean mindset, this is the best time in history to take a shot at the giants.

The Takeaway

Revenue is no longer a proxy for safety. In a world where AI drives the cost of production to near-zero, the only real moat is deep integration into a customer's workflow and a brand that people actually trust. The giants are bleeding, and the price drops are just the beginning.

Read the original at SaaStr →