April 13, 2026
AI is making software easier to build and harder to defend. The answer is not a better product roadmap. It is a better margin.
For most of the SaaS era, the defensibility argument was simple: we have millions of lines of code, years of product iteration, and a feature set that would take any competitor years to replicate. That depth was real. It worked.
AI code generation is dismantling it. A small team with the right tools can now approximate in weeks what once took years. The cost of building credible, competitive software has dropped dramatically, and it keeps dropping. According to Gartner, up to 60% of new software code will be AI-generated by 2026. If your moat is the code itself, the moat is shrinking.
This is not abstract. We saw an early signal in late 2024 when Klarna replaced Salesforce CRM with an internally built AI system, one of the first high-profile cases of a major company choosing to build rather than buy. The "SaaSpocalypse" conversation that followed reflects a genuine structural question: if AI agents can do what your software does, and if AI tools can help customers build their own version, what are they actually paying you for?
The answer has to be more than features.
Here is something that rarely gets acknowledged: SaaS has historically been a near-zero operating profit business for most companies. Gross margins are high, often in the 70-80% range. But operating margins, what is left after sales, marketing, R&D, and G&A, have been a different story.
Public SaaS companies operated with median net losses of 8-14% for most of the past three years, and before the post-2022 correction, the picture was worse. Average operating margin across public SaaS reached -31% in 2022. Many companies were burning cash at scale and calling it a growth strategy.
The logic was coherent for its time. If LTV exceeds CAC, keep investing in customer acquisition. Profitability will follow at scale. Venture capital was cheap, interest rates were near zero, and the growth-first playbook kept working as long as multiples held. Sapphire Ventures modeled that a typical high-growth SaaS company, even with strong underlying metrics, takes seven years from its Series B to reach positive operating margin. Seven years. That was considered normal.
That era is over. AI is making the cost of ignoring profitability much higher.
I co-founded HiddenLevers in 2010. We built a risk analytics platform for wealth and asset management firms, and sold the company to Orion Advisor Solutions in 2021. By the time we exited, we were running a 52% pre-tax profit margin.
That was not an accident. From day one, we ran lean. We did not raise rounds that required burning cash to justify the valuation. We built a product customers paid for, kept the team small, and operated at a margin that did not require the market to remain irrational.
Among public SaaS companies, only around 52% had positive EBIT margins at all in 2022, and most of those were at single digits. Reaching a pre-tax margin above 50% is unusual. I am not saying that to brag. I am saying it because I think the playbook is more replicable than people assume, and AI makes it more accessible now than it was when I was building.
The AI era does not just change the threat landscape. It changes the cost structure too, and lean companies are positioned to benefit.
AI tools let smaller teams ship significantly more than before. Code generation, automated testing, AI-assisted customer support, AI-driven marketing: across every function, the output per person is rising. You do not need headcount growth to maintain competitive velocity. A team of ten, using AI across engineering, support, and marketing, can operate at the output of what used to require thirty or forty people.
Companies using AI in operations report better efficiency across R&D and G&A relative to those that do not. If your cost structure shrinks while output holds or grows, margin improves. Hold price steady, and that improvement goes straight to profit. Pass some of it to customers through lower pricing, and you become much harder to displace.
That last move is the one most companies are not thinking about clearly enough.
Your customers now have more viable build-or-buy alternatives than they did two years ago. A scrappy internal team at a mid-size company can use AI tools to vibe-code a reasonable approximation of your product. It may not be as good. It may introduce security risks. But it is increasingly feasible, and some customers will do it.
The response that works is to make the calculus obvious in your favor: better features, lower price, and real margin for yourself. That is not a contradiction. It is the payoff of running lean.
If AI is reducing your cost to deliver by 30-40%, you can cut your price 20% and still improve margins. Customers who might have considered building their own solution now face a simpler math problem. The commercial vendor costs less than their internal effort would, ships faster, and carries no maintenance burden. The threat of vibe-coding their own solution mostly dissolves.
This only works if you have done the internal work first. If you are not using AI aggressively across your cost structure, your costs are excessive by the new standard. Every team still running 2022-era headcount ratios while competitors adopt AI is paying a premium that shows up as margin compression.
My target is 20% operating margins, across every function, as the floor. Not 20% someday at scale. Now.
That number is achievable for a lean team using AI across its operations. It requires applying the same scrutiny to every line item:
Marketing: AI-generated content, automated outreach, and smarter attribution can cut your cost per qualified lead significantly. This is not a future capability. It is available today.
Sales: AI tools can handle qualification, follow-up, and proposal generation with less headcount. For most B2B SaaS products, the math works.
Product and engineering: A smaller team can ship at the same velocity with AI-assisted development. I have seen this firsthand.
Back office: Support, billing, compliance, operations. All of these have meaningful AI-reducible costs if you are willing to redesign the workflow rather than just add a tool on top of the old one.
The 20% target is not arbitrary. It is the margin that gives you a real buffer when the AI era keeps shifting. A company at 5% operating margin has very little room to absorb a major competitive move, a pricing reset, or a customer segment that starts building in-house. A company at 20% can cut price, invest in a new feature, or absorb a down quarter without existential risk.
Profitability also makes you a better acquisition target. HiddenLevers commanded a strong exit in part because a profitable business is easier to value, easier to integrate, and a less risky bet for a buyer. Growth matters, but a buyer is more comfortable acquiring a business that can sustain itself without constant capital infusion.
Fraction helps SaaS teams build and ship with lean, senior, US-based engineers who know how to use AI to move fast. If you want to see what your next build would cost, no calls required.
Free and instant. Try the estimator now.
Doesn't "get profitable" conflict with investing in growth?
Not if you build the margin first. The argument is not to stop investing in product or customer acquisition. It is to eliminate the waste that has nothing to do with growth: bloated headcount in functions AI can handle, pricing that is too high to defend, and operating structures built for a world where venture capital would fund the gap indefinitely. A 20% margin with a strong growth rate is a stronger position than a 0% margin with the same growth rate.
How does AI code generation actually threaten an existing SaaS product?
The threat is not that competitors replicate you overnight. It is that the time and cost required to reach "good enough" has fallen dramatically. A customer who was never going to build their own solution in 2020 might reconsider in 2026. Maintaining a price and feature advantage that makes the commercial product the obvious choice is the practical response.
Is a 52% pre-tax margin realistic for most SaaS companies?
At exit, with a mature product and a lean operating model, yes. Earlier in the company's life, the target should be to avoid the growth-at-all-costs trap and build toward positive operating margins consistently. The 20% floor I recommend is a reasonable starting target for a company that has reached product-market fit. The 52% reflects years of compounding that discipline.
What functions benefit most from AI cost reduction?
R&D and G&A typically show the clearest efficiency gains in early AI adoption, according to SaaS Capital data. Engineering is the most discussed, but customer support, marketing content, sales outreach, and back-office operations all have significant AI-reducible costs. The honest answer is almost every function, if you are willing to redesign workflows rather than just bolt AI on top of existing processes.
Does this advice apply to venture-backed companies, or only bootstrapped ones?
The principle applies broadly. Venture-backed companies have more flexibility on timing, but the market has shifted. Investors now reward efficiency alongside growth. The Rule of 40, which balances growth rate and profitability, has become more important as multiples compress. A venture-backed SaaS company burning cash in 2026 faces different investor expectations than one doing the same in 2021.
If AI makes software easier to build, does that hurt Fraction's model too?
AI makes individual engineers more productive, but it does not eliminate the need for experienced engineers who know how to use those tools well. The gap between what AI generates and what a senior engineer who reviews, tests, and ships production code can deliver is still large. If anything, increased throughput per engineer raises the value of senior talent. What changes is that the same output now requires fewer people, which is exactly what Fraction's model is built for.
Gartner. "By 2026, 60% of new software code will be AI-generated." ProfileTree. https://profiletree.com/vibe-coding/
The SaaS CFO. "The SaaSpocalypse: AI Agents, Vibe Coding, and the Changing Economics of SaaS." https://www.thesaascfo.com/the-saaspocalypse-ai-agents-vibe-coding-and-the-changing-economics-of-saas/
Lighter Capital. "SaaS Gross Margins and How To Increase Yours." https://www.lightercapital.com/blog/saas-gross-margins-and-how-to-increase-yours
Sapphire Ventures. "What SaaS Profitability Looks Like in 2023." https://sapphireventures.com/blog/what-saas-profitability-looks-like-in-2023/
Aventis Advisors. "SaaS Valuation Multiples: 2015-2026." https://aventis-advisors.com/saas-valuation-multiples/
SaaS Capital. "AI Adoption Among Private SaaS Companies and Its Impacts on Spending and Profitability." https://www.saas-capital.com/blog-posts/ai-adoption-among-private-saas-companies-and-its-impacts-on-spending-and-profitability/
Payhawk. "How AI Reshapes SaaS Profitability and the 5 CFO Mandates for 2026." https://payhawk.com/blog/saas-profitability-cfo-mandates-2026
Related: How Much Does It Cost to Build an App? · Software Cost Estimation for Non-Technical Buyers
