Venture capitalists may value a solid $15M revenue company at zero. Their model is not built on backing good businesses, but on funding 'upside options'—companies with the potential for explosive, outlier growth, even if they are currently unprofitable.
There's a strong reluctance in venture capital to fund companies that are number two or three in a category dominated by a "kingmaker"—a startup already backed by a top-tier firm. This creates a powerful, self-fulfilling fundraising moat for the perceived leader, making it unpopular to back competitors.
Relying on the once-golden 'T2D3' growth metric for SaaS companies is now terrible advice for 2025. The market has shifted, and founders with these strong historical metrics are still struggling to get funded, indicating that even elite growth is no longer a guarantee of investment.
Founders can become fixated on achieving a good burn multiple, which is a theoretical measure of fundability. However, they sometimes forget the practical reality: a great burn multiple is useless if the company runs out of cash. Cash in the bank is a material construct, not a theoretical one.
The moat for a market leader isn't just the initial VC investment; it's the subsequent, rapid follow-on rounds that create a 'wall of money.' This forces competitors to prove they can win against not just a brand name, but also a massive and compounding capital advantage.
Private market valuations are benchmarked against public multiples. Currently, public SaaS firms with 30% growth trade at 15-20x revenue, twice the historical average. If this 'bedrock price' reverts to its 7-8x mean, it will trigger a cascade of valuation drops across the private markets.
For over a decade, SaaS products remained relatively unchanged, allowing PE firms to acquire them and profit from high NRR. AI destroys this model. The rate of product change is now unprecedented, meaning products can't be static, introducing a technology risk that PE models are not built for.
The current AI investment climate feels as 'risk-free' as the 2021 bubble. Venture firms are likely using flawed loss-ratio models, underestimating how many AI 'unicorns' will fail to generate returns, just as they did with the B2B SaaS unicorns from the previous cycle.
Unlike traditional SaaS where product-market fit meant a decade of stability, the rapid evolution of AI models makes today's PMF fleeting. Founders face the risk that their product could feel obsolete within a year, requiring constant innovation just to stay relevant in a rapidly changing market.
The burn multiple, a classic SaaS efficiency metric, is losing its reliability. Its underlying assumptions (stable margins, low churn, no CapEx) don't hold for today's fast-growing AI companies, which have variable token costs and massive capital expenditures, potentially hiding major business risks.
With hundreds of unicorns and only about 20 tech IPOs per year, the market has a 30-year backlog. Consolidations between mid-size unicorns, like the potential Fivetran and dbt deal, are a necessary strategy for VCs to create IPO-ready companies and generate much-needed liquidity from their portfolios.
While merging portfolio companies is strategically sound, it's often blocked by investor incentives (e.g., diluting a 20% stake in a winner down to 8%). The process is vastly simplified when a single firm, like Andreessen Horowitz in the Fivetran/dbt case, is a major investor in both companies, which aligns incentives.
While AI-native companies burn cash at alarming rates (e.g., -126% free cash flow), their extreme growth results in superior burn multiples. They generate more ARR per dollar burned than non-AI companies, making them highly attractive capital-efficient investments for VCs despite the high absolute burn.