Analysis
The Information published an analysis arguing that the divide between AI-native software companies gaining market share and legacy SaaS incumbents losing it is becoming sharper and easier to identify than it was even a few months ago, according to The Information. The piece's core argument: early in the AI product cycle, it was difficult to distinguish companies genuinely benefiting from AI-driven efficiency and new-feature demand from those merely adding an AI veneer to an existing product -- that distinction is now showing up clearly in growth rates, retention and pricing power.
Canva's own revenue forecast cut earlier this month is a live example of the dynamic on the loser side of that ledger: a company that built genuine AI features discovered the cost of serving them at scale was eating into the growth those features were supposed to drive, forcing a downward revision even as usage grew. On the winner side, companies building AI-native from the ground up -- rather than retrofitting AI onto an existing product -- are reportedly showing cleaner unit economics because they didn't inherit legacy infrastructure costs designed for a pre-AI product.
The practical signal for software investors is a shift in diligence emphasis: growth rate alone is no longer a sufficient signal of whether a SaaS company is an AI winner, because AI feature adoption can drive usage growth while simultaneously destroying margin if the underlying serving costs aren't managed. The more informative metric, per the analysis, is whether a company's AI-feature costs are declining as a share of revenue over time, or merely being subsidized by venture or public-market capital while the company chases usage numbers.
The analysis is directional, not a definitive ranking -- The Information doesn't name a comprehensive scorecard of winners and losers, and "AI-native" companies with cleaner unit economics today are also mostly younger and smaller, meaning they haven't yet been tested at the scale where Canva's cost problem actually surfaced. It's possible today's AI-native winners simply haven't hit the user-volume inflection point yet where their own AI serving costs become the same margin problem Canva is now managing.
