Jason Lemkin

AI Business: From -$94M to +40% Margins! Unpacking Costs #shorts

The video dissects how AI startups are turning around disastrous gross margins by slashing inference expenses, highlighting a shift from a -94% margin last year to a positive 40% this year. The speaker attributes the swing to economies of scale in model inference, noting that while margins have improved dramatically, reaching the 50% target will likely require additional time and cost discipline. Projections suggest a possible climb to 70% over the next two years, though the ceiling remains uncertain. “From negative 94% margins last year to positive 40% this year” underscores the rapid financial transformation, while the comment “it may take two years to get to 70%” illustrates the cautious optimism about future profitability. For investors and operators, the trend signals that AI ventures can achieve sustainable cash‑flow positive models, but they must manage inference spend and define scalable operating structures before the market fully rewards them.