Inside AI’s $10B+ Capital Flywheel — Martin Casado & Sarah Wang of a16z

AI funding blurs venture and growth stages

Startups raise billion-dollar rounds pre-monetization, requiring growth-scale capital and complex financial analysis despite being early-stage founder bets.

Compute negotiations now happen at inception

Founders must negotiate equity-for-compute arrangements and strategic partnerships worth hundreds of millions within months of starting, unlike traditional seed-stage deals.

Dollars translate directly to capability

Unlike the dot-com bubble's unused fiber, scaling laws allow tracing capital investments directly to model improvements, ensuring real demand exists for the infrastructure.

Frontier models may consume the entire stack

These companies can raise three times more capital each round than their entire app ecosystem combined, potentially outspending and absorbing all layers built atop them.

Infrastructure and application lines are dissolving

Model companies operate simultaneously as horizontal platforms and vertical applications, breaking traditional assumptions about value accrual at different stack layers.

Tension between research and revenue

Founders face critical resource allocation decisions between spending limited GPUs on long-term AGI research versus near-term product development that generates revenue to fund that research.

Unprecedented founder mobility and talent wars

AI founders face $5 billion poaching offers and intense public scrutiny creating a 'fishbowl effect,' resulting in turnover rates unseen since the early semiconductor era.