Brad Gerstner points out that the current impact of AI is mainly on the cost side, achieving ROI through layoffs and efficiency improvements
Brad Gerstner points out that the current impact of AI is mainly on the cost side, achieving ROI through layoffs and efficiency improvements, while shorts ignore the imminent explosive potential on the revenue side.
AI platforms like Anthropic will enter a mission-critical phase; once the revenue side breaks through, companies will find it difficult to remove their embedded applications. NVIDIA is already using AI to design new chips, creating a recursive cycle of "machines making machines."
Gerstner predicts that if Anthropic reaches a $100 billion run rate by the end of 2026, it could grow 3-5 times to $300-500 billion in revenue, an unprecedented scale of expansion.
ABAB AI Insight
As the founder of Altimeter Capital, Brad Gerstner has long invested in technology and has accurately captured multiple cycles. This time, he emphasizes the shift of AI from cost optimization to revenue-driven transformation, continuing his judgment on platform stickiness.
On the capital path, institutions are investing heavily in platforms like Anthropic to build enterprise infrastructure, motivated by the desire to secure long-term competitive advantages, strategically using recursive AI to accelerate chip and model iterations, forming a self-reinforcing loop.
Similar to NVIDIA's transition from GPU to AI full-stack, this event is in the stage where generative AI moves from experimental deployment to core productivity, with Anthropic at the forefront of large model competition.
Essentially, this is a technological substitution: AI replaces and enhances human cognitive labor, with the mechanism of recursive self-optimization breaking traditional ROI cycles, concentrating capital on a few platforms to achieve exponential revenue leaps.
ABAB News · Cognitive Law
Costs decrease first, revenues explode later, and the turning point determines victory or defeat.
The deeper the embedding, the harder the exit; the platform is the moat.
Machines making machines; history has never seen an exponential closed loop.