Marc Andriessen Says AI Is Moving 10 Times Faster Than Moore's Law. Here's Why That Matters.
Marc Andriessen, co-founder of venture capital giant Andriessen Horowitz (a16z), declared that artificial intelligence (AI) represents the most dramatic technological transition he has ever witnessed, with performance improvements accelerating at a pace that dwarfs previous computing revolutions. Speaking at Applied Intuition's Physical AI Day in Sunnyvale, California, Andriessen emphasized that AI is not merely a short-term phenomenon but the culmination of roughly 80 years of accumulated research finally reaching critical mass .
How Fast Is AI Really Accelerating Compared to Past Technology Shifts?
The speed of AI advancement has fundamentally outpaced historical computing paradigms. Andriessen explained that if Moore's Law, the principle that computing power doubles every 18 months, represented the old standard, AI scaling now operates at a dramatically different velocity.
This means AI performance improvements are occurring roughly 60 times faster than the computing industry's historical baseline ."If Moore's Law was a structure that doubled its performance every 18 months, AI scaling now moves at 10 times a year," stated Marc Andriessen.
Marc Andriessen, Co-founder at Andriessen Horowitz
This acceleration is not happening in isolation. Andriessen noted that the cost of intelligence is falling sharply, transforming AI from an expensive luxury into a basic resource. As demand explodes while costs plummet, a compounding effect emerges. The combination of large-language models (LLMs), which are AI systems trained on vast amounts of text data, expanding into multimodal systems that process images, audio, and text together, and eventually evolving into world models that can understand and predict reality, creates what Andriessen called a "welfare structure" that further accelerates innovation .
Why Is the Shift to Physical AI the Real Game-Changer?
While software-based AI has captured headlines and venture capital attention, Andriessen argued that the genuine revolution will unfold when AI moves beyond digital systems into the physical world. He emphasized that the next stage of the AI revolution depends on implementing "physical AI" that moves and operates in real environments, which could trigger unimaginable changes across industries .
This transition from thinking technology to moving and operating technology represents a fundamental shift in where AI creates value. Rather than remaining confined to data centers and software applications, physical AI will reshape entire sectors.
The focus of technology competition is moving beyond digital systems into real-world industries including manufacturing, logistics, and transportation, where autonomous vehicles, drones, and robots will apply AI capabilities ."The next stage of the AI revolution is the physical world. Over the next five years, there will be a level of innovation and spread similar to software in the real world," explained Marc Andriessen.
Marc Andriessen, Co-founder at Andriessen Horowitz
Ways AI Growth Will Continue Even Without Breakthrough Models
- Distribution of Existing Intelligence: Even if cutting-edge model development stopped immediately, the industry would continue expanding by distributing intelligence that has already been secured across more applications and users.
- Optimization for Smaller Models: Existing AI capabilities can be refined and optimized to run on smaller models and chips, reducing computational requirements while maintaining performance.
- Adaptation Across Industries: AI systems developed for one sector can be adapted and deployed across manufacturing, logistics, transportation, and other industries, creating sustained growth regardless of new breakthroughs.
Andriessen directly challenged recent skepticism about AI's sustainability, warning that betting against this pace of growth is "very dangerous." His comments serve as a rebuttal to the emerging "AI bubble theory," which suggests that current enthusiasm and investment in AI may be unsustainable .
The a16z co-founder's perspective carries significant weight in Silicon Valley. Andriessen previously developed the Mosaic graphical web browser as a student at the University of Illinois and founded Netscape, playing a key role in popularizing the internet. After the dot-com bubble, he established cloud company LoudCloud, which sold to Hewlett-Packard for $1.65 billion. In 2009, he co-founded a16z with Ben Horowitz, and the firm has since become a leading venture capital investor by backing major technology companies including Facebook, Twitter, Airbnb, and GitHub .
The implications of Andriessen's assessment extend beyond venture capital circles. If AI performance truly continues scaling at 10 times annually, and if physical AI deployment accelerates over the next five years as he predicts, the economic and social disruption could rival or exceed previous technological revolutions. The convergence of faster scaling, lower costs, and real-world applications suggests that AI's impact is only beginning to materialize, making the next decade a critical period for businesses, policymakers, and workers preparing for this transition.