Silicon Valley venture capital firm a16z founder Marc Andreessen recently stated in an interview video that AI is a civilization revolution surpassing the scale of the internet and comparable to the “invention of electricity.” Currently, AI companies are experiencing unprecedented revenue surges, but the industry is undergoing a dynamic transformation of “large models leading, small models catching up,” evolving into a national competition between the US and China. He also observed that although the public is panicked about AI replacing jobs, data shows people are fully embracing AI in their daily lives. Finally, he shared that deliberately maintaining an outspoken, even controversial, company image is to attract entrepreneurs who are equally brave and have clear stances, in order to maintain industry leadership.
The scale and nature of the AI revolution: a civilization leap spanning 80 years
Andreesen pointed out that AI is the biggest technological revolution he has seen in his lifetime, with influence far beyond the internet, comparable to electricity, microprocessors, and even the invention of the wheel. Reflecting on the 1930s emergence of computer science, he noted a key debate at the time: one faction advocated building logic and arithmetic-based “adding machines” (represented by IBM’s predecessor), while another attempted to simulate human neural networks.
Ultimately, due to technological and hardware limitations at the time, the computer route dominated for 80 years, turning computers into cold, mathematical machines. Neural networks became marginal academic pursuits. It wasn’t until three years ago, with the advent of ChatGPT, that this 80-year dormant “artificial cognition” path officially returned to the spotlight. Andreessen emphasized that this wave of growth is entirely different from previous tech bubbles; it is a real revenue explosion, with customer demand directly translating into bank deposits. He bluntly stated, “This fills the missing soul of the computer industry over the past 80 years, allowing machines to finally communicate with us in a human way. This growth rate is unprecedented in my career.”
Business models and market trends: AI as downloadable intelligence
On the commercial application front, Andreessen analyzed that AI possesses a unique “carrier characteristic.” Unlike the internet revolution, which required decades to lay physical infrastructure like fiber optics and base stations, AI is embedded in the existing global network, allowing everyone to access the latest technology through downloads. He advocates for a “pay-as-you-go” (Tokens by the drink) model, which is a liberation for startups: early-stage companies no longer need to bear huge fixed capital expenditures, and developers can call upon top global intelligence as easily as buying a drink.
Andreessen further predicts that the price of AI will fall much faster than Moore’s Law, representing an ongoing “super-deflation” phenomenon. He offered a unique insight into the GPU market anxiety: in a capitalist market, any physically replicable resource driven by huge profits will eventually lead to overcapacity. He is confident that, “As long as physically replicable things experience shortages, history shows they will eventually shift to surplus, and unit costs will fall like stones hitting the ground over the next decade.” This means intelligence will shift from an expensive scarce resource to a cheap public good with infinite demand elasticity.
Technological trends: large models lead, small models catch up rapidly
The technological trajectory is showing a dynamic cycle of “big models paving the way, small models harvesting.” While the top-tier “God Models” continue to push the ceiling of intelligence, small models are rapidly compressing performance. Andreessen predicts that future industry structure will be pyramid-shaped: at the top are a few massive cloud-based neural hubs, with billions of embedded specialized small models beneath, integrated into various physical devices.
The rise of small models will free AI from expensive cloud services. Andreessen believes that NVIDIA’s current high profits are a “signal of attack” to global competitors, calling for all forces to join in and democratize computing power. He observed a startling shift: “For example, China’s Kimmy model demonstrates reasoning abilities comparable to GPT-5 but can run on just two MacBooks. This is very attractive to enterprises because they can choose to run locally instead of paying high cloud fees.” This decentralization of intelligence will trigger unprecedented application layer explosions.
(a16z Outlook 2026: Startups remain key drivers of consumer AI, LLM segmentation taking shape
US-China Competition and Regulatory Policies: Leading equals national security
In terms of geopolitics, Andreessen believes that the AI race is essentially a survival battle between the US and China, with Washington realizing that it involves the ultimate foundation of military and economic security. China’s strength in open-source models and robotics supply chains leaves no room for retreat in this competition. However, Andreessen expressed deep concern about internal US policy trends: currently, there are 1,200 AI-related proposals across US states, some of which are very unfriendly to innovation, such as California’s SB1047, which mimics the EU’s suffocating regulatory framework. The EU’s legislation has already stalled local development, with Meta and Apple hesitant to launch new features in Europe.
He believes that excessive regulation prevents nonexistent threats but kills current competitiveness. This is not just a technical issue but a strategic misjudgment. He stated, “SB1047 attempts to impose unlimited downstream responsibility on open-source developers. That’s basically suicide; it’s self-binding in the race.” The federal government must reclaim regulatory sovereignty, or else the US risks self-sabotage in the marathon, handing over its lead advantage to competitors.
Venture capital perspective and personal philosophy
Faced with this chaotic and high-risk revolution, Andreessen shared a16z’s approach: taking “multiple bets” amid uncertainty. He also pointed out an interesting societal disconnect: polls show the public is panicked about AI, but actual behavior indicates people are increasingly dependent on AI. This phenomenon of “verbal opposition, bodily support” reveals the irreversibility of technological adoption.
Andreessen summarized that a16z’s stance of speaking frankly on public issues is to attract brave entrepreneurs. He believes that silence at this critical moment is a form of absence from the future: “Polls show voters are panicked, thinking AI will destroy everything. But when you look at ‘explicit preferences,’ they are all using AI… they use it to code, see doctors, even help date.” In Andreessen’s view, the true winners will be those who see through human nature’s real needs and dare to define their destiny amid the noise.
This article a16z founder: The AI technological revolution surpasses the internet of the past! Revealing industry profit status and regulatory bottlenecks first appeared in Chain News ABMedia.
This page may contain third-party content, which is provided for information purposes only (not representations/warranties) and should not be considered as an endorsement of its views by Gate, nor as financial or professional advice. See Disclaimer for details.
A16z Founder: AI Technology Revolution Is Greater Than the Internet of the Past! Revealing Industry Profitability and Regulatory Bottlenecks
Silicon Valley venture capital firm a16z founder Marc Andreessen recently stated in an interview video that AI is a civilization revolution surpassing the scale of the internet and comparable to the “invention of electricity.” Currently, AI companies are experiencing unprecedented revenue surges, but the industry is undergoing a dynamic transformation of “large models leading, small models catching up,” evolving into a national competition between the US and China. He also observed that although the public is panicked about AI replacing jobs, data shows people are fully embracing AI in their daily lives. Finally, he shared that deliberately maintaining an outspoken, even controversial, company image is to attract entrepreneurs who are equally brave and have clear stances, in order to maintain industry leadership.
The scale and nature of the AI revolution: a civilization leap spanning 80 years
Andreesen pointed out that AI is the biggest technological revolution he has seen in his lifetime, with influence far beyond the internet, comparable to electricity, microprocessors, and even the invention of the wheel. Reflecting on the 1930s emergence of computer science, he noted a key debate at the time: one faction advocated building logic and arithmetic-based “adding machines” (represented by IBM’s predecessor), while another attempted to simulate human neural networks.
Ultimately, due to technological and hardware limitations at the time, the computer route dominated for 80 years, turning computers into cold, mathematical machines. Neural networks became marginal academic pursuits. It wasn’t until three years ago, with the advent of ChatGPT, that this 80-year dormant “artificial cognition” path officially returned to the spotlight. Andreessen emphasized that this wave of growth is entirely different from previous tech bubbles; it is a real revenue explosion, with customer demand directly translating into bank deposits. He bluntly stated, “This fills the missing soul of the computer industry over the past 80 years, allowing machines to finally communicate with us in a human way. This growth rate is unprecedented in my career.”
Business models and market trends: AI as downloadable intelligence
On the commercial application front, Andreessen analyzed that AI possesses a unique “carrier characteristic.” Unlike the internet revolution, which required decades to lay physical infrastructure like fiber optics and base stations, AI is embedded in the existing global network, allowing everyone to access the latest technology through downloads. He advocates for a “pay-as-you-go” (Tokens by the drink) model, which is a liberation for startups: early-stage companies no longer need to bear huge fixed capital expenditures, and developers can call upon top global intelligence as easily as buying a drink.
Andreessen further predicts that the price of AI will fall much faster than Moore’s Law, representing an ongoing “super-deflation” phenomenon. He offered a unique insight into the GPU market anxiety: in a capitalist market, any physically replicable resource driven by huge profits will eventually lead to overcapacity. He is confident that, “As long as physically replicable things experience shortages, history shows they will eventually shift to surplus, and unit costs will fall like stones hitting the ground over the next decade.” This means intelligence will shift from an expensive scarce resource to a cheap public good with infinite demand elasticity.
Technological trends: large models lead, small models catch up rapidly
The technological trajectory is showing a dynamic cycle of “big models paving the way, small models harvesting.” While the top-tier “God Models” continue to push the ceiling of intelligence, small models are rapidly compressing performance. Andreessen predicts that future industry structure will be pyramid-shaped: at the top are a few massive cloud-based neural hubs, with billions of embedded specialized small models beneath, integrated into various physical devices.
The rise of small models will free AI from expensive cloud services. Andreessen believes that NVIDIA’s current high profits are a “signal of attack” to global competitors, calling for all forces to join in and democratize computing power. He observed a startling shift: “For example, China’s Kimmy model demonstrates reasoning abilities comparable to GPT-5 but can run on just two MacBooks. This is very attractive to enterprises because they can choose to run locally instead of paying high cloud fees.” This decentralization of intelligence will trigger unprecedented application layer explosions.
(a16z Outlook 2026: Startups remain key drivers of consumer AI, LLM segmentation taking shape
US-China Competition and Regulatory Policies: Leading equals national security
In terms of geopolitics, Andreessen believes that the AI race is essentially a survival battle between the US and China, with Washington realizing that it involves the ultimate foundation of military and economic security. China’s strength in open-source models and robotics supply chains leaves no room for retreat in this competition. However, Andreessen expressed deep concern about internal US policy trends: currently, there are 1,200 AI-related proposals across US states, some of which are very unfriendly to innovation, such as California’s SB1047, which mimics the EU’s suffocating regulatory framework. The EU’s legislation has already stalled local development, with Meta and Apple hesitant to launch new features in Europe.
He believes that excessive regulation prevents nonexistent threats but kills current competitiveness. This is not just a technical issue but a strategic misjudgment. He stated, “SB1047 attempts to impose unlimited downstream responsibility on open-source developers. That’s basically suicide; it’s self-binding in the race.” The federal government must reclaim regulatory sovereignty, or else the US risks self-sabotage in the marathon, handing over its lead advantage to competitors.
Venture capital perspective and personal philosophy
Faced with this chaotic and high-risk revolution, Andreessen shared a16z’s approach: taking “multiple bets” amid uncertainty. He also pointed out an interesting societal disconnect: polls show the public is panicked about AI, but actual behavior indicates people are increasingly dependent on AI. This phenomenon of “verbal opposition, bodily support” reveals the irreversibility of technological adoption.
Andreessen summarized that a16z’s stance of speaking frankly on public issues is to attract brave entrepreneurs. He believes that silence at this critical moment is a form of absence from the future: “Polls show voters are panicked, thinking AI will destroy everything. But when you look at ‘explicit preferences,’ they are all using AI… they use it to code, see doctors, even help date.” In Andreessen’s view, the true winners will be those who see through human nature’s real needs and dare to define their destiny amid the noise.
This article a16z founder: The AI technological revolution surpasses the internet of the past! Revealing industry profit status and regulatory bottlenecks first appeared in Chain News ABMedia.