@FeitengLi: a16z's latest podcast is titled "AI, the Internet in 1997, and What Happens Next," featuring guest Benedict Evans—former a16z partner, author of the annual must-read "AI Eats the World" for countless investors and executives, widely regarded as one of the most widely read analysts in tech.

X AI KOLs Timeline News
podcast ai-investment economics historical-analogy infrastructure industry-analysis

Summary

a16z podcast invites former partner Benedict Evans to explore the analogy between AI and the internet in 1997, pointing out that current AI infrastructure investment is massive but ROI is unclear. Historical experience suggests value will shift upward, and models themselves will find it difficult to achieve differentiated profits.

a16z's latest podcast is titled "AI, the Internet in 1997, and What Happens Next," featuring guest Benedict Evans—former a16z partner, author of the annual must-read "AI Eats the World" for countless investors and executives, widely regarded as one of the most widely read analysts in tech. As the title suggests, the tone is set: over the course of an hour, he systematically dismantles almost everyone's current certainty about AI: Pay $20 a month, yet burn through $10,000 worth of tokens; someone else toils for days and gets hit with a $10,000 bill, completely stunned. This isn't a bug. It's a script that played out once before in 2009. He sets the tone with an analogy: "We're in 1997, and I'm trying to predict Uber and Airbnb." In 1997, you could see the web and search, but not ride-hailing and short-term rentals—those took another decade, waiting for smartphones and payments to be ready. So today, betting on which AI app will win likely means you're still betting on "the web itself." So what is he certain about? He's certain that history will rhyme. Return to that $20 paradox—it's exactly what happened in 2009 when AT&T offered unlimited data plans for the iPhone. What we should really look at is the outcome: · Mobile data traffic grew about 1,500 to 2,000 times · Telecom stocks have been flat for two decades · "All the cool stuff was done by others"—because value all shifted upward So where is the money burning this round? This year, the combined capex guidance for Microsoft, Meta, Google, and Amazon totals around $700 billion. Evans casually draws a comparison: that's close to the entire global oil and gas industry's annual capex. They know the ROI is unclear but still spend, because for these giants, it's a "life-or-death" level of FOMO—you can't let others do it while you sit out, otherwise you become the Microsoft of the 2000s or the IBM of the 90s. But he also draws a line: "We can't spend $10 trillion a year on AI infrastructure, because there simply isn't $10 trillion to spend." Can the models themselves make money? He doesn't assert, but lays out a chain of reasoning inviting rebuttal: frontier models have no network effects, no differentiation, and no leverage over the tech stack. Then he drives it home with a profit comparison: the combined profits of Google, Meta, Amazon, Microsoft, and Apple exceed the entire telecom industry. He closes with a 1950s IBM ad: a group of engineers holding slide rules, captioned "One IBM electronic calculator equals 150 additional engineers." Every generation of technology thinks it's unprecedented. And his judgment on all this is just one sentence: "It will be magic. Twenty years from now, we'll just say: of course it was like this." Full bilingual transcript + chapter summary + subtitles: https://lattifai.com/zh/podcasts/a16z/ktl8mNiWqMM?dub=zh…
Original Article
View Cached Full Text

Cached at: 06/17/26, 08:03 PM

The title of this a16z podcast is “AI, the Internet in 1997, and What Happens Next.” The guest is Benedict Evans — former a16z partner, whose annual “AI Eats the World” report is a must-read for countless investors and executives, and who is widely considered one of the most-read analysts in tech.

The tone is clear from the title: over the course of an hour, he systematically dismantles just about every certainty people currently hold about AI.

Pay $20 a month, but burn through $10,000 worth of tokens in compute. On the other side, someone tinkers for a few days and gets a $10,000 bill — completely stunned.

This isn’t a bug. It’s the exact same script that played out once before, in 2009.

The framing he gives the current moment is a single analogy:

“We’re in 1997, and I’m trying to predict Uber and Airbnb.”

In 1997 you could see the web and search, but you couldn’t see ride-hailing or short-term rentals — those took another decade, waiting for smartphones and payments to be built out first. So if you’re betting on which AI application will win today, you’re mostly still betting on “the web itself.”

So what is he certain about? He’s certain that history will rhyme.

Back to that $20 paradox — it’s exactly what happened in 2009 when AT&T offered unlimited data plans for the iPhone. And the real lesson is in how it ended:

  • Mobile data traffic grew roughly 1,500 to 2,000 times
  • Carrier stock prices barely moved in twenty years
  • “All the cool stuff was built by someone else” — because all the value migrated up the stack

So where is the money burning this time? This year, the combined capital expenditure guidance for Microsoft, Meta, Google, and Amazon totals roughly $700 billion. Evans casually drops a comparison: that’s close to the entire global oil and gas industry’s annual capex.

They’re spending despite unclear ROI because for these big tech companies, it’s existential-level FOMO — you can’t let others do it while you sit out, or you become Microsoft in the 2000s, or IBM in the 90s. But he also draws a line: “We’re not going to spend $10 trillion a year on AI infrastructure, because there simply isn’t $10 trillion to spend.”

Can the models themselves make money? He doesn’t declare a verdict, but lays out a chain of reasoning and invites rebuttal: frontier models have no network effects, no differentiation, no leverage over the stack. Then he drives it home with a profit comparison — the combined profits of Google, Meta, Amazon, Microsoft, and Apple exceed the entire telecom industry.

He closes with a 1950s IBM ad: a group of engineers holding slide rules, with the caption “One IBM electronic calculator equals 150 additional engineers.” Every generation of technology feels like nothing before it. And his only judgment on all of this is a single line:

“It will be magic. Twenty years from now we’ll just say: Of course it was always going to be this way.”

Full bilingual transcript + chapter summaries + captions: https://lattifai.com/zh/podcasts/a16z/ktl8mNiWqMM?dub=zh…

Benedict Evans: The Economics of AI Usage and the Future of SaaS | lattifai.com

Source: https://lattifai.com/zh/podcasts/a16z/ktl8mNiWqMM?dub=zh LAI〉LAI〉

LattifAI

LattifAI: Next-Gen AI Audio/Video Processing Agent

Similar Articles

@oran_ge: https://x.com/oran_ge/status/2116337749171360048

X AI KOLs Timeline

A Chinese-language in-depth opinion piece exploring whether AI has truly created incremental value. Drawing on the recent sharp selloff in US equities (Intel down 28.6%, Google down 12.6%, Reddit down 29%) alongside rising hardware prices, collapsing token prices, and disappearing traffic, the author argues that AI is destroying value by making intelligence no longer scarce, and believes we are on the eve of an AI bubble bursting, with companies like ByteDance and Microsoft reinvesting nearly all of their profits into AI infrastructure.

@PandaTalk8: Recommend following @lexfridman. Lex Fridman has interviewed a large number of top founders, chief scientists, and KOLs in the AI field. These conversations are often deep, 2-5 hour brainstorming sessions, basically making this the most cutting-edge AI thought podcast platform, bar none. As of 2026…

X AI KOLs Timeline

Recommend following Lex Fridman's podcast, which features in-depth interviews with many top figures in AI, covering major AI companies such as OpenAI, Anthropic, Google DeepMind, and is considered the most cutting-edge podcast platform for AI thought.

@WSInsights: https://x.com/WSInsights/status/2052986400740638991

X AI KOLs Timeline

A Chinese analysis article covering Sequoia Capital's 2026 AI Ascent closed-door summit, summarizing key insights from attendees including Demis Hassabis, Andrej Karpathy, and Greg Brockman: AGI has arrived, 2026 is the year of Agents, AI will reshape white-collar work, and a 6-step action plan for ordinary people to adapt.

@leslieloser_: Had the privilege of meeting @Zhm20220917, the best at AI transformation in Jiangsu, Zhejiang, and Shanghai, for a few hours. Became even more certain about the following --In the AI era, those closer to production who understand the industry will reap huge startup dividends; understanding AI and boundaries is 20%, understanding production and industry is 80% --Small teams refuse inno…

X AI KOLs Timeline

The article shares insights on entrepreneurial dividends in the AI era, emphasizing that understanding industry and production is more critical than mastering AI technology. Companies prioritize actual problem-solving capabilities over the models themselves.

@VincentLogic: After a two-year hiatus, this blogger came back with a bombshell. He broke down the entire AI industry chain into 12 layers, from the bottom-most energy and chips all the way to the future "AI-native economic ecosystem." This video is worth watching repeatedly, especially the final definition of "AI Native companies" – it's very insightful.

X AI KOLs Timeline

Blogger VincentLogic released a video that deconstructs the AI industry chain into 12 layers, from energy and chips to the AI-native economic ecosystem, providing a systematic analytical framework.