The 18x Midas Lister Betting $3B on AI (and calling most of it fake) | Navin Chaddha, Mayfield

<p>Navin Chaddha is the Managing Partner at Mayfield. He’s made the Forbes Midas List 18 times, and thinks a lot of the AI revenue everyone&#39;s chasing right now is fake.</p><p><br></p><p>Mayfield is a 56-year-old firm that backs founders at the paper-and-pencil stage. They’re investing $3 billion into AI, but Navin warns the market is overcapitalized by a factor of 10x.</p><p><br></p><p>We get into what’s actually going on with the $1B+ funding rounds, the $25 trillion of value AI has to justify, the dangers of FOMO, how he separates vibe revenue from real revenue, backing vertical models instead of horizontal ones, why inference will dwarf training, how a startup actually beats a $100 billion incumbent, the people x-ray behind his founder bets, what cricket taught him about running a company, lessons being the last founder to IPO before the Dot Com Crash, what he learned working with Satya Nadella, and the unfinished business still driving him.</p><p><br></p><p>Thanks to this episodes sponsors!</p><p><br></p><p>Numeral: Sales tax on autopilot https://www.numeral.com</p><p>Flex: Premium banking, 60-day credit, 0% APR https://home.flex.one/referral/bananacapital</p><p>Amplitude: AI analytics https://www.amplitude.com</p><p>Merge: Every model, one API https://www.merge.dev/turner</p><p>Monaco: The revenue engine for startups https://www.monaco.com/</p><p><br></p><p>Timestamps:</p><p>0:00 Lumilens: Zero to $3B revenue in 14 months</p><p>0:50 Connecting GPU&#39;s is AI&#39;s next bottleneck</p><p>5:07 Mayfield: investing $3B in AI and semiconductors</p><p>9:39 Where a $1B round actually gets spent</p><p>12:51 The six-layer AI stack, and who needs mega-rounds</p><p>15:00 Why AI is overcapitalized by 10x</p><p>17:47 FOMO is for sheep</p><p>21:13 Vibe revenue vs real revenue</p><p>22:45 Backing vertical models</p><p>24:22 The best firms have one North Star</p><p>27:18 Are semiconductors still cyclical?</p><p>29:10 Why inference will dwarf training</p><p>31:19 What happens after every infra build-out</p><p>36:27 2 billion Gemini users isn&#39;t real AI adoption</p><p>38:21 What a correction does to AI stocks</p><p>40:46 How FOMO pulls VC&#39;s into hot categories</p><p>44:13 What Navin looks for in founders</p><p>50:23 Why &quot;everyone hates this category&quot; can be a buy signal</p><p>54:12 The argument against the cloud everyone got wrong</p><p>57:20 What white-collar work AI teammates will take</p><p>1:01:45 How AI startups beat incumbents</p><p>1:06:54 Why startups die of indigestion</p><p>1:11:49 Mayfield’s secret formula: people-first</p><p>1:19:18 What cricket taught Navin about building companies</p><p>1:23:17 Dropping out of Stanford to start VXtreme</p><p>1:29:22 Blitzscaling to blitz-failing: the last IPO before the Dot-Com Crash</p><p>1:31:25 Joining Mayfield instead of starting a 4th company</p><p>1:33:29 Unfinished business (backing a $1T company)</p><p>1:35:54 Could you tell Satya would run Microsoft?</p><p>1:38:58 Investors he respects, founders he missed</p><p><br></p><p>Referenced</p><p>Mayfield: https://www.mayfield.com/</p><p>Lumilens: https://lumilens.com/</p><p>Lumilens Raises $700M: https://www.wsj.com/tech/startup-raises-700-million-to-replace-data-center-wires-with-light-adc74358?mod=e2twd</p><p>Built to Last by Jim Collins: https://www.amazon.com/s?k=built+to+last+-+jim+collins&amp;adgrpid=186020621003&amp;hvadid=779535177756&amp;hvdev=c&amp;hvexpln=0&amp;hvlocphy=9218885&amp;hvnetw=g&amp;hvocijid=9037174021105194159--&amp;hvqmt=e&amp;hvrand=9037174021105194159&amp;hvtargid=kwd-362242264527&amp;hydadcr=21907_13365950_10662&amp;mcid=f1dd2c5deb5539b7afc6bcdfee5613c8&amp;tag=googhydr-20&amp;ref=pd_sl_4b3f3t1l23_e</p><p><br></p><p>Follow Navin</p><p>LinkedIn: https://www.linkedin.com/in/navinchaddha</p><p><br></p><p>Follow Turner</p><p>Twitter: https://twitter.com/TurnerNovak</p><p>LinkedIn: https://www.linkedin.com/in/turnernovak</p><p><br></p><p>Subscribe to my newsletter to get every episode + the transc