The world's only digital futurist & speaker - talk to my digital twin | living and breathing cutting-edge technologies to inspire Fortune 500 companies | Global innovation keynote speaker | TED | 5x author
Has the AI boom turned into a colossal financial pitfall? ➡️ The AI industry, once the darling of tech investors, now faces a stark reality check. What started as a $200B question has ballooned into a $600B dilemma. Despite Nvidia's meteoric rise, there's a massive gap between investment in AI infrastructure and actual revenue. ➡️ While Nvidia's new B100 chip promises better performance, the fundamental issue remains: where is the return on this staggering capital expenditure? ➡️ The oversupply of GPUs, lack of pricing power, and rapid depreciation of technology are turning the AI gold rush into a precarious gamble, according to Sequoia. ❓ As we navigate this speculative frenzy, are we focusing too much on hardware and not enough on delivering real value to end users? Read the full story on Sequoia Capital: https://lnkd.in/gwWnGzGf #Economics #GenAI #Investments #GPUs #Capital ---- 💡 𝗜𝗳 𝘆𝗼𝘂 𝗲𝗻𝗷𝗼𝘆𝗲𝗱 𝘁𝗵𝗶𝘀 𝗰𝗼𝗻𝘁𝗲𝗻𝘁, 𝗯𝗲 𝘀𝘂𝗿𝗲 𝘁𝗼 𝗱𝗼𝘄𝗻𝗹𝗼𝗮𝗱 𝗺𝘆 𝗻𝗲𝘄 𝗮𝗽𝗽 𝗳𝗼𝗿 𝗮 𝘂𝗻𝗶𝗾𝘂𝗲 𝗲𝘅𝗽𝗲𝗿𝗶𝗲𝗻𝗰𝗲 𝗯𝗲𝘆𝗼𝗻𝗱 𝘆𝗼𝘂𝗿 𝘁𝗿𝗮𝗱𝗶𝘁𝗶𝗼𝗻𝗮𝗹 𝗻𝗲𝘄𝘀𝗹𝗲𝘁𝘁𝗲𝗿 - you can have real-time insights, recommendations (a lot more than I share here) and conversations with my digital twin via text, audio or video in 28 languages! Join >5000 users who went before and go to app.thedigitalspeaker.com to sign up and take our connection to the next level! 🚀
Mark - I wouldn’t be dictating my response to you on an iPhone while walking to see the sunrise in the country at this moment if not for the billions poured into the tech that became the.com bubble. Back then, kids in garages were being funded (often foolishly) on the hype and the hope and the fumes for startup ideas that had some potential but little tech infrastructureto make their idea into reality. Back then,the Internet was mostly something academics and the military used. Without the billions from the Ma and Pa investors, the Internet that became Webb 1.0 and then 2.0 would’ve grown but not nearly as quickly. Yes, a lot of people lost the whole lot of money, but that infusion of private cash brought us to this moment on the edge of Web 3.0 and the AR, VR, haptic, holographic, “help me Obi-Wan Kenobi you’re my only hope” post-screen future. The AI boom is at a similar inflection point. And the big investors along with the mom and Pa investors are trying to catch the wave and with returns better than their bonds and CDs. Yes, there will be losses and crashes. But AI has something that the .com days never had. Blockchain. And together what will they not transform?
Mark Brown
2w
Believe we are clearly now in the post peak hype phase for the use of the latest marketing term “AI” to describe essentially any software running on any hardware, effectively making the term “AI” meaningless and there are so many examples of truly epic fails. Here is evidence of the level of hype: https://www.oralb.ca/en-ca/product-collections/genius-x
The balloon is probably bigger than $600B. Sequoia’s napkin analysis uses a 50% capex. I could argue that GPU driven AI centric data centers can be more around 65% with perhaps 5% labor costs vs traditional 10%. That pushes the revenue gap envelope further on H100 projections, with another hefty shift with B100 close behind. Although hyperacalers expect the fastest growth (CAGR over 13%), enterprise data centers represent 44% of market. Keeping the CPU vs GPU pricing in mind, it will be interesting how soon the premiums (real vs inflated expectations) will revert/collapse.
Delivering ‘real value’ to end users was never the plan. The plan was, to harvest the knowledge of the ‘end users’, and then displace them by undercutting them in the market, effectively selling workers skills to their former employers for 30% of the value of that function. Now, when *that* much yield is forecast our army of little rat brained VC’s try and use the only tools at their disposal, the ‘loss leader’, regulatory capture and network effect. What we have now, is a significant proportion of all global investment being locked up in long runways and technical advances coming so quickly no one can actually figure out how to safely and reliably implement Ai in real applications because the professional body of knowledge is so far behind this week’s state of the art, that makes real investors (who care about working products), very reluctant to proceed. VC guys, whose only true interests are monopoly,enclosure, arbitrage and theft will keep happily playing that game of musical chairs till it’s inevitably ugly conclusion.