In April 1993, the world wide web made its public debut. Seven years later, in 2000, the biggest beneficiary appeared to be Cisco which had pioneered the routers, servers, and software to connect computers on a network. As the arms supplier for the internet race, Cisco hit a peak market cap of $546B, making it more valuable than Microsoft at the time. If we were to inflation adjust that for today, it would be worth about $2 trillion, perhaps more. Cisco was knocked off its pedestal, driven by three primary factors: First, competition poured into the market, reducing prices by as much as 90% or more. Second, the newer servers could manage more data and the software improved such that laptops could act as servers increasing efficiencies of a company’s data center spend. Third, and likely most important, venture dollars stopped flowing and the result was Cisco’s revenue fell by 20% in 2001. The Company never fully recovered. Like Cisco, NVIDIA’s stock has rocketed, fueled by the influx of venture money into AI startups and the market hype around a new technology. Like water flowing downhill, competition enters markets with growth and outsized profit pools. Meta, Microsoft, and Google, are all getting into the chips business. AI as an industry has also gotten much better at cost-efficient compute, using old chips when performance is good enough and developing more cost-efficient techniques for training and fine-tuning. Lastly, we are seeing improved software solutions to manage data and processing. There will be AI winners, and Nvidia has been the biggest to date by miles - and good for them. Some of us remember buying Nvidia chips for our computers so we could play games - that was the graphics processor. The challenge ahead for Nvidia is to keep innovating and stay relevant as peak demand for compute subsides. Cisco got stuck milking its cash cow rather than continuing to innovate and launching product that competes with its core business. We have seen Jensen Huang pivot a few times, and hopefully he can continue to move with the AI market.
Matt Cochard
3mo
Ironically, we all had front row seats to the slow motion car crash at RS. The question is whether it’s really different this time? [quietly buys more puts] #ThanksOldEconomy
Saul Cooperstein
3mo
Itd be interesting to see them overplayed on one chart through Cisco peak.
Brilliant post
Anu Gupta
3mo
Great perspective Steve Sarracino !
Principal @ Inovo.vc ➡️ [email protected]
4moIMO, comparing Cisco and Nvidia is not the case - cisco was doing pretty easy-to-replicate stuff back then. Nvidia on the other hand does chips that consist of thousands of parts combined thx to years of research, so defensibility seems to be strong. Time will tell:)