AI and Dot-com Economics

Abstract:

The advent of the internet inspired a wave of entrepreneurs to build fully online "virtual shops," where middlemen expected to profit from the spread between retail prices and the cost of goods and delivery -- both outsourced to third parties. Between 1995 and 2000, the NASDAQ surged more than 600% fueled by consumers investing far beyond their means in technologies they barely understood. As expected, the bubble burst. By 2002, the NASDAQ had collapsed 78% as most dot-com companies failed to convert hype into real earnings.

Yet a small handful of companies not only survived -- they became giants.

Fast forward to late 2022. The release of ChatGPT shocked consumers and reignited dreams of instant success, this time under the banner of artificial intelligence. But chatbots are largely a decoy: they entertain the public while the true customers of AI are enterprise corporations building automation, analytics, and infrastructure at scale. 

The pattern is identical: a breakthrough technology, a sudden surge of public fascination, a flood of speculative capital, and a brutal sorting of winners and losers. Once again, investors are captivated by the mystery and promise of a new technological frontier. And once again, many are investing without understanding what they are buying. The outcome for them is likely to mirror the dot-com era: heavy losses for the majority, extraordinary gains for a select few. 

However, unlike the dot-com era, today's AI boom has has a clear kingmaker: NVIDIA, whose hardware and vertically architected software define the pace at which the entire industry can grow. Every hyperscaler, every model lab, every enterprise deployment is constrained by NVIDIA's roadmap. The company is not merely supplying tools -- it is setting the boundary conditions of the AI economy.

History does not repeat, but it rhymes -- and the rhyme is getting louder. A small number of companies will rise to become Mega-corporations, entities with market valuations rivaling Nation-States and Trade Blocs, ultimately shaping a new tri-layer global economic structure. 

Bio:

Dr. Gargini received a doctorate in Electrical Engineering in 1970 and a doctorate in Physics in 1975 both with full marks and honors. In the 70s he was a researcher at Stanford University and at Fairchild Camera and Instrument in Palo Alto. In 1980, Dr. Gargini was appointed as the first manager of MPU technology at Intel (80286 and 80386). In 1985 he headed the first submicron team at Intel.  In 1992 he led the team that developed and made manufacturable the Flip Chip packaging technology that is still widely used nowadays. In 1996, he became Director of Technology Strategy and responsible for worldwide consortia research until 2012 when he retired from Intel.  During his tenure at Intel, he was a member of Sematech, SRC, EIDEC, ASET and SIA Boards, Chairman of the I300I consortium and NRI. He was one of the initiators of the EUVLLC. He is presently co-chairman of the EUVL Symposium.  Dr. Gargini contributed to the formation of IMEC core partners program and became the chairman of its scientific advisory board. Dr. Gargini led the industry-wide conversion to 300mm wafers as Chairman of the I300I multi-national consortium, under his guidance fully automated semiconductor wafer handling was developed, standardized and adopted world-wide. From 1998 to 2015, Dr. Gargini was Chairman of the ITRS sponsored by the WSC.  In 2016 he became Chairman of the IRDS sponsored by IEEE Computer Society with the support of 8 IEEE Societies and Councils (i.e., CSC, EDS, CASS, Magnetics, Signal Processing, Pels, CEDA and SSCS) and with full membership participation of Europe (SiNANO) and Japan (SDRJ).

Dr. Gargini was inducted in the VLSI Research Hall of Fame in 2009, IEEE Fellow in 2009, I-JSAP Fellow in 2014 and IEEE Life-Fellow in 2020.