Is AI a Bubble?

Random words:

Feeling deep loneliness on New York City’s bustling streets. When I was in small town, I’ve never said goodbye such frequently.

Back to the topic:

When we talk about investment, we talk about economic values. Current situation of AI is very similar to Cisco’s in 2000. Cisco as an internet company spread the capacity of the World Wide Web, but sooner people realized that there is no economic value in internet company, instead, opportunities are in e-commerce etc. AI is a tool very similar to web tech. Currently, with heightened expectations, people are allocating investments and capital expenditure in AI model development, however, end-user demand is unclear and revenue is relatively minimal. This situation makes AI look like a bubble from a very long term perspective.

Stepping closer to it, there is still room in the market to party. GPUs for training and inference are increasingly on demand. First round of beneficiaries are Cloud and Ad. Second round could be hardware or something else. Although it looks like a Capitalism’s scam which is getting more money to the big tech, as small open-source models are released, moats are expected to be disintegrated and distributed. I’ve seen more and more enterprises going to have Gen AI integrated to their business or operation now. Enterprise is going to be continuously transformed to be more efficient and productive, as well as human life with this long lasting attention on AI. This kind of long lasting attention and consistent innovation are something different from internet tech in 2000 and will probably create a momentum against bubble.