AI euphoria is in full swing. The sceptics say that this is a repeat of when the railways were overbuilt . The Panic of 1873 was triggered directly by the bankruptcy of Jay Cooke & Company; a major bank heavily invested in the Northern Pacific Railroad. The subsequent depression destroyed 121 railroads and 18,000 other businesses. Another example by the bears is the bursting of the fiber optic bubble in 2001 which partially led to the dot com collapse. AI, the naysayers lament, is in the same throes of overbuilt speculative frenzy.
I disagree. At least for the next year. The orders for GPUs and network gear are in, and the deployment of AI for inference is only beginning. The staggering investment in hundreds of billions is only matched by demand. MSFT, GOOG, Amazon and Oracle are capacity constrained and will be through 2026 . Oracle’s announcement last week of its a whopping several hundred-billion-dollar backlog shows that enterprise demand for AI is skyrocketing. However, sentiment may change when the four hyperscalers announce capital spending plans in the second half of 2026. Sentiment may change if these large buyers of compute significantly slow 2027 capital spending. Without equivalent replacement buyers such as the government of Saudi Arabia, China, Japan or the EEC, there will be downward pressure on AI stocks. Yet, I am optimistic through mid-2026. The AI story is not dead, but caution is in order.
I’ve invested in most of the beneficiaries of this relentless capital investment . My positions include Nvidia, which is still a sure bet for a 20 % increase next year, suggesting a stock price of about 210. Likewise, My other AI holdings Innodata , Amazon, Google , Lam Research , Arista Networks, Marvell, and Broadcom will also boast at least 20% share price growth next year. Finally, gargantuan, and very stable MSFT will grow by 15% while throwing you a small cash dividend. (What’s not to like with this most important company on the planet?)
In short, mes amis, you must take advantage of this AI boom. A boom based on the most transformational technology ever invented – an intelligence that makes itself more intelligent. An intelligence that is better than most PhDs in any subject, and getting exponentially better. The implications to society are mind-blowing – whether bad or good. There is no turning back, the AI genie has shot out of the bottle perhaps even laughing cynically now that we have let it out of the lamp. (I’ll be writing about AI and Governance in my next post. It speaks of the myriad risks we face with AI.)
I have some non-AI, but technology forward plays that should also grow by at least 20%. These include SiTime, Palo Alto , and Samsara . I also own Trade desk , a cool advertising tech company , but for some reason it has fallen into disfavor on the street. I’ll hold on to this stock just for amusement through next year. Sadly, I’m down 45% on this puzzling holding.
Finally, I should say that the market has gone up nicely since May 2025. So, trim any holdings if they’ve gone up by 20% or more this year. Don’t be a pig. The market is at an all-time high, and valuations remain frothy. The Nasdaq and S&P may go higher, but they could very well drop 30% or more. Greed isn’t becoming. Trim now if you’ve reached that 20% gain threshold in any stock.
It’s rare that such a massive technological change occurs within one’s lifetime. Some enjoyed the marvels of fire, the wheel, or the printing press. We are lucky to have had three in our short spans -- 1) The PC, 2) The internet and 3) AI. So, enjoy the ride. This is special.