My "blogging" was in hibernation last 8 months due to my self-imposed restraint given the environment as well as built-in inertia to get started despite so many interesting events and markets reaching all time highs after taking a big dump around "Liberation Day" in Apr...Around that time I had the blog ready that it would be repeat of Mar/Apr 2020 panic and recovery during onset of Covid Pandemic. The hunch happened to be correct and I was glad that I could keep and take some positions which I am still holding especially around AI theme.
But that was then...as 2025 is about to wrap up in 10+ weeks, let's look at what's in store for rest of 2025 and 2026. And what's better time than to start writing again just before one of the most important week on the calendar with multiple key events coming up next week...
- Fed meeting to decide the course of interest rates - it's almost guaranteed that Fed will cut rates by 25 basis points (2nd time in 2025) and could also indicate another cut in Dec due to weakened labor conditions despite economy running slightly hotter to Fed's comfort. But given that Fed is coming to realization that with inflation "3% is new 2%" (I wrote about this about 2 years ago)...
- Earnings from big tech players of "MMAANG" group - MMAANG = Microsoft, Meta, Apple, Amazon, Nvidia, Google. (it's play on hindi word माँग means demand or request). The markets demand a strong earnings from these companies which most probably they will deliver. On top of their earnings, markets would be looking for their Capex spending plans for 2026 and even more importantly how they are able to convert these huge outlays into actual revenues and earnings. And this is where "token" economy comes in play - more on this later
- USA and China trade discussions with Nov 1st deadline...today morning news are saying that treasury secretary Bessent is making progress and 100% tariffs on China due to their blockade of export of rare earth materials are on hold. So cooling of these tariffs wars between two biggest economies should help continue market's "hot streak"
As markets are reaching all time highs, many pundits and newspapers have started publishing articles about "AI bubble" similar to articles in 1998/1999 about dot.com bubble...which eventually did burst in 2000-2001.
So let's look at "Will 2026 Echo 1998's Boom or 2000's Bust"?
At end of 2024, I wrote about "2025 = is it going to be 1997 or 2000" and 2025 is turning out to be like 1997. Hopefully many of my readers invested in ETFs like XLK, IGV, SMH etc which handily beat S&P - SMH is up nearly 47% in first 10 months. So logically 2026 should be like 1998.. would it be?
There are key differences between Internet heydays days and today's AI enthusiasm...
- Click vs Token pricing model - Internet boom was based on "clicks" with no clear plans on how to convert "clicks" into $ because end consumer was not directly paying for clicks - companies had to depend on users clicking on ads and getting paid. AI boom is based on "tokens" which is getting converted into real $ with subscriptions ranging from $20 per month to $200 per month for some high-end enterprise plans.
- The business model behind companies laying out fiber plants to connect the websites were flawed with no immediate revenue conversion..Hence most of these companies like Global Crossing etc went belly-up. Compared to these, the neo-clouds and Cloud titans investment in buildout of AI infrastructure is based on actual "token" consumption instead of "all you can consume" model. So as long as "token" consumption continue to grow exponentially, the business model behind these large Capex investments will work
- Demand - Here is where we could take some lessons from Internet days - even in depth of busting of internet bubble, internet traffic was 30+ times larger than when bubble was forming
- 1998: Global internet traffic was approximately 12 petabytes (PB) per month.
- 2002: Global internet traffic reached approximately 405 PB per month.
- No one would doubt that token consumption would be multi-fold higher in 4 years from now than it is today given adoption of AI in our daily lives as well as small and large businesses to improve their productivity, revenues and earnings.
- There is no froth in IPO market like internet bubble days when stocks of flimsy companies with no revenues were doubling on opening day and more later. In fact most of the recent IPOs such as Klarna (KLAR), FIGMA (FIG) have billion $ plus revenues and still markets are skeptical about their valuations - they are trading to close to their IPO prices after first day enthusiasm. This is healthy
We can keep going on factors why so called AI bubble may be different or we may still be in early cycle. Hence 2026 would feel more like continuation of 2024/25 with higher volatility the way Internet driven markets were in 1998/1999.
So what should one do with their portfolios? - simple advice: Stay invested and do systematic time based (weekly/monthly) investments in ETFs focused on secular trend of AI, Energy and productivity boosting industries while also allocating some capital for dividend producing ETFs (I personally prefer JEPI/JEPQ)
I am reading latest book by Andrew Ross Sorkin titled "1929"..it's well timed and well written book. Still early in my reading so not ready to provide analogy to Roaring 20s of 100 years back which led to biggest market crash of 1929...
In meanwhile - Happy Diwali (delayed) and Happy Halloween!
/Shyam
Comments