- Markets Mirage
- Posts
- One week / one topic: Tokenmaxxing
One week / one topic: Tokenmaxxing
Through the mansions of glory in suicide machines
What happened?
Like it or not, AI capex is now a macro variable – not ‘just’ an equity theme.
Alphabet, Amazon, Microsoft and Meta reported earnings within minutes of each other and confirmed that demand for compute is very real.
If anything, it is supply that is constrained. The hyperscalers are essentially sold out.
The proposed solution? Together, these four companies are spending ~ $600bn a year to fund the AI arms race.

Source: Bloomberg. Data as of 01/05/2026
Despite the cheap talk of ‘bubble’, investors remained quite discerning and rewarded platforms where AI investment translated into visible revenue growth or backlog (Google Cloud, AWS), while penalising those where spend rose faster than monetisation (notably Meta).
In other words, equity dispersion widened as the market focus shifted from “spend” to “returns”.

Source: Bloomberg. Data as of 17/04/2026. Past performance is not a guide to future performance.
This reinforces a broader regime change in equities: AI is no longer a uniform beta trade but increasingly a stock‑ and sub‑sector‑specific call.
That said, the combined impact of very concentrated equity indices, truly massive investments and second-order effects is enough to meaningfully affect the macro picture.
AI development is currently investment‑intensive, not efficiency‑driven: more power, more chips, more physical infrastructure.
This combination supports higher real activity but also adds pressure to energy, commodities and long‑end rates – helping explain the persistent disconnect between buoyant equities and stubbornly elevated bond yields.

Source: Bloomberg. Data as of 01/05/2026. Past performance is not a guide to future performance.
We sit therefore in an increasingly uncomfortable equilibrium.
Yes, every day that the Strait of Hormuz remains closed adds inevitable pressure to energy supply and delays to infrastructure repair.
At the same time, we are in the middle of a major technological revolution – or perhaps just at the beginning.
How can we square the two, then? Should we even try?
Our observations
Fundamentals: We seem to have forgotten that 15 years of US Tech outperformance were boosted by massive share buybacks.
Now that these very same companies are instead opting for enormous capital expenditures – as opposed to returning cash to shareholders – opining on their ‘fair valuation’ has become much more difficult.
Price action: Media reports love headlines that say “X billions of dollars vaporized today” whenever there is a market crash.
On the other hand, one company (Alphabet) added $440bn of market cap in one day on Thursday. Either way, we are playing for keeps here.
Investor beliefs: Hyperscalers results gave equities free rein to maintain their cheerful disposition.
However, ever tighter oil inventories and rising government bond yields loom ominous. Time for that much-famed ‘executive put’, maybe?

Bloomberg. Data as of 01/05/2026
So what?
While there are many important ongoing developments, it feels like a good time to exercise restraint when making portfolio decisions.
Fast-moving prices can tempt you to go for ‘hero trades’, but not making big mistakes can be as important over the long run.
With all that in mind, we stick with our conviction positions in the context of a balanced risk exposure.
The day will come for taking larger bets, as it always has in the past.
Mood music: Bruce Springsteen – Born to Run
By popular demand, here is the One week / One topic playlist
The information provided should not be considered a recommendation to purchase or sell any particular security.