I remain convinced of my bottleneck thesis regarding Next-Gen Technologies (NGT). The expansion of AI infrastructure is encountering real bottlenecks in memory, photonics, power supply, cooling, packaging, materials, and other areas.
As far as I’m concerned, nothing fundamental has changed about this technological thesis. What has changed is the environment in which it must unfold.
The massive expansion of AI infrastructure requires enormous amounts of capital. Hyperscalers and other companies are financing ever-larger investment programs and are thus competing for funds in the capital market.
At the same time, the U.S. government also requires ever more capital. High and still-rising government debt thus collides with an additional, enormous need for financing in the private sector.
This can create a vicious cycle: High demand for capital can keep yields on the bond market high or drive them even higher. This makes financing more expensive—for the government as well as for companies. And the higher the discount rate, the less corporate profits that lie far in the future are worth today. Growth stocks, in particular, are sensitive to this.
Added to this is inflation. If inflation persists or rises again due to high government spending, tariffs, or other policy measures, the central bank’s leeway to lower interest rates is reduced. In the worst-case scenario, higher key interest rates .
The interesting thing about this is that: The AI narrative can remain fundamentally intact, and yet AI stocks can still come under pressure. Not because AI doesn’t work. Or because there’s a lack of demand. Or because supply bottlenecks are disappearing.
But because capital is becoming more expensive, investments are harder to finance, and future profits are being valued at lower levels.
And then there’s a second risk factor that I hadn’t fully accounted for: the economic and geopolitical policies of the U.S. government.
Tariffsthat are imposed, modified, or rescinded on short notice. Personnel decisions at the central bank and government agencies that call political independence into question. A Foreign policythat unsettles allies and turns supply chains into bargaining chips. To me, this is no longer just the usual political uncertainty that’s always been around. This is unpredictable and risky behavior with a direct impact on the cost of capital, supply chains, and predictability—and I have absolutely no influence over it.
This brings me to Aschenbrenner. His Situational Awarenessfund has also run into trouble because he bet on his AI thesis using leverage. A risk he chose and created himself.
My risk is different. I don’t use leverage. But I’m sitting on a position in NGT that’s relatively large by my standards in a market whose valuation framework is partly determined by a government that I can neither elect nor control nor reliably assess. This isn’t a bad decision that I can correct: it’s powerlessness, that I have to deal with.
For my overall private portfolio , this means I won’t be increasing my NGT exposure any further for the time being. New capital is now flowing more heavily into defensive and diversifying investments (e.g., iShares STOXX Global Select Dividend 100 $ISPA (+0,64%)). I am also reducing some of my weaker or riskier NGT positions slightly.
This is not the end of my NGT strategy, and certainly not the end of my AI thesis. It is risk management against a risk that I cannot simply ignore just because I don’t like it.
For my ScaleLimits wikifolios , nothing about the fundamental strategy is changing. The bottleneck thesis remains intact, and I will continue to actively manage the portfolios according to my NGT methodology, including rebalancing, cluster maintenance, and position selection. A “retreat” would be the opposite of active management. That is precisely why I remain committed and am closely monitoring the environment instead of simply waiting it out.
Because for me, the following still holds true: You don’t have to abandon a good investment thesis just because market conditions have worsened. But you also shouldn’t pretend that the risk hasn’t changed. And you should honestly identify where that risk comes from.
Does this worry you, too?
