Investors made their choice last week, and it wasn’t the usual one. Over the past 12 months, investors have continued to select the hottest names in AI, that includes the companies that are building the infrastructure it takes to power AI, those supplying power, those supplying semiconductors, and more. Investors turned their back on those very stocks last week after years of little to no movement, why?
The Cost of Money & Circular Financing
AI has proven its ability to enhance productivity and do amazing things, but the cost of compute is not cheap, nor is the buildout it takes to power it. Investors were willing to ignore the costs in the first half of 2026, thinking that The Fed would step in and cut interest rates by September’s FOMC meeting, but the opposite happened. The fallout from the U.S.-Iran conflict forced The Fed to raise interest rates instead, taking the entire AI buildout by surprise. With an estimated $1.5 trillion to be spent by large cap tech on the AI buildout, that cost is now set to be more expensive because financing costs are higher.
Moreover, circular financing was a “buzzword” a few months ago, and although the idea has been forgotten, it’s actually become more blatant. There has been four recent different occasions which Nvidia has been directly provided financial support for the very companies that are placing massive orders for its GPUs. Examples? Here:
- Firmus (Australia): After abandoning its planned $5 billion IPO due to weak investor demand, Firmus is reportedly discussing hundreds of millions in additional funding from Nvidia.
- ZanKore (Indonesia): Secured $3.1 billion in financing to purchase Nvidia GPUs, with Nvidia providing credit support through a revenue sharing arrangement.
- Nscale (UK): Raised $3.36 billion in September, with Nvidia participating as an investor to support the company's AI infrastructure expansion.
- GMI Cloud (US): Raised $668 million in financing, including an equity investment from Nvidia, to expand its GPU powered infrastructure.
Think about it…Nvidia sells the GPUs, invests in companies buying those GPUs, and sometimes helps them secure financing. This creates enormous demand for Nvidia's products, but raises an important question: How much of that demand would exist without Nvidia's financial backing? It doesn't necessarily mean demand is artificial, but it creates a highly interconnected system dependent on continued AI spending. The biggest risk is what happens if AI demand fails to justify the billions being invested today. Only time can tell.
Our prediction is that it’ll take 5-7 years for AI to really have a meaningful positive impact on productivity.
Bond Yields Cooled, Why Didn’t Stocks Go Up?
We’ve written multiple analysis in recent weeks suggesting that long term Treasury Yields are putting pressure on growth stocks, but Yields did cool last week, yet growth stocks continued to sell-off. The reason is the very same reason that the tension between the two investments exists. Investors choice to buy long term bonds at 5%+ rather than putting their money in the uncertainty of AI stocks.
Plus, the market is much more than AI and tech. Looking at the market as a whole, we can see that certain sectors are still very much alive and making a comeback. Take these for examples:
- Healthcare ($XLV)
- Staples ($XLP)
- Discretionary ($XLC)
Investors are going after more traditional businesses with predictable revenue, and more importantly, less AI infrastructure costs.
FOMC Meeting this Week
October’s FOMC meeting is coming up, which means The Federal Reserve will vote and decide whether or not to raise, keep, or lower interest rates. Lowering interest rates is out of the picture, and it’s looking more and more like The Fed will raise rates by 25 BPS again. That especially puts pressure on growth (usually tech) stocks and companies are rely on debt to grow. For this reason, our recent watchlists and positions are focused on companies that are less rate sensitive. We’ll update the events as they unfold in real time this week.