/ October 06 / Weekly Preview
-
Monday:
N/A
---
Tuesday:
Various Fed Speakers
---
Wednesday:
FOMC Minutes
---
Thursday:
Initial Jobless Claims (223K exp.)
---
Friday:
Michigan Consumer Sentiment
-
Monday:
Constellation Brands Inc
Aehr Test Systems
---
Tuesday:McCormick & Company, Incorporated
---
Wednesday:
N/A
---
Thursday:
Pepsico, Inc.
Delta Air Lines, Inc.
---
Friday:
N/A
Are we in an AI Bubble?
Got a question about the markets, your investments, or a topic you’d like us to discuss in an upcoming article? We read every message and may feature your question in our daily write-ups!
Email: andrei@signal-sigma.com
Follow & DM on X: @signal_sigma
Markets closed Q3 on a high note, choosing to brush off uncertainty related to the Government shutdown, as well as Jerome Powell’s comments related to equities being “fairly highly valued”. The shutdown also caused an economic data blackout, as the Bureau of Labor Statistics halted the release of several influential reports, including the jobs report from last week. Investors are now navigating without the usual bearings, but one thing is for certain related to shutdown events: the market usually ends up higher 85% of the time according to Carson Research:
The Nasdaq (QQQ) ended the week up +1.21%, as mega-caps, especially AI related, continued their upward trajectory. The markets have been higher for five consecutive months, a very long streak by historical standards.
Overwhelmingly positive sentiment managed to overshadow fears related to valuations, and we’ll explore both the bull and the bear case in today’s newsletter. Since valuations are not a useful market timing tool, we can count on… something else to disrupt the rally, if it comes to that.
Goldman’s top traders expect October to be more difficult. Their desk commentary indicates that although overall flows continue to be positive, there is significant concentration in technology and momentum stocks, which makes the market susceptible to even minor shocks. Their Sectors outlook reflects a cautious stance on Consumer Discretionary (XLY) and Industrials (XLI), as these sectors appear exposed to slowing economic growth. Conversely, Energy (XLE) and healthcare (XLV) are expected to perform relatively better.
Medium-Term options positioning from market makers somewhat reflect Goldman’s analysis, with many sectors facing adverse risk-reward equations, especially Transports (XTN) - closely related to Discretionary spending.
Overlaying this outlook with the seasonality for October, we can spot a large variance of potential outcomes, ranging from +15% to -10% by month end (+1% on average). Yet the second half of October also tends to align with the start of the seasonally strong period of the year, as Q3 reporting season starts and supports asset prices. Corporate buybacks also return at the end of the month.
There is still a cohort of professional investors underweight the market and lagging in performance, as the chart below shows. In the bottom panel, we plotted the Median Dark Pools Index 1W (smoothed) for the top 1000 stocks in the market. Professional investors accelerated accumulation into the April decline (bullish divergence), while flows have receded since reaching a relative top in late June. Since then, while levels are nowhere near as low as December 2024, the trend has been toward more distribution than accumulation.
We would argue that this setup favors a “buy-the-dip” environment, as large traders will eventually need to play “catch-up” in terms of performance.
A phenomenon which caught our eye during the past weeks and that deserves its own discussion is the recent trend of “recycling dollars” in the A.I. space.
This cycle happens when capital raised by one group of companies—through stocks, convertibles, or bonds—is reinvested into other companies, continuing the process. In the current market, AI-focused companies are securing funds to finance substantial AI capital expenditures, generating revenue for a select group of infrastructure providers specializing in GPUs, networking, power, and data centers.
These reported revenues then support additional capital raises, higher valuations, and even larger expansions — a dynamic which effectively demonstrates reflexivity in real time.
For example, Nvidia’s data-center business has achieved remarkable revenue growth, reaching $30 billion in Q2 FY2025 and $46.7 billion in Q2 FY2026. This revenue is supported by rapid expansion from hyperscalers and “neoclouds” scaling their operations aggressively.
At the same time, capital investment further down the chain is accelerating. Citigroup forecasts AI infrastructure spending to hit approximately $490 billion by 2026 and climb to $2.8 trillion by 2029.
Let’s break down the vendor-financing cycle:
Bond Market (cash inflow) → Oracle
Oracle sells $18 billion in bonds to raise capital;
This cash is explicitly used to build AI cloud capacity (data centers, compute power) for large customers like OpenAI or Microsoft;
Oracle → NVIDIA (cash for chips)
Oracle uses the bond proceeds to buy AI chips from NVIDIA to power its AI cloud data centers;
NVIDIA benefits from this demand, generating revenue and funding its own expansion
NVIDIA → OpenAI
NVIDIA gives OpenAI cash for non-voting shares, on the condition that OpenAI uses NVIDIA systems for 10 GW infrastructure;
OpenAI → Oracle
OpenAI Contracts Oracle for AI Cloud Services through “Stargate” initiative
Deal funds Oracle's debt servicing and NVIDIA chip buys
The last time we saw this type of deal-making was in the late 90’s telecom boom, as equipment makers extended aggressive credit to carriers, effectively funding the purchase orders that produced the vendors’ reported revenues. Eventually, as several vendors defaulted, the loop broke, since vendors started to write down the value of receivables.
However, this doesn’t mean that today’s AI spending is fake. It’s just that part of it has a “reflexive” capacity. As long as “the music plays”, it’s all well and fine. Problems appear when investors start demanding higher returns on their capital outlays than the technology can produce. Delays can also be caused by very real power and supply constraints.
Like the late 90’s, today’s environment is being driven by outsized growth assumptions that stretch linearly into the future. Back then, Cisco traded upwards of 100x earnings on the belief it was selling the “backbone of the internet.” Obscure companies with questionable business models sprang up like mushrooms (Pets.com and Webvan being most notorious). They eventually collapsed, as financial models were proven completely unsustainable.
The psychology of investors today also bears resemblance. Similar to the internet back then, the AI narrative is impossible to ignore, attracting a lot of capital by simple fear of missing out. Investors are more concerned about being left behind in the AI arms race than losing money if it doesn’t prove to be profitable. The less risky trade appears to be investing than not investing in AI, especially for dominant players like Meta, Microsoft and Google.
Although the similarities are notable, the differences are just as important. Unlike the dot-com stars of the 1990s, today’s AI leaders are not pre-revenue firms, burning cash, without a clear route to profitability. Nvidia, for example, earns tens of billions in quarterly revenue thanks to its dominance in GPUs. Microsoft, Amazon, and Alphabet operate substantial non-AI businesses generating free cash flow that supports their AI investments. These companies are not speculative ventures; they rank among the largest and most profitable corporations in history.
By contrast, in the late ’90s, investors piled capital into companies with no earnings or customers. That is definitely not the case today.
Historically transformative technologies (automobile, radio, the internet for example) take decades to succeed. While the overall trend may be correct, investors will most likely have a hard time picking the winners from the first generation of companies. Studebaker, Duesenberg, Packard, DeLorean, Saab are some examples from the automotive space of once illustrious companies which failed to survive. Similarly, Yahoo and AOL were dot-com era darlings which eventually receded from once dominant market positions.
Today’s tech leaders are very much aware of history and are doing their best to avoid the same fate. Their actions are inflating a bubble - which may yet last.
While many were “right” about the Dot.com bubble, they were also “wrong” in their investing. The bubble kept inflating without their participation and profits. As investors, we need to play this game of “musical chairs” that makes us money on the way up, during the “inflation phase”.
We are also aware that not all players in the AI space can keep their promises of success. A recent piece in The Guardian caught our attention:
“Investors have assumed that every major US player in [large language models (“LLMs”)] will be a winner. This assumption is essential, as the monopolies that power big tech – such as Microsoft’s Office suite, Google Search, Gmail, and Docs, and Meta’s Facebook – are, without exception, approaching the end of their useful lives. The vast majority of customers believe that these products have gotten worse – and made users less productive – over the past decade or more. Each big tech company needs a global monopoly in AI to sustain their success and market value. They are not all going to get one.”
Roger McNamee
“If something cannot go on forever, it will stop.”
This is the theme for this week’s Technical Analysis part, as the backdrop is as bullish as ever, with the inherent warnings that spring up whenever markets get extended. With little to spoil the party, bulls remain firmly in control. No the market cannot keep rising at 20% growth rates every year, as the longer term technical trend suggests.
For now, support remains just below the last close, at the $655 level (M-Trend), while shifting sentiment could trigger a full 50-DMA retracement on an “unexpected” catalyst (or plain old profit taking). We are still of the belief that the overall environment is one of dip-buying, which should foster upside into the end of the year. Resistance stands at $695.
Our Trading Strategy
Managing risk is more important than anything at this stage. Rebalancing positions, keeping track of profit and stop targets while being aware of potential downside is critical.
Most importantly, we could take some defensive actions at the next rebalancing interval in order to overweight Staples and Basic Materials in favor of our extended positions in Tech. A rotation can happen sooner or later, but it’s only a matter of time to a “rug-pull” in the high flying AI space. Ideally, we would like to position ahead of the move.
Separately, we’ll take stock of our automated strategies behavior tomorrow, looking at Enterprise for macro guidance. So far, this model has nailed the risk-on approach to markets since mid-May.
Signal Sigma Research & PRO members will be notified by Trade Alert of any live portfolio changes (if subscribed). If you’re not on this plan yet, you can get a free trial when you join our Society Forum. If you need any help with your trading strategy (or would like to implement one on your account), feel free to reach out!
Disclosures / Disclaimers: This is not a solicitation to buy, sell, or otherwise transact any stock or its derivatives. Nor should it be construed as an endorsement of any particular investment or opinion of the stock’s current or future price. To be clear, I do not encourage or recommend for anyone to follow my lead on this or any other stocks, since I may enter, exit, or reverse a position at any time without notice, regardless of the facts or perceived implications of this blog post. I currently do not own or plan to own any position, long or short, in the securities mentioned.
I am not a financial advisor licensed in the United States. Nor am I providing any recommendations, price targets, or opinions about valuation regarding the companies discussed herein. Any disclosures regarding my holdings are true as of the time this article is written, but subject to change without notice. I frequently trade my positions, often on an intraday basis. Thus, it is possible that I might be buying and/or selling the securities mentioned herein and/or its derivative at any time, regardless of (and possibly contrary to) the content of this blog post.
I undertake no responsibility to update my disclosures and they may therefore be inaccurate thereafter. Likewise, any opinions are as of the date of publication, and are subject to change without notice and may not be updated. I believe that the sources of information I use are accurate but there can be no assurance that they are. All investments carry the risk of loss and the securities mentioned herein may entail a high level of risk. Investors considering an investment should perform their own research and consult with a qualified investment professional.
I wrote this blog post myself, and it expresses my own opinions. I do not have a business relationship with any company whose stock is mentioned in this blog post. The information in this blog post is for informational purposes only and should not be regarded as investment advice or as a recommendation regarding any particular security or course of action.
The primary purpose of this blog post is to share industry expertise and research and receive feedback (confirmation / refutation) regarding my investment theses.