2023-05-26 08:15:00 ET
Summary
- Everyone is talking about Generative AI since Microsoft mentioned its use at the beginning of 2023, with some of QQQM's top holdings also making big announcements to this effect.
- The problem is that investors have been rewarding companies that cut down on employee numbers while announcing investments in AI without actually analyzing the potential returns.
- Going beyond the hype paradigm, this results in stereotypes and overvalued stocks, with Nvidia figuring as the "Epitome of AI-stereotyping"
- Also, investors are oblivious to the regulatory risks which pose real challenges to the growth of the AI market.
- All this AI stereotyping has led to inflated tech valuations and, in my view, offers an opportunity to make money through shorting QQQM.
In this thesis, I cover the Invesco NASDAQ 100 ETF (QQQM) which has rallied by more than 25% since the beginning of this year as shown below, but, it should gain even more, driven by Nvidia's (NVDA) latest financial results . In contrast to current market enthusiasm, my aim is to justify shorting the ETF, namely by questioning the rhetoric that AI has the capability to boost tech's revenue and profitability in a context where stakeholders seem to be stereotyped.
Comparison of Nasdaq-tracking QQQ and S&P 500 (seekingalpha.com)
To make my point, I will use the "stereotype" concept which has its roots in psychology, with an example being gender stereotyping where daughters were previously given only dolls to play with.
However, with social and cultural evolution, things have changed as seen in the image above, but, stereotyping has made a comeback in the tech world somewhat like the tech bubble twenty years back.
Big Announcements and Risks of Repackaging Old Technology with New Wordings
In fact, tech stocks started to outperform when to compensate for sluggish revenue growth in Q2-FY2023 caused by weakness in its cloud business, Microsoft ( MSFT ) announced its bet on generative AI. That was on January 24. Meta Labs ( META ) also followed suit with its stock subsequently rallying more than 100% year-to-date as shown below.
Value investors were also attracted when mega caps announced tens of thousands of jobs cuts, which normally augurs well for profitability. In this respect, another use of AI is improving the efficiency of corporate functions including eliminating certain back office functions as per IBM's ( IBM ) CEO.
The problem is that most are focused on the visible part of the iceberg or what Microsoft is doing to beef up its Bing search engine with ChatGPT's Generative AI. However, if one takes the time to go through earnings call transcripts, AI is supposed to help its cloud business, through more value creation for Azure customers for each dollar they spend, while at the same time, improving the software giant competition position and conferring it with more pricing power.
In this context, it must be mentioned that Microsoft already has an intelligent cloud segment and that both Alphabet ( GOOG ) and Amazon ( AMZN ) also propose AI services as part of their GCP and AWS infrastructures respectively.
Now, with ChatGPT there has been a renewed interest in using natural language for report-writing purposes as you no longer have to be an expert with years of training to work with intelligent algorithms, and anyone can access the tool and see the benefits. As a result, there has been a lot of interest from corporations in AI in general, but, the real challenge is to transform interests into business opportunities.
For this purpose, some companies may repackage existing products that have been around for years with Generative AI wordings to increase their appeal and take advantage of people becoming stereotyped. An example is Recommendation AI which is about studying the profile of buyers to recommend products that are most likely to be chosen. This technology has been around for years, used by some including Alphabet and Tencent (TCEHY), but since it overlaps with the sentiment analysis features that ChatGPT proposes , marketers can be tempted to brand all new products as Generative AI.
Thus, probably led by their marketing departments to show that they are not left behind, there was a flurry of announcements by the QQQM's other holdings, adding to AI stereotyping and feeding into Nvidia's valuations.
Nvidia is the Epitome of AI stereotyping
Talking processors for supporting AI workloads, there is Nvidia with its A100 GPUs which drive OpenAI's algorithms. Now, whether Microsoft succeeds in creating an AI-first cloud platform or not, Nvidia will surely benefit as more computing power gets added to Azure and to support wider usage of Bing. Other large corporations may also want to build their own private chatbots. Thus, the company expects second-quarter fiscal 2024 sales to be around $11 billion , or 53% above the $7.18 billion analysts were expecting.
Now, this is great, but the question is whether it can be sustained after the initial frenzy of orders coming from customers who do not want to miss the train. Also, does a 50% increase in sales justify a trailing price-to-sales multiple that is more than 900% above the median for the semiconductor sector?
This rich valuation in a way represents the epitome of AI stereotyping, as if only Nvidia's chips can drive intelligent algorithms. However, it is far from having a monopoly status as processors from Alphabet and Intel have also been found to be suitable and even better in certain instances by benchmarking organizations as I detailed in a recent thesis .
Valuations Grade (www.seekingalpha.com)
Along the same lines, Amazon does not entirely depend on Nvidia, and, as the largest public cloud provider it has developed its own purpose-built Trainium processors for deep learning. Also, with the announced launch of two new language models intended for customers including Bedrock, the hyperscaler is certainly in a sweet spot, but again one needs to see returns before investing.
Furthermore, Alphabet has launched its Bard Conversational AI to compete with Microsoft, prompting people to turn their attention to the rivalry between the two giants, instead of focusing on what value add the technology brings to the Google ecosystem or how the ad business will perform during an economic slowdown.
Noteworthily, in contrast to peers, Tesla ( TSLA ) and Broadcom ( AVGO ) have not made big announcements but nonetheless have prospects. Now, as one of OpenAI's major investors, Elon Musk's company has been using machine learning to improve the degree of autonomy of its vehicles through prediction algorithms, and Broadcom is not only a semiconductor manufacturer but has also diversified into the software business with its acquisition of CA Technologies and Symantec. As a result, it offers some of the monitoring and data protection solutions that will come in demand as hackers use ChatGPT to i mprove their ability at committing cybercrime.
Now, there are certainly a lot of prospects with a market expected to be worth over $50 billion in 2028 from only $11 billion this year, but as per these two companies' valuation grades of D, stereotyped investors seem to have already rewarded them more than what is potentially achievable.
Investors Oblivious of Regulatory Risks
This enthusiasm based on the preconceived idea that AI will not only boost revenue growth but also improve profitability without actually seeing any sign of improvement makes no sense, but, worst, the market seems not to have priced in the problems.
For this matter, like most innovative technologies, ChatGPT also comes with problems, in this case, data privacy concerns, with some users being able to access others' chats. As a temporary measure, the developer OpenAI has tweaked its application and disabled the history feature while a more permanent fix is found.
Now, this tweak addresses privacy concerns but, on the other hand, reduces the set of data from which ChatGPT can learn. This in turn diminishes its efficacy, as AI applications in general can be envisioned as data guzzlers which feed on vast amounts of data to generate the most comprehensive report.
Consequently, regulatory risks pose a real challenge, far beyond the millions or billions of dollars that big tech are fined from time to time by authorities in the U.S. and E.U. This may sap the appetite for building large AI platforms which can invariably digest all the consumer data at hand, in turn impacting demand for intelligent chips.
Unaware of those risks, the market seems to have elevated tech mega caps which were themselves previously regarded as risky and overpriced to safe haven status during the banking crisis, on the back of AI. Hence, QQQ's holdings all gained more than 10% during that period, while, the Dow Jones index which holds stocks of the more traditional sectors of the economy in including energy fell.
QQQM's Holdings (www.invesco.com)
For this matter, QQQM has a Price-to-Earnings ratio of 25.08x and a price-to-cashflow of 15.41x. Compare this to the Energy Select Sector SPDR ETF ( XLE ) with a P/E of just 10.55x, a P/CF of 3.39x, and the dividends paid to shareholders to have an idea of the degree to which tech is overvalued.
Now, the NASDAQ-100's top holdings except for PepsiCo represent about 55% of its overall weight. Therefore, considering that the ETF gained 25.45% YTD, 55% of this figure comes to 14% (25.45 x 0.55).
Short QQQM
Therefore, after a 25% rally, I estimate that the ETF can fall by about 14% which is the portion of the upside that can be attributed to AI stereotyping. Thus, by shorting QQQM, gains of $19 per share are possible based on the current stock price of $136.43 falling to $117.33 (136.43 x 0.86).
Noteworthily, this represents only a slight correction compared with the bursting of the Internet bubble when between April 2000 and October 2002, the Nasdaq composite lost 740% of the 800% it had gained between 1995 and March 2000. At that time, the technology sector was in a particularly bad shape as in addition to being overvalued, it was associated with uncertain financial prospects, and conducive to unpleasant surprises, which is not the case today.
As for the timing, you may want to wait before placing your bet as a pause by the U.S. Central Bank in early June may provide catalysts for additional upside.
Last but not least, there is Apple ( AAPL ) which has refrained from making any big announcements, but, proceeded more rationally by advertising for several jobs specifically looking for expertise in Generative AI. Still, the market has rewarded the company with a 37% YTD gain probably due to its rich device ecosystem used to access AI applications. This shows the high level of expectations, but, I remind investors that unlike in 2020 during the Covid pandemic when digital transformation boosted tech earnings, there is not likely to be any AI revolution this time.
For further details see:
AI Stereotyping Justifies Shorting The QQQM ETF