Ever since OpenAI's release of ChatGPT in late 2022, interest in the AI industry has evolved at a breakneck pace. From a $10 billion investment by Microsoft (NASDAQ:MSFT) into OpenAI, to Google's launch of its Bard AI chatbot, there's been no shortage of growth catalysts.
The growth hit new levels on Tuesday, May 30th after Nvidia hit a $1 trillion dollar market cap after shares reached a high of around $409. The company cashed in on the AI gold rush after a stellar earnings report, detailing incredible profit and sales growth along with strong guidance for the upcoming year. Underpinning these results was high demand for its chips, which led to strong gains in technology sector ETFs and semiconductor ETFs last week.
The ETF industry has been busy too. On May 18th, Roundhill Investments launched the Roundhill Generative AI & Technology ETF (CHAT). This launch capped off a busy year for Roundhill that saw the release of numerous thematic ETFs capitalizing on current trends, including the Roundhill BIG Bank ETF (BIGB) which benefitted from weakness in smaller regional banks.
Let's take a look at the overall generative AI theme and look at what CHAT has to offer.
Generative AI explained
To get an answer from the subject matter expert, I asked ChatGPT-4 the following question: "Define 'generative AI" for me in a paragraph". Here's what it had to say:
"Generative AI refers to a subfield of artificial intelligence focused on creating new content, often by learning patterns from a set of input data and then generating output that follows these learned patterns. This could involve creating anything from written text, to music, images, or any other kind of content. Generative AI models, like those based on Generative Adversarial Networks (GANs) or transformers like GPT-3 and GPT-4, work by training on vast amounts of data and learning to mimic the structure and style of that data."
Here's an easier way to understand it: imagine you've asked a bunch of artists to draw a picture of a cat. You'll get lots of different pictures, but they'll all resemble a cat in some way. Then, you give these pictures to a new artist, and ask them to draw a cat based on these examples.
Generative AI works in a similar way. You feed it lots of examples - this could be pictures, text, music, anything really - and it learns the patterns and structures. Then, when you ask it to create something new, it uses what it's learned to generate new content that fits within those patterns and structures.
This is why generative AI can create a lot of different things. For instance, it can write a piece of music in the style of Beethoven if you feed it enough of Beethoven's music, or it can create a realistic image of a cat if you give it enough pictures of cats. It can even write a story or an article if you provide it with enough examples of the kind of writing you want (there goes my job security).
However, it's important to remember that generative AI isn't consciously creative or imaginative in the way humans are. It's just using mathematical methods to generate new combinations based on the data it's been trained on. There have also been issues with factual inaccuracies, particularly in cases where generative AI straight up invents a fictional source or citation.
CHAT ETF analysis
Roundhill's bull thesis for AI growth centers around a few highlights.
- Firstly, they project that generative AI enterprise software has a total addressable market (TAM) of approximately $121 billion.
- Additionally, they cite Goldman Sachs (NYSE:GS)' Macro team, which stated that AI could stimulate roughly $7 trillion in worldwide economic growth in the coming decade.
- Finally, to provide evidence of the rapid uptake of AI applications, Roundhill cites how ChatGPT achieved a milestone of 100 million users just two months post-launch, which ranks it among the fastest apps to reach this user base, a record previously held by TikTok.
Unlike some of the existing AI and robotics ETFs on the market, CHAT is actively managed. That is, it does not track an index by an external provider like Solactive (a common provider of thematic indexes). The ETF has a globally diversified, large-cap focus, classifying its current 30 holdings as either:
- AI platforms: companies that develop, train, and sell large language models (LLMs) and application programming interfaces (APIs) for third-party developers.
- Infrastructure: companies providing hardware and semiconductor used in AI applications.
- Software: companies that build and sell enterprise and consumer AI applications.
Unsurprisingly, CHAT's current largest holding as of May 29th is Nvidia with a 9.37% weighting. Following it is a smattering of familiar large-cap U.S. tech stocks such as Microsoft, Alphabet (NASDAQ:GOOGL), AMD, and Adobe (NASDAQ:ADBE). The ETF charges a 0.75% expense ratio and has attracted around $13 million in assets under management.
This content was originally published by our partners at ETF Central.