BMO Global Innovators Fund Active ETF Series-BGIN-Monthly Commentary
Stay in May
The market adage of “Sell in May and Go Away” advises investors to reduce their stock exposure as the Summer months approach and reinvest in the Fall. However, thanks to the powerful secular spending trends we outlined in our AI Investment Cycle monthly, we decided to “Stay in May” and stay invested in our AI-related investments in anticipation of excellent earnings results. That turned out to be excellent advice we expand upon in this monthly.
An Update on the AI Investment Cycle
Semiconductors | Cloud Capex | Applications |
|
|
|
Semiconductor Readthrough from NVIDIA’s1 Results
NVIDIA is the largest position in the BMO Global Innovators Fund and it is at the forefront of AI development. Its results serve as an important bellwether for the entire AI industry. Despite elevated expectations from investors, they did not disappoint and results confirmed that we are still in the virtuous stage of the AI Investment Cycle.
NVIDIA reported an increase in Q1 revenue to $26 billion, up 18% sequentially and up 262% year-on-year, outperforming their anticipated guidance of $24 billion for the quarter. Growth was driven by robust demand for the NVIDIA Hopper GPU computing platform. Data center revenue was up 23% sequentially and up 427% year- over-year reaching a record $22.6 billion with both training and inference growing significantly. NVIDIA’s full stack of technology continues to give clients the ability to successfully deploy fast, efficient, generative AI applications. NVIDIA also benefits from the sale of networking products that are attached to their AI semiconductor infrastructure and noted strong year-on-year growth for their InfiniBand networking gear and new announcements for Ethernet-compatible products.2
Cloud customers continue to spend to develop their own internal AI-based applications and to build infrastructure that they can rent out to others. Management noted that for every $1 spent on NVIDIA AI infrastructure, cloud providers can earn $5 in GPU instant hosting revenue over four years.
Additionally, customers are expanding beyond traditional data centers with new demand coming from countries around the world that are investing in Sovereign AI by building up domestic computing capacity to create AI- applications using data from their country and key industries.
Driving all of this is the productivity gains from AI applications with almost every industry noting significant productivity improvements with their AI-based pilot and commercial applications.
New AI Product Launches from Major Players (Google, OpenAI, Microsoft)
Google’s I/O Conference – The End of Search?
On May 14th, Google’s I/O Conference began with a number of notable announcements during the keynote speech. These included3.
- AI Overviews: Google announced they will be rolling out a fully revamped search experience in the US called AI Overviews. Now when you search on Google, instead of seeing the links related to your search, AI Overviews provides an AI-generated summary with key information and links to dig deeper at the top of the search page (refer to Figure 1 for an example). From a monetization perspective, at the bottom of the AI Overview summarization, there will be sponsored advertisements related to the search topic, followed by the typical search links. The main debate with this change to search is how much more expensive search becomes given it requires more compute to produce these responses. Additionally, we’ve seen several hallucinations within AI overviews in the last couple of weeks. Google will need to continue to refine their models to lessen the hallucinations.
- Context Window Update: Google announced that they will be bringing an improved version of Gemini 1.5 pro to all developers globally now with a 1 million context window. They will also be expanding the context window to 2 million, which is now available in private preview. The increase in context window to 2 million is a massive improvement and helps to set Gemini apart from other models with lower context windows. When responding to queries, AI models typically have a limit on how much data can be considered to generate a response. The limit on the text size it can consider is called a context window. A larger context window is particularly helpful for tasks that require a deep understanding of context, such as summarizing long articles or answering complex questions. A larger context window can help improve the performance of an AI model, reduce hallucination, and significantly increase accuracy. Google’s ultimate aim is to unlock an infinite number of context windows.
- Project Astra: Google is in the midst of exploring the future of AI assistants that can process multimodal information, understand the context, and respond naturally in conversations. One of the challenges here is in getting the response time down to something conversational. Google has been working to improve how their models perceive, reason, and converse to make the pace and quality of interaction feel more natural.
- Generative media: Introduced (1) Imagen 3, which is a text to image model that produces high quality images in a wide range of formats and styles (refer to Figure 2 for an example), and (2) a professional music AI tool to help artists expand their creativity. They also announced Veo, which is a generative video model able to produce videos in different cinematic styles. Veo will generate a 49 second video, with the option of extending the video length.
Figure 1: AI Overviews Example
Figure 2: Imagen 3 Example of Image Generation and Prompt
OpenAI - ChatGPT Update Demonstrates Multimedia Capabilities4
On May 13th OpenAI launched GPT-4o (“o” for “omni”), which is an AI model that can work with audio, images, and text in real time simultaneously. This moves AI capabilities towards natural human-computer interaction. The model can accept any combination of text, audio, image, and video and generates any combination of the inputs as outputs, responding in a timeframe of a human (an average of 0.3 seconds). This is a huge improvement over past models that were slow to respond to voice commands with delays of 5 seconds on average, which made them infeasible for human voice-to-voice conversations. Capability-wise its performance is like OpenAI’s flagship language model GPT-4 in terms of efficacy and accuracy of text outputs for English and coding languages. The demos of the product are amazing5 with examples of two GPT-4os interacting and singing, a sample interview prep use case and we could personally relate to the math homework example with Imran Khan of the Khan Academy.
Microsoft’s Build Conference – AI-Enabled PCs are Coming6
The most interesting announcement from this conference was new Surface laptops and tablets designed with generative artificial-intelligence capabilities right on the device. Using Qualcomm’s Snapdragon processors and an optimized version of Windows that together are designed to perform powerful AI-computing functions right on the PC versus the cloud. Edge AI vs Cloud AI. This may help solve some of the latency, privacy, and power issues associated with AI. We expect to hear more announcements in this product category from the Computex show in Taiwan that we are attending.
Apple in June
We expect to hear more new product announcements are expected to come in June with Apple’s WWDC (Worldwide Developers Conference).
Summary
This has been an incredibly active May and it will be followed by a slew of exciting announcements in June including keynotes from Computex in Taiwan and our takeaways from the Collision conference held here in Canada in June. Stay tuned for more news to come. You may be relaxing as the breezy days of Summer come this year but developments in the AI space are showing no signs of slowing down.
1 Nvidia Corporation is an American multinational corporation and technology company.
2 Source: Nvidia, Bloomberg, and BMO GAM. May 31, 2024.
3 Source: Alphabet, Bloomberg, and BMO GAM. May 31, 2024.
4 Source: OpenAI, Bloomberg and BMO GAM. May 31, 2024.
5 https://openai.com/index/hello…
6 Source: Microsoft, Bloomberg, and BMO GAM. May 31, 2024.
Disclaimer
Commissions, management fees and expenses (if applicable) all may be associated with investments in mutual funds. Trailing commissions may be associated with investments in certain series of securities of mutual funds. Please read the fund facts, ETF facts or prospectus of the relevant mutual fund before investing. Mutual funds are not guaranteed, their values change frequently and past performance may not be repeated. Distributions are not guaranteed and are subject to change and/or elimination.
For a summary of the risks of an investment in the BMO Mutual Funds, please see the specific risks set out in the prospectus. ETF Series of the BMO Mutual Funds trade like stocks, fluctuate in market value and may trade at a discount to their net asset value, which may increase the risk of loss. Distributions are not guaranteed and are subject to change and/or elimination.
BMO Mutual Funds are managed by BMO Investments Inc., which is an investment fund manager and a separate legal entity from Bank of Montreal.
BMO Global Asset Management is a brand name under which BMO Asset Management Inc. and BMO Investments Inc. operate. The viewpoints expressed by the Portfolio Manager represents their assessment of the markets at the time of publication. Those views are subject to change without notice at any time without any kind of notice. The information provided herein does not constitute a solicitation of an offer to buy, or an offer to sell securities nor should the information be relied upon as investment advice. Past performance is no guarantee of future results. This communication is intended for informational purposes only.
As the fund is less than one year old, the actual Management Expense Ratio (MER) will not be known until the fund financial statements for the current fiscal year are published. The estimated MER is an estimate only of expected fund costs until the completion of a full fiscal year, and is not guaranteed.
Any statement that necessarily depends on future events may be a forward-looking statement. Forward-looking statements are not guarantees of performance. They involve risks, uncertainties and assumptions. Although such statements are based on assumptions that are believed to be reasonable, there can be no assurance that actual results will not differ materially from expectations. Investors are cautioned not to rely unduly on any forward-looking statements. In connection with any forward-looking statements, investors should carefully consider the areas of risk described in the most recent simplified prospectus.
“BMO (M-bar roundel symbol)” is a registered trademark of Bank of Montreal, used under licence.
Publication Date: June 2024