A lot has happened in the year since I wrote my last commentary, but the most-common question topic over that time continues to revolve around artificial intelligence (AI). This wasn’t a surprise to us (or I’m assuming anyone reading this), since AI continues to dominate headlines, dinner conversations and, increasingly, investor thinking. Many of the financial advisors we speak with are now being asked by clients how AI will impact the world and their portfolios.
Across our recent conversations, AI questions generally fall into three buckets:
How is EdgePoint using AI?
How are our portfolio companies using AI?
How could advisors benefit most from the AI trend?
I’ll use this commentary to answer each question.
How is EdgePoint using AI?
Large language models (LLMs) like ChatGPT are most effective at synthesizing and summarizing written information. They’re not great at math (yet), but they’re very good at helping us parse vast quantities of text faster.
At EdgePoint, we think of AI as a co-pilot – one that helps us cut through the noise and focus on what matters. A few examples:
Getting up to speed on a name or industry
We’ve used AI to take years’ worth of earnings call transcripts, investor-day decks and filings, then load them into a searchable tool. From there, we can ask targeted questions, such as the reasons for first-quarter margin compression, what synergies were management targeting with an acquisition, or how did the tone change over time. It’s like having an analyst who’s read everything – and remembers it all instantly.
Comparing companies or industry trends
Want to know which competitor is growing fastest and the reason why, according to management? Or how different semiconductor CEOs talk about tariffs? With AI, we can ask those questions and get quick, organized responses that previously would’ve required hours of work.
Assembling complex narratives from public data
One of our CEOs recently sold stock, citing stress on his personal real estate portfolio. We used another tool to validate this. In under an hour, we had a list of his properties, public appraisals from state websites and evidence of declining values. Historically, that might’ve taken a day or more.
Learning new concepts quickly
Whether it’s the 10-year Brazilian R$/US$ currency history, the difference between MRDIMM vs. CXL memory architectures, or what “sediment-hosted copper deposits” means, AI can explain it all. Even better, it can do it in a format that adapts to how you learn. It’s like having a personal tutor with an infinite syllabus.
AI is a tool, meaning it’s important to look at how and why it’s being utilized. We’re using AI to complement our curiosity to ask better questions, improve our research abilities and learn faster. This effectively buys us something incredibly valuable – time. It allows us to spend more of our time thinking about ways a business might have unrecognized positive change, rather than just searching.
How are our portfolio companies using AI?
AI is a massive tailwind for many companies. Some of the names are obvious (which means the benefit of AI is already factored into the share price), but some are less so. Here are a few businesses we own that can benefit from the adoption of AI:
Applied Materials, Inc. and Rambus Inc. – the picks and shovels of AI
Everyone’s heard of NVIDIA Corp. But as with any gold rush, the people selling picks and shovels often generate very pleasing returns because they get a cut of profits from their customers (whether or not they hit paydirt). In the case of AI, semiconductor tools and memory interfaces are critical.Appliedi builds the equipment used to manufacture semiconductors – some of the most advanced machines on earth. To put this into perspective, a modern chip can have 60 miles of copper wire packed onto something the size of your fingernail and require thousands of steps over at least three months to produce. AI is driving demand for more chips, which drives demand for more of Applied’s tools.
Given the complexity of these tools, there are only a small handful of companies globally that can make them. AI is accelerating the demand for improvements in chips and their performance, increasing the complexity of both the chips themselves and their protective casing (packaging). This ultimately drives more demand for the most advanced tools, such as those manufactured by Applied.Rambusi designs chips that help solve a bottleneck in computing called the “memory wall.” As AI models get bigger, the flow of data between the processor and memory becomes a limiting factor. Rambus chips help break through that wall and help AI work faster and more efficiently, especially in data centres.
As mentioned, AI is accelerating the pace of innovation for semiconductors. Rambus’ end market for buffering chips historically worked on a two-year upgrade cycle. AI is pushing semiconductor manufacturers to improve their chips annually. This acceleration is leading to meaningful growth for Rambus’ end market and benefiting suppliers like Rambus that can meet these accelerated delivery schedules through increasing market share.
Roche Holdings, Inc. – speeding up drug discovery
Pharmaceutical innovation has historically been expensive, slow and risky – a combination that makes AI’s promise incredibly valuable. Rochei was an early AI adopter, partnering with NVIDIA years ago.
One example of how AI helped Roche was identifying vixarelimab, a promising treatment for pulmonary fibrosis. AI didn’t just find the needle in the haystack – it helped sift the hay.
Roche is now expanding its AI capabilities across drug discovery and diagnostics. Speeding up the process means more shots on goal and a higher probability of success, all while improving outcomes for patients.SAP SE – making enterprises smarter
SAPi is the global leader in enterprise resource-planning (ERP) software – the digital command centre of a business. Like an air traffic control tower for a company, it helps oversee operations, finance, supply chains and more. With all of a company’s data in one place, it gives SAP a unique opportunity to introduce AI tools to help customers understand their data and make better decisions.
Some examples of areas SAP is looking to introduce AI:Predictive forecasting – helping companies anticipate inventory needs or cash flow shortfalls before they happen.
Automated workflows – using natural language prompts to trigger complex actions like generating purchase orders, approving expenses or flagging anomalies in financial reports.
AI copilots for employees – giving sales, finance or procurement teams a virtual assistant that can answer operational questions in real time by drawing on company-wide data (e.g., “What’s our top-selling product in the northeast this quarter?”).
Compliance and audit support – Automatically identifying outliers or suspicious transactions to reduce risk and improve governance.
There is a vast opportunity for SAP to offer more value to customers through AI features, and these features could be monetized and create significant shareholder value.
Dayforce, Inc. – rethinking payroll and HR from the ground up
Dayforcei is a disruptive provider of payroll and human capital management software. What sets Dayforce apart from its peers is its unified architecture – the entire platform is built on a single, integrated database. While most competitors have stitched together multiple systems over time, Dayforce offers a single source of truth.
That architecture unlocks meaningful advantages:Real-time pay calculation – instead of waiting until the end of a pay period to calculate wages, Dayforce computes pay continuously. This gives employers better visibility into labour costs and makes features like same-day pay or daily wage payouts possible.
Better data, better AI – Because all payroll, benefits, HR and time-tracking data live in one place, Dayforce is uniquely positioned to leverage AI. Its Copilot tool allows employers to upload documents like training manuals, benefits plans and HR policies to train the AI to answer employee questions. It’s like having a 24/7 digital HR assistant that never forgets a policy or misses a detail.
Monetizing early – Dayforce is already benefitting by charging US$3 to US$5 per employee per month for its AI Copilot, with expectations that this could rise to US$10 to US$12. For context, this would represent a meaningful increase on top of their existing revenue base.
While AI is still in its early innings, most of the value creation so far has been concentrated at the front of the stack – semiconductors, infrastructure and foundational models (AI-training models based on large data sets). But over time, as the technology matures and spreads, we believe its impact will broaden across industries and use cases.
We own several businesses that are already beginning to benefit from this shift – some in obvious ways, others in more nuanced ones. What’s particularly exciting is that in many cases, we believe we aren’t paying for these AI tailwinds at their current valuations. These are companies we own for other reasons, with AI offering a compelling free option on future growth. In other words, we’re able to buy growth without paying for it.
How could advisors benefit most from the AI trend?
Over the last year, our advisors have mentioned several AI platforms they’re using to make their practices more efficient. Here are some examples of the most recommended ones that could benefit our other advisor partners (as long as they adhere to their firm’s policies).
A few that stand out:
NotebookLM (Google)
A powerful tool for searching across uploaded documents. Imagine uploading fund facts, compliance manuals or client-onboarding guides, and being able to ask natural-language questions like: “What’s the risk rating of Fund A?” or “Where do we disclose our referral policy?”ChatGPT (OpenAI)
The Swiss army knife of AI tools. For US$20/month, you get a tool that can:Summarize client meetings from your voice notes
Transcribe written notes using photos
Generate marketing ideas or client event plans
Help with Know Your Product (KYP) documentation
Compare portfolios or fund fees with natural-language queries
Deep Research (OpenAI)
Great for prospecting or due diligence. Let’s say you’re considering a new office space or event venue. Deep Research can aggregate location info, customer reviews, pricing and hours of operation, all at once.
This is just scratching the surface. If you’re an advisor curious about where to start with AI, don’t hesitate to reach out.
The more things change…
The core of our investment approach has stayed consistent through countless technological innovations. We’re optimistic about the long-term potential of AI. Not because of hype, but because of how it helps us do our job better – continuing to identify businesses with underappreciated growth potential. AI can also be one source of that growth and, if it is, we believe we’ve invested in companies that are ready to take advantage of it.
Finally, it’s important to remember that using AI is like Jeopardy! – the questions matter. Knowing what you want to know is as important as the tool you’re using. AI is only as good as the prompts that are being written. We’ve found that AI is just another way to boost the natural curiosity that we’ve nurtured over the years from researching businesses we want to own.
Thank you for your continued trust and support. We work hard every day to be worthy of it.