Every investing year is interesting in its own way, but 2025 provided a decidedly distinct waypost to mark DM’s first quarter century in business. The inaugural year of a US presidential term is typically strong for stocks but, as we learned during its first go around, this administration doesn’t often lend itself to historical comparison. In fact, it seemed that any chance at a favourable start was dashed in early April when a nuclear scale tariff bomb was dropped on the world and, once investors calculated what this punishing levy would do to global commerce, the S&P 500 quickly shed a fifth of its value.
In a series of retreats, however – the frequency of which earned him an acronym nickname matching the quintessential Mexican street food – the president backtracked on his initial decree multiple times, reducing or delaying tariffs for several countries and exempting entire product categories. Relieved, markets recovered and soon animal spirits took over with exceptional force, not only boosting prior market leaders but also putting a charge into virtually anything related to artificial intelligence and flooding capital into stocks with no earnings and startups which offered little more than a vague business plan (see chart).
In a rare divergence, just as investors were expressing boundless optimism in the economic and technological future, they were also bidding up the ancient doomsday asset; in fact, the price of gold jumped by more than half in 2025, posting its best year since 1979 when US inflation was running at 13%. The reasons for the yellow metal’s rise are likely myriad, not least of which being the rapidly deteriorating US government balance sheet. Regardless, this collective backdrop created something of a “barbell” scenario in investment markets, where both aggressive growth and traditional hedging assets were in vogue at the same time, with much of what resides in between largely ignored. While this has created fertile hunting ground for a manager targeting reasonably priced cash flow growers which might not be prominent on the mainstream radar, it also made keeping up with benchmark returns unusually difficult over the past 12 months. And so, what was a very good year for markets turned out to be somewhat less good for us.
Canada
With a relatively high representation by metals and mining stocks, the TSX performed very well in 2025, even surpassing the S&P 500 in both CAD and USD terms. Owing to our core investment philosophy and for the reasons listed below, however, we have tended to markedly under-weight such companies in our equity allocations:
- commodity producers are capital intensive businesses, which means that a significant portion of available cash must go back into the ground to maintain production, leaving less available to generate growth, pursue acquisitions, or provide shareholder distributions;
- when they do reinvest funds, resource companies have a well-earned reputation for underestimating the cost of capital projects, leading to budget overruns and balance sheet pressure;
- the assets of resource extractors tend to be geographically fixed (and often in politically less desirable regions), limiting operational flexibility and leaving little room to escape adverse developments;
- commodity producers are always price takers instead of price setters; this isn’t a problem when prices are rising but when they fall, there is little opportunity for the company to maintain margins based on attributes like brand affinity, product quality, or network economics.
In 2025, the TSX materials sector (which is nearly three quarters weighted in gold companies) jumped by more than 100% and, not surprisingly, was the largest contributor to aggregate market return. Though we hold a modest position in Wheaton Precious Metals (which avoids most of the shortcomings listed above because it is a streamer earning royalties and not a producer developing mines), a single stock allocation wasn’t enough to overcome the massive precious metals rally.
US
Just as it was difficult in the first half of the year to discuss the backdrop for US stocks without referencing the White House, it’s now virtually impossible to do so without mentioning artificial intelligence and the influence that its buildout is having on both the economy and the investing environment. The attention that AI is getting is more than warranted, however, especially when its potential impact is fully appreciated. For example, labour currently accounts for about 55% of total private sector expenses in the US and, according to Blackrock, if AI and related technologies were able to reduce this share of corporate costs by just 5%, it would translate into about $878 billion in incremental after-tax corporate profit each year.
Given the potential scale of this economic opportunity, we certainly want our equity mandates to participate but, as stewards of capital, we also need to think about risk in the management of your assets. Last quarter, we outlined some of the things that we thought might precipitate a pause in AI growth and a resultant re-rating of associated share prices, including the accelerating trend toward “circular financing” by industry players. Under such arrangements, firms act as some combination of customer, supplier, lender, and equity investor to each other, which can not only muddy the determination of what capital flow is attributable to whom but naturally raises the risk of contagion, where a setback for one becomes a problem for many.
We also referenced the potential for infrastructure bottlenecks to cause a slowdown in AI growth, as the massive power requirements of data centres collide with an already stretched grid. Not surprisingly, existing rate payers are starting to push back against these incursions and even ordinarily pro-development Florida recently proposed an “AI Bill of Rights”, which would protect local communities’ right to block data centre construction. Though it runs afoul of a subsequently issued executive order preventing “excessive state regulation of AI”, the bill is emblematic of increasingly negative consumer sentiment, which is now showing up in town hall meetings and will likely rear its head in the upcoming mid-term elections. With this in mind, it’s not unfathomable that a bump in the AI path might come from a source much more prosaic than the challenges of processing power, system integration, or chip supply.
Though much of the AI buildout has been funded through the prodigious cash flow of senior tech and communications companies, debt financing is increasingly being utilized to pay for these enormous investments. And, because the obligation to repay interest and principal on borrowings is fixed, but the timing and scale of AI revenue is not, the risk of a cash flow mismatch is elevated. When Oracle announced its $300bn deal with OpenAI, for example, the market initially cheered, sending its share price soaring through $300 (which is roughly where we sold our entire position). Once investors contemplated how much the company would have to borrow to fulfill its part of the agreement, however, the stock plunged by 40%. Elsewhere, some companies are creating exotic “off balance sheet” structures to keep AI-related debt away from their books and though this isn’t necessarily a red flag yet, it does warrant consideration, especially when we recall how such novel financial engineering schemes have fared in the past.
Apart from these concerns, picking the eventual winners in any period of transformative innovation is really hard (Netscape, Myspace, or AOL anyone?). In less than a year, in fact, Alphabet provided a condensed vignette of this tendency when the stock was initially battered in Q1 on concerns that artificial intelligence would displace the need for traditional internet search and undermine the company’s most important revenue source. By the second half of the year, however, the stock was the darling of the Mag-7, as its release of the Gemini 3 AI assistant received rave reviews and was soon scaling faster than ChatGPT (see chart below).
Whereas nine months ago, strategists were speculating about what would happen to Google’s revenue base if a big chunk of ad spend was grabbed by OpenAI, a better question today might be, “how durable is OpenAI’s revenue stream if Alphabet decides to offer Gemini for free, just as it does with search?” Though it would have been perfectly justifiable to move on from Alphabet in early 2025, that would have also meant missing out on a 65% calendar year gain.
Aside from its leadership in artificial intelligence, we believe that Alphabet provides our portfolios with a risk-managed way to gain exposure to the burgeoning sector. First, although AI will undoubtedly be an important profit driver for the company going forward, Alphabet already has substantial eggs in other baskets, including Google Cloud and advertising through YouTube and Google Search (in fact, ad revenue from search hit a new high in September of $590m … per day). Second, Alphabet is a platform company which provides (controls?) access to a vast number of consumers and businesses, almost certainly offering a coveted gateway to the monetization of artificial intelligence (on cue, just before this note was published, Alphabet announced a partnership with Walmart which will allow customers to complete purchases directly through the Gemini interface).
The same potential can be found in DM portfolio holdings Amazon (e-commerce, robotics, cloud), Meta (social), Microsoft (enterprise, cloud), and Apple (devices). While Alphabet and the first three on this list have deployed many billions developing their own AI products, Apple has been noticeably absent from the spending spree. Much like Alphabet, Apple was maligned early in the year by investors and analysts who grumbled that the company was falling behind and would ultimately miss the AI boat.
That impression has begun to shift, however, with several now suggesting that Apple has been cagey in sitting back and may be able to capitalize on its restraint by reaping the benefits of everyone else’s work to apply the best advances and features to its very sticky user base at a fraction of the cost and with significantly reduced development risk. And, with 2.4 billion iOS devices and 1.5 billion iPhones in service, Apple’s opportunity to monetize AI may be greater than that of anyone – in the end, the winner in artificial intelligence may not be the smartest or earliest application, but rather the one that’s already there when you unlock your phone.
The year ahead
About every half decade or so in our 25-year history we’ve experienced a performance setback and, while all of our mandates performed well in absolute terms in 2025, none exceeded its benchmark. This unfortunately is the downside of maintaining a strict equity discipline: it’s impossible for a consistent management philosophy to be in favour at every stage of a market cycle.
That said, the last weeks of 2025 were encouraging for us, as US market participation began to broaden beyond the S&P’s biggest companies and in Canada, several of our first half laggards staged powerful comebacks in the closing quarter. As our investment committee prepares for 2026, we’re confident that market concentration will continue to fall in the US, improving the backdrop for stock picking and allowing some of the forgotten middle market to catch up with last year’s performance leaders.
Though it isn’t naturally in our DNA to own commodity-oriented companies, we can’t deny the shifting sands beneath global finances and the possibility that gold’s climb is still in its early innings, despite its prodigious run in recent months. This, along with evaluating overlooked cash flow growers in the US and outside North America, will likely occupy much of our investment committee time in 2026.