Client Update, September 2026: Paying Attention
“Remember that the art comes in editing after the invention, not before.”
Ben Thompson
Dear Clients and Friends,
I hope you have all had a wonderful summer and have managed to stay cool despite the intense heat covering much of the country.
Despite taking August off from this monthly letter myself, the world kept turning, and we return this September with a lot of news to touch on. For that reason, I thought I would do something a little different this month. Rather than trying to force every noteworthy event of the summer into a single theme, I thought I would simply talk through some of what has been on my desk lately, along with the implications of each for the markets, our portfolio, and the broader environment.
To start, let’s touch on the recent eye-opening activity in the bond market.
Let the Bond Market Speak
I have written several times over the past year and a half about the importance of the bond market as a potential guardrail on government policy. The bond market has an ability to discipline political and business actors alike, even more than the equities market. As the Financial Post puts it, “Treasury yields are the center of the US financial universe, serving as the benchmark for everything from corporate debt to housing loans.”
So, when bond traders start demanding higher interest rates in exchange for their money, as we have seen recently with the 10-Year Treasury yield reaching 5%, even the government must take notice.
Something similar happened dramatically last year. After Trump announced his so-called Liberation Day tariffs, the bond market reacted poorly, and Trump subsequently pulled back. Here’s what I wrote in May 2025:
That’s because to the credit investor, who is pricing in the sovereign risk of the U.S., the U.S. suddenly looked a lot more economically fragile and a lot riskier. If you’re buying U.S. sovereign credit, your perception and expectations around U.S. economic growth have changed in light of these higher tariffs and potential retaliation and counteractions, like a potential U.S.-China trade war. Ultimately, it was the credit markets’ repudiation of Trump’s tariffs that triggered the 90-day pause [on those tariffs].
Today, the specifics are different, but the trends are the same. Tariffs remain at least part of the picture (and American businesses clearly hate them), and we are now facing down a potential trade war with Canada (more on that in a moment), ongoing war in Iran, persistent inflation, and a national debt that has climbed past $40 trillion.
And today, the bond market is once again sending a message. Investors who lend the United States money want to be compensated for inflation, fiscal risk and uncertainty.
The Trump Administration initially tried to reassure increasingly nervous bond traders. For example, Treasury Secretary Scott Bessent discussed expanding Treasury purchases of longer-dated government bonds. The mechanics are straightforward: introduce a large new buyer into the market, and all else being equal, bond prices should rise and yields should fall.
Markets responded accordingly at first, but not for long.
That brings me to an excellent recent Wall Street Journal op-ed from Stanley Druckenmiller, one of the most accomplished investors of the past several decades. Druckenmiller argues that policymakers should be extremely reluctant to intervene simply because they dislike the price the bond market has established. “Governments defending prices against fundamentals always lose,” he argues. “The only variable is how much they spend before conceding.”
I agree.
We need to distinguish between ensuring that Treasury markets remain liquid and functional during a genuine crisis and intervening because policymakers would prefer lower borrowing costs. If investors are demanding higher yields because they see greater inflation or fiscal risk, suppressing those yields does not eliminate the underlying problem. In fact, it can make matters worse by removing the pressure to address it.
This is one reason I continue to view the bond market as something of a great equalizer. Markets do not need to win elections, curry favor, or make anyone feel good. At some point, price simply must reflect reality.
That is also why policymakers should listen carefully, even (or especially) when the message is uncomfortable.
When Our Largest Trading Partner Stops Trusting Us
Speaking of uncomfortable, another item I have been following closely is the deteriorating relationship between the United States and Canada.
The Atlantic recently published an excellent analysis of how trade negotiations between the two countries fell apart. The immediate dispute is about tariffs, and the failure of negotiations has now triggered retaliatory tariffs between the two countries.
However, I think the larger story is about something more fundamental: Canada increasingly concluded that it was not negotiating with a consistent, predictable, or credible partner.
The trends were already pointing this way back in June. Then, a Pew poll found that fewer Canadians than ever viewed the U.S. favorably (only 33%, a steep drop from 54% in 2024).
Now the trade war is risking making permanent a rupture that might have otherwise proven to be only transitory. “The collapse of the tariff talks points to the fact that the old Canada-U.S. relationship is over and, for many Canadians, it also confirms the perception that Canada can’t trust the Trump administration,” Daniel Béland, a political science professor at McGill University in Montreal, told AP News in late August.
There are plenty of legitimate disagreements the United States can have with its allies. We can debate trade balances. We can argue that European countries have historically spent too little on their militaries relative to GDP and relied too heavily on American defense spending. Those are reasonable policy conversations.
What has happened with Canada is different. After talks broke down, Canadian Prime Minister Mark Carney commented, “We recognized that sometimes [America’s] signature was written in pencil.”
As an American, that hurts to hear from a close trading partner, but it especially hurts because there’s truth to it. There was a time, not long ago, when the United States could be relied on to be a steadfastly trusted partner to its closest allies. That does not seem to be true anymore. Carney has argued that the only rational thing for Canada to do now is reduce its exposure from a risk management standpoint. He is correct.
This is ultimately to the detriment of Americans. The economic relationship between the United States and Canada is extraordinarily integrated, particularly across the Midwest and Ontario. Supply chains, energy markets, manufacturing and employment cross the border in both directions. Michigan, Wisconsin, and other states are not economically isolated from Ontario simply because a line appears between them on a map.
That means a trade war intended to exert pressure on Canada can inflict significant damage on American businesses and communities as well.
The Cost of Lost Credibility
Zooming out for a moment, this situation is illustrative of an unhappy turn in the U.S.’s global position. For decades, one of the United States' greatest competitive advantages has been something difficult to place on a balance sheet: credibility.
Our credibility as an economic engine and partner is a major reason why, for example, the U.S. dollar serves as the world’s primary reserve currency.
Our allies did not always agree with us. Our trading partners certainly did not always get everything they wanted. But there was enormous value in the belief that the United States generally meant what it said, honored its commitments, and remained a reliable economic and strategic partner.
This is where I think several seemingly unrelated stories begin to intersect.
The bond market is ultimately making judgments about the credibility of American fiscal and economic policy, even in the face of promises from Treasury Secretary Bessent and Trump himself. Canada is making judgments about the credibility of the United States as a trading partner. Other allies are making similar calculations about our reliability as a strategic partner.
No single decision in the reordering of a global economic system necessarily looks consequential. In fact, some of the most economically destructive consequences of current U.S. policy may go completely unnoticed by the vast majority of Americans.
When a foreign student chooses to attend a university somewhere else because we have made it harder to study here, our economic system loses brainpower that might otherwise contribute to future growth. When a talented engineer decides, or has no other choice, to build a career in another country, we lose that much more capacity to innovate. And when governments like Canada develop new trading relationships because they no longer want to rely as heavily on Washington, it is the American economy that will suffer in the long run.
This is why credibility is such a valuable asset and why damaging it can be so expensive. Relationships take years to build, and once businesses or countries have spent enormous amounts of money creating alternatives, those decisions cannot necessarily be reversed with a change in rhetoric or even a change in administration.
Some of the current damage is undoubtedly recoverable. The United States remains the world's largest economy, home to extraordinary businesses, universities, capital markets and technological capabilities. Countries will continue to take our calls. But that should not be confused with an unlimited ability to dictate terms.
For more on this topic, I also wrote about the consequences of being an “unreliable partner” in the February 2026 letter.
Checking In on AI
On another (but perhaps not entirely unrelated) topic, the past month or so has seen a flurry of cautionary warnings about AI coming from within the technology and AI sector itself, including the CEOs of Anthropic and OpenAI themselves. Former Microsoft founder and CEO Bill Gates also weighed in with an August op-ed in The New York Times, where he wrote, “AI is changing so quickly that it isn’t clear exactly what will happen next. We’re facing big questions raised by the way the current technology works, the ways people will use it for ill intent, and the ways AI will change us as a society and as individuals.”
My take: it is good for the AI industry and major AI players to be thinking through potential consequences to ensure that we, as a country, are already positioned to deal with issues as they arise. Up to this point, there has been an awful lot of optimism coming out of Silicon Valley without a countervailing appreciation of the fact that highly disruptive technologies bring with them a certain amount of unpredictability and, inevitably, some degree of “creative destruction.”
In other words, the outcomes here are unknowable, and it is critical to approach this next cycle with a level of humility that lets us see clearly rather than looking at AI solely through rose-colored lenses or as unmitigated doomsayers.
We know that AI can be deployed in business to great effect. Airbnb (NSQ: ABNB) offers a great example. In recent earnings reports, Airbnb reported terrific results, but one thing that really caught my eye: Airbnb CEO Brian Chesky described AI as “the best thing” to have happened to their company and attributed recent improvements in revenues and margins more to AI than anything else. Specifically, AI now writes or coauthors 60% of Airbnb’s code, even as revenues have risen 17% despite headcount remaining flat. That kind of operating leverage is quite rare.
And, in the end, doing more with less is a core function of what a capitalist society does. If AI can enable more businesses to do that, great.
That said, if jobs are going to get re-routed (or outright destroyed) on a large scale, we must be prepared for that with active preparation and actionable response plans.
As far as investing in AI goes, we retain the cautious approach we have described for the past few years, with an emphasis on avoiding valuation risk (we can’t make money if we don’t buy at a discount). Airbnb and similar companies notwithstanding, we are concerned that many organizations might be moving too quickly without any real understanding of how this will all shake out; or, like Nvidia (NSQ: NVDA), are pushing into novel territory, like trying to turn chips into an asset class on par with real estate or vehicles.
And importantly, companies do not necessarily need to finance the AI infrastructure boom themselves to benefit from it. Some of the biggest long-term winners may simply be businesses that selectively incorporate these increasingly powerful tools into their existing operations.
But with so much money flying freely around, the overall investment landscape around AI continues to remind me of other Capex bubbles. We tend to think of technology companies as not capital-intensive businesses. Many, like Airbnb itself, aren’t. But for those organizations building data centers and other infrastructure, those are huge capital projects that involve significant economic (and now political) risk.
In short, AI can be great if harnessed and used correctly and in targeted ways, as Airbnb shows. But we also have to be mindful of companies over-leveraged to this Capex cycle. What happens if that funding dries up? The podcast Invest Like the Best recently spoke with business analyst Ben Thompson on that very question, where they discussed, among other things, “whether the massive AI infrastructure buildout can generate returns before the capital runs out.”
A Quick Rivian Update
Finally, a few more interesting items crossed my desk recently involving Rivian (NSQ: RIVN).
I remain fascinated by Rivian’s forward momentum. The R2 continues to receive encouraging reviews, and Rivian is preparing to add a second production shift while raising its production and delivery expectations. They are also still pushing to beat even Tesla (NSQ: TSLA) to full driving autonomy, as this fascinating article in IEEE Spectrum magazine discusses. I also enjoyed a recent interview with CEO RJ Scaringe and a separate conversation with Rivian's Chief Design Officer about the development of the R2.
One detail I found particularly interesting is the company's effort to dramatically simplify the vehicle. Rivian has reportedly reduced the R2's parts count by approximately 45%, an enormous change that has implications for manufacturing complexity, cost and sustainability.
Nevertheless, building cars is extraordinarily difficult, and Rivian still has a tremendous amount of execution ahead of it. For that reason, I continue to think about the company in terms of a marathon rather than a sprint.
One Last Thought
There is a temptation in an environment this noisy to believe that successful investing requires predicting every major development.
I do not think it does.
I cannot tell you exactly where Treasury yields will settle, when relations with Canada will improve, how much of today's damage to American credibility will prove permanent, or precisely how the AI investment cycle will unfold.
What we can do is pay attention and remain disciplined when uncertainty is high.
The bond market is providing information. Canada's response to American trade policy is providing information. Airbnb's operating results are providing information. Rivian's production progress is providing information.
Our job is to listen without feeling compelled to make a bet on every answer.
Sometimes the most valuable investment decision is recognizing an extraordinary opportunity. Other times, it is recognizing that an outcome is unusually difficult to predict and choosing not to stand in the middle of it.
As always, thank you for your trust and the opportunity to invest on your behalf.
With warm regards,
Glenn
