AI Boom Leverage, Oil Supply Risks, Travel Boom, Healthcare Stocks, Treasury Bond Buybacks, Home Insurance Deductible & More

The AI boom is starting to look a lot more leveraged than investors realize

There is a growing risk in the AI infrastructure buildout that isn’t getting nearly enough attention: how much of this spending is being financed, and how much of the risk is sitting off the balance sheet.

The headline numbers around capital expenditures are already staggering, but what concerns me more is what sits underneath them: joint ventures, off-balance-sheet financing arrangements and leases that haven’t even commenced yet.

In other words, some of the financial obligations associated with this AI buildout aren't necessarily showing up in today's debt figures. And the spending is enormous. Goldman Sachs analysts estimated that hyperscalers have combined lease commitments for data centers, R&D facilities, offices and equipment of $1.5 trillion, up from about $200 billion five years ago. This includes about $1 trillion of “uncommenced” lease commitments, which are not yet shown in financial statements but will result in future payments. This pairs with consensus forecasts that hyperscaler capital spending alone will surpass $1 trillion per year from 2027 onward and there are with no clear signs of moderation. According to a multi-asset credit strategist at PIMCO, the AI capex cycle is, adjusted for inflation, on track to be the largest investment cycle since the 19th-century railway construction.

The problem is what happens if the revenue doesn't grow fast enough to justify the investment. This is where Steve Eisman’s warning is particularly interesting. Eisman, who became famous for betting against the housing market ahead of the financial crisis, believes the AI boom has become increasingly dependent on just two companies: OpenAI and Anthropic.

According to Eisman, those two companies account for roughly 70% of AI-related revenue at Microsoft, Amazon, Alphabet's Google and Oracle, and potentially 25%–35% of their overall cloud revenue.

That creates a concentration risk that investors shouldn't ignore. If OpenAI and Anthropic continue growing rapidly, the economics of all this infrastructure can work. But what if they don't?  Eisman believes one of the biggest threats could come from China.

Chinese open-source and open-weight AI models are significantly cheaper, and if they continue gaining market share, the industry could face something that investors haven't really modeled into these enormous infrastructure investments: an AI price war.

If the price of AI inference and cloud computing falls dramatically, the companies that have spent hundreds of billions building capacity could find themselves with a serious problem.

The infrastructure doesn't disappear just because pricing does. The debt doesn't disappear. The leases don't disappear. And the depreciation expense certainly doesn't disappear.

Another major concern given all the commitments from OpenAI is the turnover the company has seen. The company recently announced that Chief Revenue Officer Denise Dresser is leaving less than a year after joining the company. Dresser had brought more than a decade of Salesforce experience and was viewed as someone with important enterprise expertise as OpenAI tried to compete with Anthropic.

She isn't the only senior executive to leave. Fidji Simo stepped down from her product and business role, and several other executives including COO Brad Lightcap departed earlier this year.

Executive turnover doesn't necessarily mean something is wrong. Fast-growing companies go through enormous amounts of change. But when two companies are potentially responsible for such a large percentage of the revenue supporting an enormous AI infrastructure investment cycle, leadership stability becomes much more important.

There are a lot of things that need to go right to justify the enormous amount of spending in the AI space. And increasingly, there seem to be more and more question marks that investors need to consider. I’m not saying the AI boom is over. I’m saying investors should spend a lot more time asking who is financing this boom, who is ultimately responsible for the obligations, and what happens if the economics of AI change.

Refined oil could be in jeopardy over the next 6 to 12 months

U.S. refineries are currently operating at historically high utilization rates at around 96.5%. Aside from July 25 of this year, when utilization briefly reached 97.2%, the last time refineries were operating at this level was in 2018, when utilization hit 96.6%.

Part of the problem is our own doing. California politicians deserve a significant amount of blame. Over the past 20 years, nine of the 12 refineries that have closed in the United States have been located in California. At the same time, the push toward electric vehicles led many refiners to avoid investing the billions of dollars required to build new refining capacity. Now, we not only lack significant new capacity, but some existing refineries are also in need of repairs and upgrades.

That leaves us particularly vulnerable considering we are in hurricane season, which runs from June 1 through November 30. A major hurricane hitting the Gulf Coast could knock out anywhere from 10% to 30% of U.S. refining capacity, depending on the severity and location of the storm. That could put enormous pressure on already-tight supplies of gasoline and diesel.

And supplies are already below normal. Global inventories of refined fuels, which primarily consists of gasoline and diesel, are estimated to be roughly 130 million barrels below normal levels for this time of year.

This isn't just a U.S. problem. Gasoline and diesel are globally traded commodities, and the global refining picture has changed significantly because of the war in Ukraine. Ukraine has reportedly knocked out roughly 30% of Russia's refining capacity, while Russia has also reduced exports of refined products as it prioritizes its own domestic needs.

The United States is a free market, and American businesses can trade refined products on the global market. U.S. refineries currently export roughly 900,000 barrels per day of gasoline, while diesel exports recently reached a record 1.9 million barrels per day.

This is why having millions of barrels of crude oil doesn't necessarily solve the problem. You can have all the oil in the world, but if you don't have the refining capacity to turn it into gasoline and diesel, that oil is of limited use to consumers.

As an investment firm, we're always looking for the other shoe that could drop. This is one that concerns me. If we get a major hurricane over the next few months and refining capacity is reduced even temporarily, the impact on gasoline and diesel supplies could be significant. A disruption lasting only a week could be enough to send energy markets into a tizzy, particularly given how tight inventories already are.

The irony is that we spent years aggressively pushing toward electric vehicles while underinvesting in traditional refining capacity. EV adoption hasn't progressed as quickly as many expected, but the refining infrastructure we depend on for gasoline and diesel hasn't magically expanded either.

Now we're heading into hurricane season with historically high refinery utilization, below-normal refined fuel inventories, limited new refining capacity and a global market that is already facing disruptions. That's a combination worth paying attention to.

Americans are traveling more than ever

If you’ve noticed how busy airports have been lately, there’s a reason: Americans are traveling more than ever, and there’s little sign of that slowing down.

One reason is wealth. Americans collectively hold roughly $100 trillion in wealth. They’re also living longer and, perhaps more than any previous generation, are choosing to spend their later years enjoying life, traveling, and creating experiences.

Back in the 1970s, 80s, and even the 90s, Americans seemed more content to stay home, spend time with family, and enjoy their homes. Fast-forward to today, and travel has become a much bigger priority.

Trips to Europe reached a record 24 million in 2025. While some Europeans certainly aren’t thrilled with the influx of American tourists, those visitors are having a major economic impact. Americans accounted for roughly 15% of luxury sales across Europe.

Of course, not everyone is happy about the crowds. Barcelona, which sees roughly nine times as many visitors as it has residents, has seen protests against tourism, including protesters spraying tourists with water. I guess on a hot day, that might not be the worst thing.

The change in travel habits is pretty remarkable. As recently as 1990, only about 5% of Americans had a passport. Today, that figure is around 50%, giving Americans far more ability to travel internationally.

So who is doing all this traveling? Women 55 and older account for roughly 24% of travelers to Europe and other international destinations. I don’t know about you, but that doesn’t surprise me.

What does this mean going forward? If this trend continues, it could have a meaningful impact on the economy. Airlines, hotels, restaurants, and other businesses tied to travel should continue to benefit from Americans prioritizing experiences.

I do believe we’ll eventually see an increase from the historically low levels of spending on home remodeling and repairs. But I also wonder if that trend could eventually be constrained as people choose to spend $10,000 on a trip to Europe rather than $10,000 on a kitchen remodel. And with all these Americans traveling around the world, I have to wonder how many have taken the time to see the incredible places we have right here in the United States.

I’m talking about the Grand Canyon, Yellowstone, the giant redwoods of Northern California, or our nation’s capital in Washington, D.C. We may be traveling more than ever but maybe we should remember that there’s still plenty to explore right here at home.

Do you have any healthcare stocks in your portfolio?

You may be thinking, “What a boring investment.” And it’s true, healthcare stocks have performed poorly over the last few years. But if your portfolio is heavily concentrated in high-risk areas like AI and technology, you may want to consider adding a couple of healthcare stocks.

I know that could mean giving up some performance as tech stocks skyrocket… or, I should say, if they continue to skyrocket. But healthcare as a portfolio diversifier could be a good option for people that don’t want to sell all their tech winners.

Look at the recent history as it seems when tech stocks rallied, healthcare struggled, but on the other side of the coin, from late June to late July, as the AI rally cooled, healthcare outperformed technology by roughly 30 percentage points.

Part of the reason is that semiconductors are highly cyclical businesses, while healthcare is much less cyclical. We need healthcare in good times and bad.

We saw another great example in 2022. Inflation was rising, interest rates were climbing, and the market sold off, with the Nasdaq falling more than 30%. Yet healthcare outperformed by roughly 30 percentage points.

This is why when we build portfolios for our clients, we don’t just look at the individual companies. We look at how those companies correlate with one another and how they could react under different market conditions.

It’s also why we don’t want to be overly concentrated in any one industry. You should take a look at your portfolio and ask yourself: Is it balanced, or is everything likely to fall at the same time when the next downturn comes? Sometimes the most “boring” investment in your portfolio can be one of the most important.

The Treasury just doubled its bond buybacks. But how much does it really change?

The Treasury announced this week that it will at least double the maximum size of its long-term bond buybacks from $2 billion to at least $4 billion per operation, beginning in September. The move is designed to improve liquidity in the longer-dated Treasury market after long-term yields surged.

The market reaction on Wednesday was significant, but I think it’s important to put the size of this move into perspective.

The U.S. national debt has now topped $40 trillion. So while going from $2 billion to $4 billion sounds substantial, $4 billion is just 0.01% of $40 trillion.

This isn't debt reduction. The Treasury is essentially buying back certain longer-term securities and managing the composition and liquidity of the debt. It doesn't address the underlying fiscal problem.

And this is where I think the bigger issue gets interesting. The government has increasingly leaned toward issuing more shorter-duration debt. That can make sense when short-term borrowing costs are lower, but it also means a larger portion of the debt needs to be refinanced more frequently.

That creates interest-rate risk. If rates remain elevated, the Treasury has to continually roll over maturing debt at higher rates. The government may save money today by borrowing shorter, but it potentially increases its exposure to what happens to interest rates tomorrow. It's similar to choosing a short-term adjustable loan over locking in a long-term rate. You might get a lower rate initially, but you have to refinance much more often.

We have already seen how the refinancing risk impacts the government as the average interest rate on government debt climbed from 2.23% in 2016 to 3.45% in 2026. During the craziness of Covid in 2021, the average interest rate was 1.61%. These higher interest costs and larger debt balance have pushed interest expenses above $1.2 T and servicing the debt is now more costly than major categories like defense and Medicare. At that level it is the second-largest federal expense behind only Social Security.

The big problem is there doesn’t appear to be an end in sight as the government, which includes both political parties, continues to spend money and low-interest rate debt will continue to mature. As of Q3 of Fiscal Year 2026, close to 33% of US publicly held debt was set to mature within 12 months and the average maturity as of June 2026 was 71 months.

The Treasury has plenty of tools to manage the debt market, and this buyback could certainly help liquidity and temporarily reduce pressure on long-term yields. But there is a big difference between managing the debt and solving the debt problem. At $40+ trillion, the numbers are simply too large for a $4 billion-per-operation buyback program to materially change the underlying fiscal picture. The real solution isn't finding a better way to refinance $40 trillion. It's eventually getting the growth of the debt and deficits under control.

Financial Planning: Do You Have the Right Home Insurance Deductible?

Homeowners insurance across California has been an ongoing issue, with premiums rising sharply and some policies being canceled or not renewed by insurers. If your goal is to reduce the cost of homeowners insurance, one strategy worth considering is choosing a higher deductible and paying for smaller losses out of pocket. Homeowners insurance is generally most valuable for protecting against low-probability, high-impact events, such as a major fire or severe property loss, rather than functioning as a reimbursement program for routine repairs and relatively small claims. Filing a small claim may provide an immediate financial benefit, but it also becomes part of your insurance history and can affect future premiums or your ability to obtain coverage. In other words, the reimbursement from a small claim may ultimately be offset, at least in part, by higher future insurance costs or the difficulty of finding affordable coverage. For homeowners who have sufficient savings to absorb smaller losses, accepting a higher deductible can therefore be a sensible way to lower annual premiums while preserving insurance for the truly catastrophic losses that could otherwise threaten their financial security.

The Whatnot app appears to be more gambling than shopping

At first glance, I wondered: What’s the difference between Whatnot and eBay? They’re both auction platforms, so how could people actually lose money?

Well, that assumption appears to be way off base. If you’re unfamiliar with Whatnot, the platform launched in 2019 and allows people to bid around the clock on trading cards, toys, designer goods, fashion, and even instant ramen. In 2025 alone, the platform added 20 million new accounts, and the company is reportedly on track to reach $1 billion in revenue in 2026.

In 2025, roughly $8 billion worth of goods passed through the platform, while creators hosted more than 550,000 hours of live shows each week. It’s mind-boggling what the company has built and how effectively the platform can pull people into a high-pressure auction environment.

The average auction lasts just 10 to 30 seconds. That’s where the excitement begins. One example I came across involved a trading card that started at $1. Within 45 seconds, the auction closed at $170. That’s exciting because everything is moving so quickly. But Whatnot also gives users the option to swipe a yellow bar at the bottom of the screen, which automatically submits a new bid a few dollars above the current offer.

In other words, it becomes incredibly easy to keep bidding without really stopping to think about what you’re actually paying. In the case of that $170 card, a quick online search showed that similar cards regularly sell for less than $80.

There have been similar examples involving people trying to find undervalued Morgan Silver Dollars, only to send them to professional grading services and discover that the coins were worth far less than what they paid on Whatnot.

Then there’s another feature called “breaks.” Buyers pay a set amount for a designated share of an unopened box of trading cards. They have no idea what card they’re going to receive, but they’re hoping to pull something worth hundreds or even thousands of dollars. The person running the break is typically a high-energy salesperson who keeps the excitement going and encourages people to continue bidding.

As the energy builds, it becomes incredibly easy to get caught up in the moment and lose sight of how much money you’re actually spending. And this is what concerns me. We already live in a world where gambling and high-risk speculation have become increasingly normalized from options trading to prediction markets to sports betting.

Now you can add platforms like Whatnot to the conversation. I’m not saying every purchase on Whatnot is gambling. Plenty of people are buying legitimate products and getting fair value. But when you combine ultra-short auctions, automatic bidding, mystery products, high-pressure sales tactics, and the possibility of a huge payoff, the line between shopping and gambling starts to look pretty blurry.

There have been plenty of stories of people getting in over their heads financially and ultimately having to sell investments or other assets to deal with their debts. Whatnot is still a private company, so investors can’t currently buy the stock.

But I think it’s worth paying attention to companies like this not just because of their growth, but because of what their business models tell us about how increasingly comfortable people have become with gambling and speculation. The question isn’t just whether Whatnot can reach $1 billion in revenue. It’s how much of that revenue is being driven by genuine consumer demand and how much is being driven by the excitement of the next bid.

Why Do Lumber Prices Keep Spiking?

Lumber is an important material for building homes, as well as for remodeling and other construction projects. While prices are nowhere near the crazy highs of around $1,400 per 1,000 board feet reached about five years ago, lumber is still trading well above historical norms, around $550–$620 per 1,000 board feet.

So why are lumber prices rising? This time, it’s not because of the overwhelming demand we saw during the housing boom and pandemic. Instead, the bigger issue is tight supply and low inventories.

The tariffs on Canadian lumber have contributed to the supply constraints. The good news is that U.S. producers have added more than 8 billion board feet of sawmill capacity over the past decade and now account for roughly 75% of the domestic market. Meanwhile, Canada’s share has fallen from around 30% to 19%.

And there is an easy way for Canada to potentially reduce some of these trade pressures: buy more American goods. In 2025, the U.S. ran roughly a $46 billion trade deficit with Canada. More balanced trade could benefit both countries.

But tariffs aren't the only problem. A number of U.S. sawmills have closed in recent years after the prolonged housing downturn reduced demand enough to make some operations uneconomical. That has left the industry with less capacity just as supply constraints have become more important.

Then there are natural supply disruptions. Wildfires have destroyed timber, while the mountain pine beetle has damaged trees by boring into and killing pine trees.

There isn’t an easy solution to all of these problems. We can't simply eliminate wildfires, and controlling the pine beetle is difficult. But trade is something we can influence.

If the U.S. and Canada can find a way to make trade more balanced while maintaining adequate domestic lumber production, it could help increase supply and ultimately put downward pressure on prices.

For now, though, the lumber market is a good reminder that prices don't always spike because demand is booming. Sometimes, it's simply because supply has been constrained for too long.

The Price of Chicken Is Falling and It Could Keep Falling

It doesn’t seem that long ago maybe a year or two when we were talking about the rising price of chicken and how it was creating challenges for restaurants and consumers.

Fast-forward to the summer of 2026, and it appears the major poultry companies may have overplayed their hand. After fighting through low inventories caused by bird flu, chicken production is now far outpacing demand.

During the first six months of 2026, approximately 4.9 billion chickens were slaughtered, an increase of 3% from the same period last year and 6% from five years ago. Production has benefited from better weather and relatively few cases of bird flu.

And because chickens have such a short production cycle, supply can change quickly. Approximately 90% to 95% of chickens are ready for slaughter within just six to eight weeks.

But there’s another problem for the chicken industry: Americans still love their beef. Even though beef prices have reached record levels over the past year, beef sales were still up 2%.

So what does this mean for consumers? It means you can get a high-quality source of protein at lower prices as the huge supply of chickens works its way through the market.

But don’t expect these low prices to last forever. Over the next six to nine months, falling prices should eventually stimulate demand, and chicken producers may respond by pulling back on production. Higher demand combined with lower supply could eventually give chicken companies the ability to raise prices again.

For now, though, it might be a good time to enjoy some chicken. Maybe chicken Cordon Bleu for dinner tonight?

The Bottled Water Wars Put the Cola Wars to Shame

The Cola Wars happened back in the mid-1970s through the early 1980s, with Coke and Pepsi as the two main players, along with a handful of other competitors.

Today’s bottled water wars make the Cola Wars look simple. There are roughly 2,800 bottled water companies in the United States, and new brands seem to pop up every day. Make no mistake: bottled water is big business. Americans buy roughly 50 billion individual bottles every year, generating an estimated $45–$50 billion in revenue. That’s more than 16 billion gallons of water.

Younger generations, in particular, are increasingly focused on drinking what they believe is the purest water possible. And they’re willing to pay some pretty ridiculous prices for it.

One of the newest brands gaining popularity among affluent and health-conscious consumers at private clubs is Loonen Still Water. A bottle costs around $4 in stores, but at high-end clubs such as Casa Capriano in Los Angeles, you could pay more than $6 for a single bottle.

The funny thing? This isn’t some magical water from a remote island. It’s filtered water from a Southern California spring.

And Loonen isn’t even the most expensive bottled water out there. Check out a six-pack of Hallstein water, and you could be paying around $70. Granted, the bottles are 25 ounces, compared with the more standard 16-ounce bottle.

But is there actually a health benefit to paying these prices? According to Roger Fielding, a professor at Tufts University, there’s no evidence that these high-end bottled water brands provide any biological advantage over other bottled water.

And it gets even more interesting. Samantha McBride, a professor at the University of Pennsylvania, says it’s unlikely that any bottled water is completely free of traces of microplastics. That’s especially ironic because many consumers are paying a premium specifically because they believe they’re avoiding contaminants.

Yes, Loonen uses glass bottles, but even the caps contain a plastic lining. I do believe bottled water can taste better than much of the tap water in the United States, and depending on where you live, there can be legitimate reasons to prefer filtered or bottled water.

But the bottled water industry has become a perfect example of the power of marketing.

Companies have convinced consumers that water, which is essentially one of the most basic commodities on Earth can somehow become a luxury product worth $4, $6, $10 or even more per bottle.

The Cola Wars were about choosing between Coke and Pepsi. The Bottled Water Wars are about convincing us that not all water is created equal.