Right now, a massive amount of capital and attention is flowing into one story: artificial intelligence. At the same time, large parts of the economy, energy, food, timber and wood products, homebuilding, metals and mining, paper and packaging, chemicals, and traditional physical retail, continue to generate real cash flows from assets and operations that are relatively well understood. These are not exciting businesses. They do not promise to change the world. But when uncertainty around the dominant narrative is unusually high, they become worth examining more closely.
People are currently acting like we already know how artificial intelligence will play out, who the big winners will be, how much value it will create, and which existing businesses will be disrupted. That level of confidence is unusual. The reality is that the range of possible outcomes remains wide, and a lot of current valuations are pricing in a very optimistic version of events that has not yet been proven.
The Problem with the Current Narrative
When this kind of concentration happens, it is worth looking at the parts of the economy where the key variables are still more observable. Energy, food, timber and wood products, homebuilding, certain transportation and physical retail models, metals and mining, paper, packaging, and chemicals. These are not new or exciting. They do not promise to reinvent daily life. But many of them still generate real cash flows from assets, operations, and customer behaviors that can be analyzed without having to make big guesses about technologies whose commercial shape is still unclear.
This is not an argument that these sectors are full of hidden gems or that they will magically outperform anything related to AI. A lot of them are cyclical, capital intensive, and have long histories of management teams doing dumb things with money when times are good. The point is narrower. When uncertainty is unusually high in one area of the market, businesses where you can at least estimate a reasonable range of outcomes become relatively more interesting to look at.
Why These “Boring” Sectors Deserve Attention
You can examine cost structures, capacity utilization, returns on capital through different parts of the cycle, and competitive positions without needing to forecast how large language models will evolve over the next decade.
Let us be clear though. Most of these businesses are not cheap right now either. In many cases, free cash flow yields remain below what you can get from government bonds. That does not make them uninteresting on principle, but it does mean any margin of safety has to come from somewhere else, whether that is asset values, normalized earnings power across a cycle, or genuinely durable competitive advantages. Simply being “not AI” is not enough.
A Reality Check on Valuations
Note that nothing in this piece constitutes investment advice. Here is a broader look at where current prices sit relative to estimated fair value across three different valuation methods:
| Company | Current Price (mid-Jun 2026) | OE Fair Value | DCF Fair Value | BZ Fair Value |
|---|---|---|---|---|
| ExxonMobil (XOM) | ~$118 | $72 | $105 | $88 |
| Chevron (CVX) | ~$152 | $68 | $98 | $82 |
| Shell (SHEL) | ~$57 | $58 | $92 | $75 |
| Nestlé | ~$98 | $95 | $125 | $108 |
| PepsiCo (PEP) | ~$168 | $145 | $185 | $162 |
| Coca-Cola (KO) | ~$65 | $58 | $78 | $66 |
| Weyerhaeuser (WY) | ~$29 | $32 | $42 | $36 |
| Rayonier (RYN) | ~$26 | $28 | $36 | $31 |
| UPM-Kymmene (UPM) | ~$35 | $38 | $48 | $42 |
| D.R. Horton (DHI) | ~$135 | $115 | $145 | $128 |
| Lennar (LEN) | ~$148 | $125 | $158 | $140 |
| Union Pacific (UNP) | ~$248 | $195 | $245 | $218 |
| Maersk | ~$1,950 | $1,850 | $2,350 | $2,050 |
| Costco (COST) | ~$870 | $620 | $780 | $695 |
| Walmart (WMT) | ~$72 | $68 | $85 | $76 |
| Seven & I Holdings | ~$2,650 | $2,850 | $3,550 | $3,180 |
| BHP Group | ~$55 | $52 | $68 | $58 |
| Rio Tinto (RIO) | ~$62 | $58 | $75 | $64 |
| Nucor (NUE) | ~$155 | $145 | $185 | $162 |
| International Paper (IP) | ~$45 | $42 | $55 | $47 |
| Packaging Corp of America (PKG) | ~$175 | $165 | $205 | $182 |
| Linde (LIN) | ~$455 | $420 | $520 | $465 |
| BASF | ~$52 | $48 | $62 | $53 |
What this shows is straightforward. A small number of these companies are trading at or reasonably close to estimated fair value on at least one of the methods. Most of them are not. Some, particularly the higher-quality names with stronger moats or asset backing, look less expensive than the broader market on a normalized basis. Others remain expensive once you strip away any growth assumptions and just look at current cash generation and required returns. This is normal in cyclical, capital-intensive industries.
The Real Work Most People Won’t Do
The real reason to spend time on these businesses is not because they are safe or because they will beat the AI trade. It is because you can actually do the analysis. You can look at historical returns through different economic environments. You can assess whether a company has pricing power or cost advantages that have held up over time. You can examine how much capital is required just to stay in place versus how much is left for growth or shareholder returns. These are hard questions, but they are answerable with data that already exists.
In contrast, a lot of the current excitement around AI requires investors to take strong views on things that are still very difficult to quantify. How much productivity improvement will actually show up in corporate profits? Which business models will be protected and which will be commoditized? How much capital will ultimately be required to achieve and maintain leadership? These are legitimate questions, but the answers are still speculative for most companies. When large amounts of money are being deployed based on assumptions that have not been stress-tested yet, it creates distortions elsewhere.
Risks Are Still Very Real
That does not mean the traditional sectors are free of problems. Far from it. Many of them have terrible track records when it comes to capital allocation. Management teams often overinvest at the top of the cycle and then struggle when conditions normalize. Some of these industries have seen their competitive positions weaken over time due to globalization, regulation, or technological change that happened gradually rather than dramatically. Others remain highly exposed to commodity prices or interest rates. The fact that they are easier to analyze than frontier AI does not make them good investments on their own. It just makes the risks more visible.
This is why the actual work matters more than the category. Looking at a company like Weyerhaeuser or Nucor or Linde requires you to think about normalized returns, maintenance capital needs, and cycle-adjusted free cash flow. It forces you to ask whether the current price leaves room for the normal variability these businesses have always shown. That kind of analysis is possible. It is just slower and less exciting than betting on transformative technology. Most people will not do it. They will either stay in the crowded trade or assume that anything outside of AI must be undervalued by default. Both approaches tend to end badly over time.
Markets eventually punish overconfidence, even when it is dressed up as being on the right side of history. Right now, a lot of confidence is being placed in outcomes that remain difficult to predict with any real precision. That creates its own kind of opportunity, not in blindly buying traditional businesses, but in being willing to do the slower, more grounded work in places where the variables are still measurable. Whether that work actually leads to good investments still depends on price, discipline, and the ability to accept that even the more understandable parts of the economy have real risks. The narrative will not protect you from those. Only the price you pay and the quality of the analysis will.