On Grave Dancing

2025.08.10

CXII

[Navigating A Bubble, Capital Cycles, The Revenue Trap, Cash is King, Software is No Different]

Thesis: Businesses that don't need demand to go up profit more when it inevitably goes down.

[Navigating A Bubble]

A few weeks ago, I claimed that we're in an AI bubble, but I didn't really tell you what to do about it.

If you're building a business, particularly a software business, I think the solution really isn't that deep--focus on maximizing Free Cash Flow or profits, not revenue.

This prevents you from structuring a business that is contingent on demand continuing to increase at the rate that it is.

And yes, we are talking about “AI” businesses AND “software” businesses here; I think the market has generally conflated the two to the point of them being more or less the same in a lot of cases.

Now, keep in mind, I've never navigated a bubble before, so I can only tell you what we're doing that we think will work. And why do we think it will work? A mix of first principles thinking and, well, the fact that it's what everyone whose made it through a bubble recommends doing.

[Capital Cycles]

To understand what a bubble is, you can look at the most basic part of Econ 101*: Supply & Demand.

If there is more supply for some good or service than there is demand, then the person who has control of the supply can charge more for it.

Think Spice Melange in Dune, Oil & Steel in the 19th and 20th centuries, or GPUs & access to AI Models in real life.

The catch with all of these things (other than Spice, which was contingent on production on one planet) is that when there is way more demand than there is supply, other people will come in and start offering supply.

If I see Daniel Day Lewis making ungodly amounts of money selling oil... well, I'm going to start drilling for oil, and a bunch of other people will, too!

There's a few problems with this:

• Me and the other new market entrants are NOT talking to each other about how much supply we're bringing online

• There is a lag between us deciding to invest in capacity and that capacity actually being available

• Demand for the commodity is inherently unstable - will it keep going up at the same rate? Probably not...

So, all of us are making big investments on imperfect information about the market. It's okay at first, because somebody's making money.

What invariably happens, though, is that at some point supply begins to outstrip demand!

If it not so glorious and contained to a specific industry, analysts and industry specialists will likely be the only ones who notice. However, if it negatively impacts big institutions (like Lehman going under), we might shout bubble!

We also might end up shouting bubble if the “industry” (AI?) is the one that mega cap companies are claiming will grow their already massive revenues by 50%…

Whether this is because an alternative is introduced, there was some distortion artificially inflating demand, all the suppliers over invested, or some combination of these things, now the market has flipped!

Everybody who got too excited about the future prospects of the space and built a business that was dependent on the price imbalance to continue is now in trouble.

On the other hand, the companies that built businesses that were in strong positions regardless of the market imbalance are actually in a really good position when such an imbalance corrects.

The competition will quickly die away and market share and underpriced assets will be their’s for the taking!

Businesses that don't NEED demand to go up profit more when it inevitably goes down.

*I can't actually tell you if this was in econ 101, rumor has it I skipped class & pregamed the final...

[The Revenue Trap]

A lot of businesses that struggle in a down cycle are the same businesses that were optimizing exclusively for revenue in the up cycle.

Some consequences of optimizing for revenue without setting constraints around profitability:

• You'll be okay with a high Cost of Acquiring Customers

• You’ll be okay with high cost of Assets, Leases, or general OpEx, as long as it leads to revenue!

• As a consequence of the above, you'll place lower emphasis on margins

• If you’re in a space that’s not exclusively commoditized, you’ll likely cut corners on quality to get revenue

• You’ll take on capital (debt, vc funding) to grow as much as possible during the upcycle, increasing your risk

Depending on how extreme these factors become in your case, your business will be more fragile to shifts in the supply / demand imbalance.

Once we hit a down cycle, you'll be prone to losing investor confidence, your unit economics will get worse, and your generally bloated operating practices will become dead weight. Thin margins will become liabilities and you'll be asking for a life raft, not rocket fuel.

If you needed another round or more debt to keep going, you'll get it on worse terms, if you get it at all!

The famous last words of Chuck Prince, the CEO of CitGroup in 2007, a year or two before their market cap went down to something like 1/10th of what it was:

Revenue growth at the cost of margins might create short term wins, but it also introduces risk that is no longer under your control.

[Cash is King]

On the other hand, the businesses that do well in a down cycle are the ones that typically prioritize Free Cash Flow (FCF), Ebitda, or some other measurement more aligned with the underlying health of the firm.

In other words, the thing they are optimizing for means they will be okay, even if everything external goes wrong.

In the up cycle, while the attractive supply / demand imbalance is tempting, these businesses don't over extend at the expense of profitability; they maintain healthy margins and keep accruing assets, but only when they are reasonably priced. They don’t raise debt or financing to grow faster than is needed.

Then, when the imbalance goes the other way, and we start hitting a down cycle, these businesses are impervious to market conditions, because they have strong balance sheets and money in the bank and a model that can generate cash, even if margins take a hit. They can purchase assets from their competitors who are fighting for survival for pennies on the dollar. They can deploy their capital to expand and grow & get even more aggressive in taking market share, because now it is less competitive and cheaper to acquire customers.

If they lag on revenue in the up cycle, it doesn't really matter, because in the long run, the compounding nature of survival keeps them afloat & let’s them thrive.

A direct opposite to Chuck Prince's quote is the eerie nickname for Sam Zelle, who was known for buying when his competitors were dying: "The Grave Dancer."

His quote, albeit tied to real estate, sums up the argument clearly:

And he ended up pretty good, selling his firm EQR for $36B to Blackstone in 2006, the largest LBO ever at the time.

[Software is No Different]

But Noah, software is nothing like real estate, or these other markets! You're drawing parallels where none exist!

Thankfully, the argument is really rooted in first principles, not parallels (reasoning from first principles is almost always superior to reasoning from analogy).

Software requires investment in the form of time, effort, money, operating costs, and leasing (apis) or building & owning & maintaining software. And, even though this is more “lightweight” than other markets, it is still very real and requires some lead time.

And, if you say GenAI makes software even cheaper, sure, but it just replaces the capital investment in the stuff human engineers build with a capital investment in GPU farms & API calls to LLMs providers that rent those farms.

That’s part of the reason I’m so comfortable conflating AI and software here; I think it’s applications and the way that market is treating it already has conflated the two.

Money is being thrown at people to make investments in software and AI right now, and builders are spending months and years trying to capture market share. This, even if it is not translating into physical oil wells, is a very real capital investment.

It is being made on the presupposition that the money will be multiplied. That presupposition is based on quite frankly a difficult to gauge demand to extend the capacity of software to replace more people or help the same people be more “efficient.” In theory, and based on how the market is acting, there should be a lot of demand for this. However, in terms of created economic value that is likely to be realized in a reasonable time frame, it’s not yet so clear…

It's okay with me if you believe that this time it's different, and demand will keep going up regardless of how aggressive investment in supply is.

That'll just create more opportunity for companies like BirdDog that prioritize sustainable business models vs revenue at all costs. Catch ya on the way down ;)

Live Deeply,