As a behavioural investor, the current episode has been fascinating and clearly challenging at the same time. The “everything bubble” as some observers have described it, has been unwinding in stages since 2021, starting with the most speculative assets (profitless technology companies and crypto) first. These assets were most vulnerable or sensitive to the rise in the cost of capital and withdrawal of liquidity. Perhaps the final phase of this episode will be the AI bubble or the “new hopium” as we have described it.
As Soros said on speculative episodes, “financial markets, far from accurately reflecting all the available knowledge, always provide a distorted view of reality. The degree of distortion may vary from time to time. Sometimes it’s quite insignificant, at other times, it is quite pronounced. When there is a significant divergence between market prices and the underlying reality, there is a lack of equilibrium conditions.”
“Every bubble has two components: an underlying trend that prevails in reality and a misconception relating to that trend. When a positive feedback loop develops between the trend and the misconception, it sets a boom-bust process in motion. The process is liable to be tested by negative feedback along the way. If it is strong enough to survive these tests, it reinforces both the trend and the misconception.”
A subtle but important point above is that there is often fundamental support for a speculative mania. That reinforces the trend (consensus beliefs) and the misconception of reality. It also encourages speculators to buy an asset on the hope that it will be worth more in the future. As we noted recently, the enthusiasm for artificial intelligence (like the internet bubble in the 2000) is warranted regarding its potential to improve productivity in a range of industries.
NVIDIA has some clear edge or advantage over its competitors. We also have no doubt that the euphoric positive speculative feedback loop behind the AI mania could push valuation multiples beyond what seems remotely plausible. However, if returns on capital and profit margins are extremely high, other companies will enter the market to compete. That is how capitalism works.
One way we thought to handicap or bound the potential euphoria is to look back at a few of the survivors of the first technology bubble. CISCO is an excellent example. The company traded as high as 64 times sales (chart 1). CISCO remains a high return (profitable) business. Trend ROE is still just under 30% (chart 2). However, CISCO remains 36% below its record high in March 2000. Perhaps using that benchmark, NVIDIA could double again? We obviously don’t know. Although at 36 times sales, the company has some extremely heroic growth assumptions.
The CEO of Sun Microsystems, Scott McNealy provided some useful context on his company trading at 10 times sales near the peak in 2000. “At 10 times revenues, to give you a 10-year payback, I must pay you 100% of revenues for 10 straight years in dividends. That assumes I can get that by my shareholders. That assumes I have zero cost of goods sold, which is very hard for a computer company. That assumes zero expenses, which is really hard with 39,000 employees. That assumes that I pay no taxes, which is very hard and assumes that you pay no taxes on dividends, which is kind of illegal. It also assumes with no R&D for the next 10 years, I can maintain the current revenue rate. Now, having done that, would any of you like to buy my stock at $64? Do you realise how ridiculous those basic assumptions are? You don’t need any transparency. You don’t need any footnotes. What were you thinking?”
The final point to note about NVIDIA and the semiconductor sector more broadly is that it remains a highly cyclical sector. Even in the current episode, the company has experienced boom and bust phases in demand related to crypto and gaming. While there might be a broader secular trend developing, the semiconductor sector tends to be correlated to the global manufacturing cycle. On the positive side, while that is extremely weak, the Philadelphia SOX index has based and turned up in year-on-year terms (chart 3). Perhaps that is a bullish development for the macro cycle over the coming months?
On the secular trend, as we noted above, market expectations could still become far removed from reality. However, what we do know is that liquidity conditions are considerably less favourable after the policy tightening last year. Moreover, once a bust commences, it tends to be short and steep, because it involves the forced liquidation of unsound positions. Manias, booms, and busts (greed and fear) have been a part of financial history since the South Sea Bubble. There is often some fundamental support. Although it is still important to consider the future growth required to achieve 36- or 64-times sales.