Aug 30, 2024
“With enough money, you can bubble sort anything.”
The Nvidia results were eagerly awaited. This was entirely natural as Nvidia’s performance accounts for about 1/3 of the YtD gains on the S&P500. However, the results were not well received, which might be considered puzzling given Nvidia’s quarterly numbers met or beat analysts’ estimates on nearly every measure. The problem was not the numbers but what was expected: the “whisper” number. Nvidia investors have grown accustomed to blowout quarters, and these latest numbers were not blowouts. Guidance was more balanced than investors might have hoped: gross margins declined slightly and the new Blackwell processor lineup has proved more challenging to manufacture than hoped. If you are used to a succession of shockingly positive surprises, anything less is going to disappoint.
One might argue that so it is with everything associated with AI. We expect such an amazing future that it is now difficult for AI to exceed expectations. Which is why the Hindenberg bear report on Super Micro (reportedly NVDA’s third biggest client) was so interesting. What if we now simply expect too much too soon? MIT Prof, Daniela Rus says the current advances we see and extrapolate from are based on old ideas applied to huge data sets with vast computing power, and the statistical methods we are using might have lower ceilings than we imagine. Would be a shame if all those PE-funded AI start-ups (some of which boast Nvidia as an investor) had down rounds.
Bob Elliott makes the case that it’s not just AI that has an expectations problem: US stocks in general have an expectations problem. Similarly, one might focus on the apparent inconsistency of interest rate and stock market expectations. The revisions to 2nd quarter GDP were taken as supportive of the Goldilocks scenario – “This economy is neither in recession nor is it at risk of an imminent end to the current business cycle” but can we really trust what we think we see?
Even if we can trust the data, this “equilibrium” might not be very stable. The Conference Board suggests that what you think about the economy might be highly dependent on how much equity you own. And stock markets can both giveth and taketh away.
Dollar General’s results made much the same point. The company flagged “financially constrained” consumers as partly to blame for a disappointing report. This, as Donald Rumsfeld might say, is now a “known, known”. The “known, unknown” we are interested in is the near-term level of policy rates, and here Bostic weighed in, suggesting it “may be time to cut,”, which to our mind is a little underwhelming given what is already priced. Large debt burdens can be difficult for markets to digest without a supportive rate environment. Which made us wonder whether this was an omen of sorts.
Calvin and Hobbes by Bill Watterson for August 16, 1986 – GoComics