Trending
This past week, I was invited onto a network at 4pm EDT. Because that time coincides with the market’s close, I came prepared not only to discuss the day’s activity but also the earnings reports that would be arriving during that timeframe. As we wrote yesterday, investors have been all too quick to punish companies who deliver EPS beats but miss on another metric. The key companies that I discussed yesterday met that fate, along with many others. Why?
For the show in question, I was asked specifically to focus on Arm Holdings (ARM), Airbnb (ABNB), and Robinhood (HOOD). I obviously could not comment on a company that competes with mine, but the first two were in-bounds. All three beat their consensus EPS estimate – some quite handily – but all are lower this morning even as markets are modestly higher. I consider this an ominous sign.
The most jarring example is ARM. EPS was $0.36, 20% above the $0.30 estimate; revenues came in at $928 million, well above the $880.4 expectation. That was the sort of result that should have encouraged AI-focused investors. But instead, they chose to harp on “disappointing” guidance. Mind you, that guidance was itself a short-term positive and at worst a longer-term push. Revenue guidance for the June quarter is $875-$925 million, well ahead of the $868 average estimate; EPS guidance for that quarter is $0.32-$0.36, also comfortably above the $0.31 consensus.
So far, so good, right. The nitpicking regards the 2025 guidance. Earnings should be OK. Guidance was for $1.45-$1.65, putting the midpoint above the $1.53 consensus. The problem, such as it is, comes from revenues. Those should be $3.8-$4.1 billion, putting the midpoint at $3.95 – thus below the $4.01 bn estimate. This would represent 26% growth over 2024 expectations, but no matter. When a company is priced to perfection, there is essentially no room for error.
Worse, have we seen the return of the dreaded “whisper number”? This was a feature during the high-flying internet era. Investors had an extra set of expectations above and beyond those published by analysts. A company could beat its numbers and get punished anyway, mainly because those extra expectations were bolstering premium valuations. This sounds all to similar to what we’re seeing this cycle.
That makes me nervous, especially ahead of Nvidia’s (NVDA) earnings on May 22nd. We have come to expect that NVDA will beat expectations, raise guidance, then beat those raised expectations next quarter. This creates an amazing virtuous cycle, one that has benefited investors immensely. But even that stock had quarters that led to modest selling after beating and raising. Sure, the stock rose 16.4% after its last report, but it fell -2.5% in the prior quarter and was essentially unchanged in the quarter before that.
It is too early to say whether there is indeed a “whisper number” for NVDA’s next report. But it would be concerning if there was even the slightest whiff of concern about a slowdown in the growth rate of AI investment. Notice I said slowdown in growth. It is premature to think that the cycle itself might be coming to an end. Yet in light of Mark Zuckerberg’s comments during that company’s post-earnings conference call, implying that it could take years for Meta Platforms’ (META) AI investments to pay off, it seems logical to expect some easing of the rapid pace of that spending as companies figure out how to get AI to improve their bottom lines. That’s what happened in the early days of the internet. It’s more of a speed bump than a wall, but NVDA and many of its peers are not priced for speed bumps. Keep your ears wide open for the whispers.