$MRVL The earnings results are out. Q2 revenue was $2.739 billion, up 37% year over year, slightly above market expectations; adjusted EPS was $0.94 versus the expected $0.93, also a modest beat. More importantly, the data center business grew 46% year over year, accelerating clearly from 27% in the previous quarter. For next quarter, the company guided revenue of $3.15 billion and EPS of $1.10, and it also further raised its full-year and next-year revenue outlook. The FY2027 revenue target was increased to about $12 billion, and FY2028 to about $18 billion. So, looking purely at the fundamentals, there isn’t really anything wrong with this earnings report.
Why did the stock fall by nearly 8% after the earnings release during the overnight session? I think it mainly comes down to expectations being too high. Before the MRVL earnings report, the stock had been rallying from below $200 all the way to above $240. This time, the EPS only slightly beat expectations, and after Nvidia just delivered explosive growth, it’s obviously not enough for the market’s appetite. At the same time, the next-quarter non-GAAP gross margin midpoint declines from this quarter’s 58.9% to around 58%, operating expenses continue to rise, and revenue growth hasn’t fully translated into EPS. In addition, the market originally expected the company to provide more information about FY2029—especially revenues from the Google custom-chip partnership—but management left more details for the October investor day. Overall, I’m more inclined to interpret this drop as a retracement of the run-up before earnings, rather than a重新 pricing of the fundamentals.
As of the current overnight level of $222, based on the FY2028 consensus EPS of $6.24 before the earnings report, that’s roughly a 35x P/E. If after the raised revenue guidance, FY2028 EPS is ultimately revised upward to what I consider a more reasonable range of about $6.7–$6.8, then at the current price the multiple would actually be only around 33x. MRVL’s Forward P/E over the past few years has generally fluctuated between the low 20s and the mid-to-high 40s, so 33–36x is about in the historical middle—slightly above the middle. You can’t say it’s especially cheap, but compared with before earnings, it’s already clearly more reasonable. And there’s still a possibility that earnings expectations can continue to be revised upward.
In terms of strategy, for now I won’t chase and buy directly at $222. I’d rather wait until around $200 to open positions again. After that, if it drops to $180, I would add significantly more.
Admittedly, MRVL is a company with solid fundamentals. Over the past three years, its growth rate has been able to stay around 50% year-over-year, and it occupies an important AI track in custom chips and optical interconnects. The market is willing to award it a high P/E of 50 or even 60 when sentiment is soaring, and when sentiment turns sour, it can drive its P/E down below 25. So my approach to MRVL is the same as to NEBIUS: I trade in ranges—take profits at highs, then enter at lower levels. This time, from 240 to 250, I’ve almost cleared all of my MRVL positions. Now I’ll patiently wait for another opportunity to get in again.
