Marvell Technologies (NASDAQ: MRVL) stock is up a cumulative 24% after the announcement of results. Some part of this is a result of those results announced - they're good, ahead of expectations. As we continually remind, what changes stock prices is new information. So, yesterday's MRVL price reflected the balance of opinions about what the coming results were going to be. Those results prove to be better than expectations and the price will rise. That's just how it works - beating expectations is that new information which changes the stock price.
Rather more important to this leap is the forecasts for future growth. As we say the earnings were good but not that good. “Marvell Technology (MRVL) late Thursday edged above Wall Street's targets for its fiscal first quarter” doesn't drive a stock price up 24%. “Q1 Non-GAAP EPS of $0.31 beats by $0.02. Revenue of $1.32B (-9.0% Y/Y) beats by $20M.” Sure,m that's good, that's nice even, but it's not good or nice enough to add 24% to a near $50 billion market capitalisation.

Marvell Technology stock price from NASDAQ
The forecasts for near term revenue growth are also nice but again not that good. What's really driving this is the realisation of what Artifical Intelligence is going to do to Marvell. Or, to be more accurate here, Large Language Models - this is really about LLM, not AI.
The basic Marvell business proposition is making systems on a chip. Or, to be more simplistic, encoding software into silicon, designing down from many pieces of silicon to fewer or even one. The major market for this is in the data market. This isn't chips for PCs or phones, this is chips for data centres. OK, so what do we know about LLMs? They're huge, vast pieces of code which run billions of iterations on vast quantities of data centre resources. As AI expands then the market for encoding much of that down into the silicon, thereby saving compute resources, is going to be vast. That's the market for Marvell that the stock market has just latched on to.
It's also necessary to be ever so slightly sceptical here. For as with any technological revolution there's no certainty that the current players are going to be the winners of the next round of the game. Sure, we can handicap the race, those already in the business have some advantages and so on. But there's no certainty here at all. After all, if there were then how did Marvell push aside Intel in this race for the data centres? It's not as if Silicon Valley is short of capital these days, nor of people willing to spend it upon an attempt to upend the established order.