There's no real way to get around it – the rollout of Nvidia's latest lineup of flagship graphics cards (the RTX 50 series) has been a bit of a mess from a PR point of view. While the same huge demand has roared up to snap them off virtual and physical store shelves, the feedback about performance has been really mixed.
Experts and consumers alike have pointed out that the performance gains over similar 40 series cards are modest, but the increased power draw can be really quite significant, making for a questionable upgrade option. At a briefing for its newest lineup of high-end laptops last week, though, MSI was keen to emphasise that the 50-series laptop GPUs might just change the story a little.
I spoke to Andy Chu, Manager of the Product Marketing department at MSI, who had plenty to say about the leaps that these new laptops were making over previous generations both in gaming terms and (crucially) in AI task management. Over the course of a lengthy presentation, he demonstrated how certain local language models could exhibit basically night-and-day differences between 40 series GPUs and 50 series ones.
I had to turn him round to the Nvidia rollout, though, and when I asked him if it had been more challenging than usual to integrate the new cards, he was forthcoming: "We have had some postponements of launch dates, but that’s not just been on Nvidia, but also us wanting to make sure that the latest version of the vBIOS can be properly integrated on our laptops."
As Chu explained, because MSI owns its own factory, this sort of delay doesn't have to cascade too badly, but it is still implicitly extremely sub-optimal. Still, discussing the challenge of this year's integrations, he said: "From the engineers’ point of view, I think it should be challenging because it’s a new architecture, with new GDDR7 [technology]. I think, already, those two factors make the barriers of the new series stronger than our previous experiences."
So, in effect, while things have been a little rocky, MSI is one of a few Nvidia partners that has enough experience to roll with the punches. When it comes to performance, he was extremely positive after I asked him about those depressing power statistics doing the rounds: "On the desktop side, I know the 5080 and 5090 draw more power, but I don’t think that’s the story of the notebooks. You can see the TGP [or Total Graphics Power] of the 5080 and 5090 Laptop is 175 watts."
That necessary internal limit, he explained, means that the amount of power variation going on in the laptop GPU is much lower, and the control that manufacturers have over performance is much tighter. Basically, you're not playing with nearly as much risk. In fact, this has a major upside, says Chu: "According to what we’ve seen so far, more power efficiency doesn’t make it weaker. From our perspective, because they’re using the same wattage, you can get more performance. So maybe power efficiency is getting even better on RTX 50 series."
Now, that little "maybe" does mean that we can't necessarily qualify all of this as an immediate win for Nvidia, showing that its laptop GPUs can turn the tide of public opinion back in its favour this time out. Still, Chu does give a useful reminder that the parallel naming conventions can't change the fact that Nvidia's desktop and laptop GPUs aren't the same cards, but rather just relatives.
So, if you've been looking at specs tables and benchmarks for 50-series cards, thinking you'll steer clear, you might want to hold out for a wave of laptop reviews that will start trickling through in the next few weeks. It's entirely possible that they'll demonstrate that the smart money is on a top gaming laptop this time around, rather than a hulking desktop beast.
