Well, Nvidia is making more money than ever but they are down about 2%. Investors just seem to want to sell tech stocks right now and for AMD to be up does stand out imo.
It is not about how much you make now. People invest based on what they think is the potential of the company. There's been a lot of concerns around the sustainability of spending on more AI hardware because all these AI companies have yet to find a way to make money out of it. Even OpenAI is bleeding cash constantly. So Blackwell is going to be faster, but companies can't just replace their entire AI hardware infrastructure because its sunk cost. Hence, companies will still buy from Nvidia, but its not going to be like a few years back where they are building up their AI infrastructure.
I’m actually wondering how long before we learn the dGPU business is done for. It’s been a huge money loser for years. I wouldn’t be surprised if it’s part of the 10B cost reduction. While technically and asset, it’s not exactly worth much on the open market.
It is losing money, but I feel is one area which Intel should absolutely focus on. As you can tell, the likes of AMD and ARM processors have fairly good iGPUs, which they can "bundle" with their CPU sales. This is especially lucrative for the mobile/ laptop space. If Intel goes back to those dumb iGPU, it is not possible for them to achieve low power usage on laptops/ mobile devices powered by Intel chips. GPUs is not always about gaming as there are other usage.
Personally, I don't think Intel will become irrelevant, but will certainly face an uphill climb going forward. All the "time bombs" set by prior CEOs are starting to blow up one after another. The main failures are, (1) sitting on their fab issues for way too long allowing competitors to overtake them, and, (2) sitting on the brand name with little innovation or change, a.k.a. playing it too safe. For example, every generation of CPU is always minor improvements, same number of cores, severely gated features that you need to pay substantial premium to unlock, etc. It was easy money for Intel, until an able competitor starts bucking the trend and throwing massive cores counts and desirable features at a lower price point. I feel the biggest misstep (personal opinion) is for Pat to triple down on fab business. All their competitors are not tied down to a fab, meaning if there's issues with a fab, they have the option to use another. Currently, the issues plaguing Raptor Lake and Intel's decision to use TSMC for their next gen CPU just tells me something is not working well with their fab despite all the fluffy fab numbers, i.e. Intel 4. Like I have been saying it is not the number that counts, but how well it works in the final product. This is going to negatively impact their effort to attract big companies to use their fab, and even if they manage to get some contracts, they won't be able to charge like TSMC. So the fab business is going to be a serious drag to Intel, like GF was for AMD back then. I do think the fab losses will widen because of the ongoing expansion plans, since they need to hire more people, obtain more equipments, and pay for running cost. Those government payouts may be cushioning some of the impact now, so I think the worst is yet to come. As the company looks weaker, it will also translate to higher borrowing cost to finance their investments. So I can understand the need to drastically reduce cost.
, now my next passion is 4K 240hz (currently using LG OLED C4 144hz)
