In the last 24 hours, I learned that Indians have a network of worldwide diaspora group chats with tens of thousands of members, which they use to hunt down any social media posts that criticize them, then will create fake, A.I. images of your face & name on a sex offender registry list, then flood your entire timeline with thousands of copy/pastes of this same photoshopped image.
Here are the comments on just ONE of my posts that they raided yesterday:
Why have memory stocks gone up so much?
Because the smarter AI gets, the more memory it consumes, the more bandwidth it needs, and the more data it has to keep moving.
The problem is that this memory cannot be produced easily, even if the industry wants more of it. HBM is constrained by process complexity, packaging, yield, equipment, and supply chain bottlenecks. DRAM capacity also cannot be expanded overnight.
As a result, AI infrastructure is now paying a massive cost. The industry calls this the “Memory Tax.”
It is the tax AI has to pay as it scales. And this is not just a short term cycle. It is increasingly becoming a structural problem.
The industry is already moving. Simply buying more DRAM is not enough, so companies are starting to explore multiple ways to use less DRAM, place memory closer to compute, or bypass traditional memory architectures altogether.
In this article, I break down how the industry is trying to work around the Memory Tax, which technologies are likely to benefit, and how the power struggle between Nvidia, hyperscalers, and the big three memory companies could unfold.
If you want to understand why memory stocks have risen so much, how much further they could go, and which memory related technologies may benefit from this shift, I believe this article will give you a clear framework.
If you are investing in memory, I highly recommend reading it.
Full article: https://t.co/QIIchTV9bz