The comeback of the dinosaurs? ๐ฆ
The market is currently in a state of classic bipolar disorder.
On one hand, Intel's report was released and the market reacted with signs of enthusiasm not seen in a long time ๐
Intel stock is up over 20% in after-hours trading,
but the broader reaction is much more interesting -
Investors reacted with great enthusiasm to AMD stock and Arm Holdings $ARM, both up about 6% in after-hours trading. ๐ผ
On the other hand, the undisputed queen, Nvidia, remains on the red and sad side of the board. ๐
Note what happened in the last month!
Intel stock +51% ๐ผ
Arm stock +51% ๐ผ
AMD stock jumped 48% ๐ผ
Nvidia? - 7% ๐
When examining performance since the beginning of the year, the gap widens even further ๐
$INTC stock rose 81% ๐ผ
$ARM stock rose almost 90% ๐ผ
$AMD stock rose 42% ๐ผ
$NVDA stock? - only rose 7% since the beginning of the year ๐ฏ
If you thought these were just dry numbers, you were wrongโ๏ธ
- We are in the midst of a gladiatorial battle between "the brain" and "the muscles," and this story is going to change everything you thought about the chip world. ๐คซ
The battle of architectures: the central processing unit versus the graphics processing unit -
To understand why the market rushed to Intel and its partners, one must discover who does what in this circus. ๐ฉ
To understand the event in depth, let's first explain the difference between the CPU and the GPU -
The CPU - the central processing unit: it is the "general",
it manages the logic - making decisions and moving forces.
The GPU - the graphics processing unit,
it is the "worker with dirty hands." ๐ทโโ๏ธ
- It is the parallel computing engine built to lift heavy weights of data. ๐๏ธโโ๏ธ
Until today, everyone wanted muscles. ๐ช
But suddenly the market remembered that a brain is also needed to manage the business? ๐ง
Nvidia's problem: when muscles are no longer enough?
In the first era of artificial intelligence, everyone was crazy about "training" models. ๐ค
At this stage, Nvidia's muscles were everything. โก๏ธ
The need for immense computing power created an absurd situation where for every one central processing unit, you needed eight Nvidia graphics processing units - simply because without them, you would be left behind. ๐
Nvidia celebrated this vacuum, and the market bought this story blindly. ๐ต
But now, is the paradigm changing? It seems the market no longer buys into this exclusive dependency. ๐ค
The big turnaround?
The real drama began when models are already trained and move to the "inference" stage - the stage where the model truly works for us. ๐ค
Here, the story already changed - the need for graphics processing units plummeted, and the ratio changes to only three or four graphics processing units per central processing unit. ๐
Is the central processing unit returning to be a significant player on the field? โฝ๏ธ
But this is still only a technical change, not a tectonic shift in the financial balance of power - because the need for GPUs is still at least four times that of CPUs. ๐ค
The real turnaround - artificial intelligence agents become the new commanders on the ground:
The icing on the cake here is the transition to a world of "artificial intelligence agents." ๐ฑ
These entities don't just output text - they perform actions, make decisions, and manage processes. ๐ฎโโ๏ธ
This requires much more logic and control and less raw computation.
In simple terms: the central processing unit becomes critical again. ๐น
The market is pricing in an extreme scenario where we will see more central processing units than graphics processing units in data centers...? ๐
This topic came up most directly in Intel's investor call:
Joshua Buchalter, an analyst from TD Cowen, asked:
How to model CPU demand in Agentic AI workloads? Are there any tips for long-term growth in unit or core count?
David Zinsner, CFO, answered him thus:
In the training phase, the GPU to CPU ratio is high, but in the inference phase, it decreases to 3:1 or 4:1.
In the Agentic AI phase, where multiple agents collaborate, the ratio may even reverse.
Who profits and who is left in the dust? ๐
AMD: leads the market with its server processors, and is positioned for the new artificial intelligence needs. ๐ค
Arm Holdings: it does not manufacture chips - it is the architect that designs the plans for energy-efficient processors - and this is what data centers now depend on to avoid burning out. ๐ฅ
Intel: is trying to make an aggressive comeback and is currently focusing precisely on next-generation infrastructure. ๐น
Nvidia: is currently perceived as "one-dimensional" - as the market moves from training hype to performance reality, it is forced to pause and recalculate its path. ๐
So what's the bottom line of the event? The market's reaction to Intel's report is not really about Intel - it's about the understanding that the artificial intelligence architecture is changing. ๐คซ
AMD and ARM stocks jumped in after-hours trading because they hold the key to the new reality created here? ๐
While Nvidia is forced to contend with the fact that even the biggest muscle in the world needs a brain to tell it what to do... ๐ง
We hope we've brought some order to things for you ๐ค