norman a gabriel

trust the process!! Seize the opportunity!! Believe in the Grind !!

[norman a gabriel @rhoman1996]

$AMD

[samurai @Samurai__Stock]

https://t.co/VLDKM32oND

[norman a gabriel @rhoman1996]

@HolySmokas $AMD if earnings keep accelerating, free cash flow keeps growing !! why not ? from a market cap comparison not even close to $NVDA MC

[norman a gabriel @rhoman1996]

@dannycheng2022 luck is when opportunity meets preparation !!Trust the Process !

[norman a gabriel @rhoman1996]

@MikeLongTerm @amitisinvesting Agree semi’s more cap ex !

[norman a gabriel @rhoman1996]

@TLAMB91 Love it ! Nothing goes up in a straight line ! Buy low sell high lol

[norman a gabriel @rhoman1996]

$NVDA market cap 5 trillion $AMD market cap 800 billion !!

[Mike @MikeLongTerm]

$AMD is winning while $NVDA is scaling back CPU-GPU Memory 🚀🚀🚀

This is a MASSIVE win for AMD because NVIDIA's last-minute SOCAMM capacity cut on the Vera CPU (in the Rubin NVL72 rack) signals execution friction, de-spec'ing, and potential ramp risks on its 2026 flagship AI platformm, exactly the kind of opening AMD's MI350/MI400 series and Helios rack-scale solutions need to gain significant market share on Training.

Rubin was sold as an evolution of Grace/Blackwell extreme co-design with Vera CPU delivering 3x memory capacity and 2.4x bandwidth over Grace, all coherently tied to Rubin GPUs. Dialing back SOCAMM makes the platform look less "unbeatable" on memory-bound workloads (large models, agentic inference, long-context). AMD has been hammering its HBM3E/HBM4 edge: MI350 series already offers 288GB per GPU (vs. Blackwell's ~192GB), and MI400 jumps to 432GB + 19.6 TB/s bandwidth. AMD's rack-scale Helios doesn't rely on the same unproven modular CPU memory tech, so it avoids this exact risk.

Hyperscalers & AI Native (Microsoft, Meta, Google, Amazon, Softbank, Oracle, OpenAI, Anthropic...) hate single-vendor risk. Any visible de-spec or delay on NVIDIA's next-gen (Rubin follows Blackwell's massive ramp) makes them to buy much more AMD chips/racks right now. AMD's MI350/MI355X already claims better TCO and memory capacity vs. B200/GB200 for inference, with software catching up fast via ROCm/UALink. A Rubin hiccup in 2026 gives AMD's MI400 (and full rack) time to prove itself without NVIDIA dominating the narrative.

SOCAMM issues echo earlier delays (originally eyed for Blackwell Ultra, pushed to Rubin). NVIDIA procures some memory directly for VVIP pricing, but cuts still hurt suppliers like Micron and expose scaling limits. AMD's strategy leans heavier on HBM (which it doesn't fully procure for all racks) and open standards. While AMD has its own DRAM exposure, it hasn't had to walk back a marquee CPU-memory feature like this. Analysts note AMD's memory-heavy designs can shine in scenarios where NVIDIA is forced to compromise on capacity for cost/reliability.

AMD has been winning customers left and right with annual AI accelerator cadence: MI350 (2025) → MI400 Helios rack with the best EPYC Venice for Agentic AI( H2 2026). We may witness violent market share gain for $AMD in 2027-2028 toward 40-60.

In the end, this SOCAMM de-spec on NVIDIA’s Vera Rubin platform hands Dr. Lisa Su and AMD a timely strategic winning in the 2026 & beyong AI accelerator race. While NVIDIA dials back ambitious high-density modular memory to hit cost, reliability, and ramp targets, AMD under Dr. Su’s leadership is delivering on a cleaner, more aggressive roadmap: MI350/MI400 series with superior HBM3E/HBM4 capacity and bandwidth (288–432GB+ per GPU, up to 19.6 TB/s), TSMC 2nm process advantage over Rubin’s 3nm, and full rack-scale Helios solutions that avoid the same last-minute compromises.

This is $NVDA execution friction that creates breathing room. Hyperscalers wary of single-vendor risk and memory-bound workloads now have stronger reason to diversify sockets toward AMD’s more straightforward high-memory designs and improving ROCm/UALink ecosystem. Dr. Su has repeatedly shown the ability to turn competitive gaps into momentum, whether through disciplined annual cadences, customer wins like META, OpenAI, Softbank, Anthropic or Oracle, or pivoting to exploding agentic CPU demand alongside GPUs.

Not Financial Advice! DYOR!

[norman a gabriel @rhoman1996]

$AMD

[Ally @treasureh8nter]

A gentle reminder that in $AMD ‘s last 13F filing, they have $6.5m worth of $MRVL shares 😉

[norman a gabriel @rhoman1996]

@oguzerkan $NVDA market cap is over 5 trillion! $AMD 800 billion ! $NVDA GROWTH priced in ?

[norman a gabriel @rhoman1996]

@LongGameEquity yes $amd reminds me of $nvda when it got started ! earnings for $amd are exploding upward ! let ur winners run !

[norman a gabriel @rhoman1996]

$AMD

[Danny cheng @dannycheng2022]

This post has aged well. When $AMD ran to $300 a couple of months ago and then smashed through $400 just weeks later, I received a flood of messages asking if I was trimming my position.

I stood firm and shared my full bullish thesis with the community: $AMD is headed toward a $1 trillion market cap as my first major target — and explained in detail why I wasn’t selling a single share.

To keep everyone aligned on the core must-hold stocks, I try my best to chart them daily and highlight the most important levels. Any post I mark “MUST-READ” is exactly what I want the community to focus on every single day, week, and month!

[norman a gabriel @rhoman1996]

chasing $NVDA market cap over 5 trillion !! $AMD

[norman a gabriel @rhoman1996]

@dannycheng2022 @chad_ventures $AMD to be clear at $470 you are not taking any profit ? in fact you are still accumulating?? thank you ? thank you for sharing your thoughts !

[norman a gabriel @rhoman1996]

$AMD

[Yiannis Zourmpanos @yianisz]

I still think the market does not fully understand $AMD’s economics. Otherwise this is probably already a $700+ stock.

1/ Agentic AI is changing the CPU-to-GPU ratio from ~1:8 toward potentially 1:1 over time.. $AMD owns both sides of the stack: EPYC CPUs + Instinct GPUs.

2/ And the GPU story itself is misunderstood too.

MI400 is not trying to beat $NVDA on raw FLOPs. The real advantage is 432GB HBM4 memory capacity vs ~288GB on Vera Rubin. For inference, memory density increasingly matters more than peak compute.

3/ Meanwhile ROCm 7 brought major improvements in inference, PyTorch integration, vLLM, SGLang, Hugging Face support, and model portability. Lisa Su literally said workloads that once took months to optimize can now take days.

If the Street fully understood the economics here, I doubt the stock would still be trading at these levels.

[norman a gabriel @rhoman1996]

$AMD market cap 733 billion $NVDA market cap 5.7 trillion lot of room for $AMD STOCK TO APPRECIATE!!

[Oguz Erkan @oguzerkan]

$AMD is approaching its $NVDA moment.

$UBS sees server CPU TAM at $173 billion by 2030.

Assuming a 25/75 allocation between ARM and x86 architectures, with $AMD capturing a 50% revenue share within the x86 segment, we are looking at nearly $65 billion in revenue from server CPUs alone.

Its whole TTM revenue is currently $34 billion.

Long $AMD

[norman a gabriel @rhoman1996]

$AMD $NVDA market cap 5.3 TRILLION , $AMD market cap 7.3 BILLION !!

[Christian Darnton @CMDarnton0]

I’m basically making annual salaries every time AMD moves 5%. Wild.

Still believe this is heading to a $1T market cap, exactly as I predicted.

$AMD to $2000

[norman a gabriel @rhoman1996]

@dannycheng2022 amen !

[norman a gabriel @rhoman1996]

@chad_ventures nothing goes straight up 4 ever !!

[norman a gabriel @rhoman1996]

$AMD consolidating !

[Antonio Linares @alc2022]

A year has passed

$AMD is trading at $440

[norman a gabriel @rhoman1996]

$AMD AGREE 100 %

[The Great Mattsby @matthughes13]

People waiting for lower will only be chasing later.

This market isnt letting anyone in anymore

$AMD

https://t.co/qFjVGsgZff

[Cole @StockOptionCole]

$AMD After 5 days of consolidation around 450s, the 8MA caught up to the low wicks of the last few days bottoms

We're shaping up for the next leg higher to $600+

Don't forget the massive Monthly breakout & Whale flow that was in 600C 3 weeks ago.

Semiconductor face ripper rally isn't over.

$TSM $NVDA $AVGO $INTC $QCOM