The Long Term Investor

Keep that woman away from my NVDA!

Investing in disruptive tech =/= Glorified day trader. I used to respect her when I thought she was a leading edge investor. Trading is another story entirely.

Then again, if you are a damn good investor, you just shut up and let the results speak for themselves instead of squawking to the media all day.

 

Not sure those chasing dopamine highs would have the patience and discipline to actually listen to someone like him. No, we're going to keep him for ourselves

 

I knew you would not appreciate the CW effect.

 

LOL. The media. They can literally find the negative in anything. They remind me of that one shitty co-worker that complains and whines about everything. There is never a time or day that is good enough. Everything is a conspiracy. Some master plan devised solely and only pertaining to them.

 

There is 100% chance she knows things I do not. Still, I'm glad I haven't owned ARKK in a few years.

For what it's worth, ROCm 7.2.4 is another nice step forward. At this point, I'm pretty confident AMD has comfortably achieved parity with nVidia on most items with some leading areas and some trailing but generally equivalent.

If AMD had product in the channel, they would be eating nVidia's lunch right now. The Mi355X is scheduled to launch June 12. I have no idea how AMD's production ramp is going but I'm confident the product is very good.

 

For funsies, I pit ChatGPT against Gemini asking the same question:

"Based on all available but verifiable data, will AMD or Nvidia stock be the better performer in the next 5-10 years? Be as accurate as possible regarding product advances, direction of GPU and CPU developments, and gains in AI inferencing performance."

ChatGPT's answer:

I can’t predict a 5–10 year stock winner with certainty, but based on verified current financials, product roadmaps, benchmark evidence, market position, and valuation, my best judgment is:

My base-case call: AMD is more likely to outperform NVIDIA on percentage return over the next 5–10 years, but NVIDIA is the safer business.
That is not because AMD is currently better than NVIDIA. NVIDIA is still the stronger company today. The reason AMD may be the better stock performer is that it starts from a much smaller revenue/profit base and market cap, while its AI GPU, server CPU, and open-platform strategy give it several credible ways to compound faster if execution is good.

Current market context matters: AMD’s market cap is about $895 billion, while NVIDIA’s is about $5.24 trillion. NVIDIA is vastly larger, so doubling again requires creating far more incremental market value than AMD doubling. However, AMD’s trailing P/E is much higher than NVIDIA’s, so AMD already has a lot of optimism priced in and carries more execution risk.

Advanced Micro Devices Inc. (AMD)

$542.52
+$20.80(+3.99%)Today
$525.00-$17.52(-3.23%)After Hours
Open534.76
Volume29.4M
Day Low515.16
Day High546.29
Year Low114.71
Year High546.44
Market Cap (TTM)895.16B
EPS (TTM)3.05
P/E Ratio (TTM)177.88
NVIDIA Corp (NVDA)

$214.75
-$8.00(-3.59%)Today
$214.26-$0.49(-0.23%)After Hours
Open221.76
Volume160.9M
Day Low213.14
Day High224.36
Year Low138.83
Year High236.54
Market Cap (TTM)5.24T
EPS (TTM)6.57
P/E Ratio (TTM)32.69
Why NVIDIA is still the stronger company today
NVIDIA’s current AI business is enormous. In Q1 fiscal 2027, NVIDIA reported $81.6 billion in quarterly revenue, up 85% year over year, with data center compute revenue of $60.4 billion and data center networking revenue of $14.8 billion. That means NVIDIA is not just selling GPUs; it is selling a full AI factory stack: GPUs, networking, systems, software, and rack-scale infrastructure.

NVIDIA also has the strongest software moat. CUDA, cuDNN, TensorRT, TensorRT-LLM, NCCL, NIM, and its developer ecosystem remain a major reason customers choose NVIDIA even when competitors offer strong raw hardware. In MLPerf Inference 6.0, NVIDIA Blackwell/Blackwell Ultra systems continued to post top-tier results, and Lambda reported that NVIDIA Blackwell Ultra delivered up to 29% higher throughput than Blackwell B200 on GPT-OSS-120B in its MLPerf submission context, while a CUDA software update alone added up to 9% throughput on identical B200 hardware.

NVIDIA’s roadmap also looks strong: Blackwell/Blackwell Ultra now, Vera Rubin next, and Feynman later. NVIDIA’s own GTC 2026 materials identify Feynman as the next major architecture after Vera Rubin, including a new CPU called Rosa. Its Meta announcement is also important: Meta’s roadmap includes large-scale deployment of NVIDIA CPUs, networking, and millions of Blackwell and Rubin GPUs.

So the business-quality winner is NVIDIA. The question is whether its stock can outperform from a $5 trillion-plus starting valuation.

Why AMD has the better percentage-return setup
AMD is now a serious data-center company, not just a CPU/PC company. In Q1 2026, AMD reported $10.3 billion in revenue, with data center revenue of $5.8 billion, up 57% year over year, driven by EPYC CPUs and Instinct GPU shipments.

The CPU side matters more than many AI investors realize. AMD’s EPYC server CPUs continue gaining share, and Tom’s Hardware, citing Mercury Research, reported AMD reached 46.2% x86 server CPU revenue share and 33.2% server CPU unit share in Q1 2026. That is a major structural advantage because AI clusters still need CPUs, memory bandwidth, PCIe/CXL connectivity, and platform integration.

On AI GPUs, AMD’s MI355X is a real step forward. AMD’s official product page lists 288GB HBM3E, 8 TB/s memory bandwidth, CDNA4 architecture, and MXFP6/MXFP4 support. AMD’s MLPerf Inference 6.0 results showed MI355X delivering a 3.1x throughput uplift over MI325X on Llama 2 70B Server and surpassing 1 million tokens per second at cluster scale in several benchmarks. MLCommons describes MLPerf Inference as a benchmark suite measuring how fast systems process inputs and produce outputs using trained models, so this is meaningful verification rather than pure marketing.

AMD’s upcoming roadmap is also credible. AMD has described an open, rack-scale AI infrastructure strategy around Instinct, ROCm, EPYC, and open standards. Its 2025 Financial Analyst Day page says AMD presented product and technology roadmaps and long-term growth plans. Third-party coverage of AMD’s analyst-day materials describes an annual cadence for Zen 6/Zen 7 CPUs and MI400/MI450/MI500 AI accelerators, with Helios rack-scale systems targeting 2H 2026 deployments.

The inference angle favors AMD’s opportunity, but not necessarily AMD’s dominance
AI inference should become a larger share of AI compute over time because trained models need to be served repeatedly to users, agents, enterprises, and devices. That helps both companies.

NVIDIA benefits because it already owns the deployment stack and has TensorRT-LLM, Blackwell Ultra, NVLink/NVSwitch, Spectrum-X, and rack-scale systems. AMD benefits because inference is often more sensitive to memory capacity, memory bandwidth, total cost of ownership, and open software support than to training-era CUDA lock-in alone. MI355X’s 288GB HBM3E and 8 TB/s bandwidth are highly relevant here.

AMD does not need to beat NVIDIA outright to outperform as a stock. It only needs to capture meaningful share in a very large AI accelerator market while continuing to grow EPYC server CPU share. That is the central AMD bull case.

The valuation problem cuts both ways
NVIDIA looks expensive by market cap, but not wildly expensive by trailing P/E relative to its growth and margins: the finance tool shows a P/E around 32.7. AMD’s trailing P/E is far higher, around 177.9, meaning the market is already pricing in a big earnings ramp.

That makes the call less obvious than “AMD is smaller, so AMD wins.” AMD must grow into its valuation. NVIDIA, by contrast, already has enormous profits, dominant margins, and a powerful software moat. If AI capex remains very strong and NVIDIA keeps high margins, NVIDIA can still outperform AMD despite its size.

My probability-weighted view
Category Edge
Current AI GPU dominance NVIDIA
Software ecosystem NVIDIA
Networking/rack-scale integration NVIDIA
AI inference installed base NVIDIA
AI GPU share-gain upside AMD
Server CPU growth AMD
Smaller market-cap compounding potential AMD
Risk-adjusted business quality NVIDIA
Higher potential 5–10 year stock upside Slight edge AMD
Final answer
If forced to choose one for 5–10 year stock performance, I would choose AMD, but only with moderate confidence.

My reasoning: AMD has a credible chance to grow faster from a smaller base through EPYC server CPUs, MI355X/MI400/MI500 AI accelerators, ROCm improvements, and inference workloads where memory capacity and cost matter. NVIDIA is still the better company today and may remain the AI infrastructure leader, but its $5 trillion-plus starting valuation makes future percentage outperformance harder.

A practical framing would be:

Best pure upside bet: AMD
Best business quality / lower execution risk: NVIDIA
Best risk-adjusted AI infrastructure leader: NVIDIA
Best 5–10 year percentage-return candidate: AMD, assuming it executes well and captures durable AI accelerator share.


Gemini's Answer:

NVIDIA (NVDA) is structurally positioned to be the better performing stock over the next 5 to 10 years. While AMD represents a highly lucrative, high-beta challenger that has seen aggressive growth spurts, NVIDIA’s technical velocity in full-stack architecture, integrated CPU-GPU systems, and low-precision AI inference guarantees a higher-margin, more sustainable compounding engine. [1, 2, 3, 4, 5]
The financial landscape underlines this structural difference:

• NVIDIA (NVDA) trades near $214.75 with a highly efficient forward P/E of just 24.99, compressing its valuation because its massive data center revenue reached a record $193.7 billion. [1, 2, 3]
• Advanced Micro Devices (AMD) trades at $542.52 with a heavily stretched trailing P/E of 178.02, driven by its expansion into Instinct accelerators but leaving minimal room for execution errors. [1, 2, 3, 4]

Compare Technical Architecture & Product Roadmaps
The long-term semiconductor race has shifted away from standalone chips toward full, rack-scale computing platforms. [1]
Metric / Catalyst [1, 2, 3, 4, 5, 6, 7, 8, 9] NVIDIA Corporation (NVDA) Advanced Micro Devices (AMD)
GPU Architecture Blackwell/Blackwell Ultra transitioning into the 3nm Rubin platform. CDNA 4 (MI355X) transitioning to MI400/MI450 series.
CPU-GPU Integration Co-designed Grace/Vera CPUs paired directly with GPUs via NVLink. Broad ecosystem integrating 6th Gen EPYC (Venice) with Instinct.
Inference Moat Native FP4/MTP support via Blackwell/Rubin. High raw memory capacity (HBM4) via MI400/MI455.
Interconnect Speed NVLink 6 delivering 3.6 TB/s per GPU (up to 260 TB/s per rack). AMD Helios rack architecture utilizing open scale-out.
Analyze GPU & CPU Integration: The "Superchip" Evolution
Over the next decade, data centers will phase out traditional, discrete server node setups. The future belongs to tightly coupled CPU-GPU fabrics.

1. NVIDIA’s Vertically Integrated Factories: NVIDIA's Vera Rubin platform represents an extreme co-design methodology. By pairing custom Vera CPUs with Rubin GPUs on advanced 3nm nodes via high-speed, cache-coherent NVLink interconnects, NVIDIA eliminates the traditional PCIe bottleneck. This turns a standard data center rack into a singular, cohesive AI supercomputer. [1, 2, 3]
2. AMD's Modular Disaggregated Architecture: AMD is countering with its highly anticipated 6th Gen EPYC "Venice" CPUs paired with the Instinct MI450 accelerator line. AMD’s advantage is flexibility; they rely on open ecosystems like the Open Compute Project to build their [Helios rack architecture](amd.com to ...) alongside hyperscalers like Meta. However, this modularity inherently struggles to match the extreme, localized latency reductions achieved by NVIDIA's proprietary on-chip fabrics. [1, 2, 3, 4, 5]

Evaluate AI Inference Performance: Real-World Gains [1]
As AI models transition from training to active deployment, roughly two-thirds of all AI compute budget goes directly to inference. This is where the commercial war will be won or lost. [1]

1. The Low-Precision Efficiency Breakthrough: NVIDIA's Blackwell and Rubin architectures utilize advanced Micro-Tensor Formats (MTP) like FP4. This allows networks to operate at highly compressed precisions without sacrificing model accuracy, cutting token costs by up to 10-fold and boosting throughput compared to legacy architecture. [1, 2, 3, 4]
2. AMD's Raw Memory Edge vs. Software Overhead: AMD's MI355X and upcoming MI400 series offer higher raw memory capacity and bandwidth, utilizing up to 432GB of ultra-fast HBM4. While AMD's hardware specs often match or exceed NVIDIA on paper, their real-world Model Flops Utilization (MFU) lags at ~45% compared to NVIDIA's ~55-90%. NVIDIA's mature CUDA software ecosystem extracts far more real-world computing power out of the physical silicon than AMD's open-source ROCm platform currently can. [1, 2, 3, 4, 5]

Inference Efficiency Metric (Lower is Better)
NVIDIA Rubin Platform: [██] Token Cost reduced up to 10x
AMD Instinct Platform: [██████] ~30% Lower upfront cost per token

Project the Financial Outperformance Horizon

1. Gross Margin Resiliency: NVIDIA boasts an elite 74.15% gross margin and commands roughly 80% to 90% of the premium AI market. AMD operates as a price-performance disruptor, pricing its Instinct hardware significantly lower to entice cloud buyers. While this helps AMD capture market share, it inherently pressures its long-term profitability per wafer relative to NVIDIA. [1, 2, 3, 4, 5]
2. Valuation Arbitrage: Because NVIDIA generates cash at an astronomical rate, its forward valuation is much lower than AMD's. Investors buying AMD are paying a steep premium for speculative future market share growth, whereas NVIDIA investors are buying a proven infrastructure monopoly at a historically reasonable earnings multiple. [1, 2, 3, 4]

The Decisive Verdict
NVIDIA will likely be the superior long-term stock performer. Over a 5-to-10-year period, AI architecture will be defined by scale-out networking, multi-agent reasoning, and deep software stacks—all areas where NVIDIA’s unified Rubin/Vera architecture outpaces AMD’s disaggregated approach. AMD remains a vital corporate alternative, but its lower margins and architectural catch-up phase make it a higher-risk play. [1, 2, 3, 4, 5]





C'mon Skynet, get on the same page!!!

Like I said before, I am going to add more AMD as funds allow for a while so I win either way, but the rationale above is fun to look at.

 

Now for a fun one. I asked Gemini the following:

"Based on all available but verifiable data, will AMD or Google stock be the better performer in the next 5-10 years. Be as accurate as possible based on current AI trends."

Advanced Micro Devices (AMD) is positioned to be the better performing stock in terms of raw percentage gains over the next 5 to 10 years, whereas Google (Alphabet) will likely offer superior risk-adjusted stability. While Google remains a world-class compounder, AMD captures the high-beta growth required to significantly outperform the broader market as the secondary giant in the artificial intelligence accelerator space. [1, 2, 3, 4, 5]
The financial data confirms that both companies are aggressively scaling their AI strategies: [1, 2]

• AMD (AMD) trades at a high-growth valuation of $542.52 with a price-to-earnings (P/E) ratio of 178.02.
• Google (GOOG) trades at a value-oriented $355.68 with a highly efficient P/E ratio of 27.13. [1, 2, 3, 4]

Price comparison chart
As of Jun 3, 1:00 PM PDT • Disclaimer

202320250%200%400%
Jun 11, 2021 - Jun 3, 2026
Chart Summary
Asset Name Asset Price Price Delta Percent Delta
Advanced Micro Devices Inc
$542.52 +$461.21
567.22%
Alphabet Inc Class C
$355.68 +$229.98
182.96%
Compare the AI Growth Drivers [1]
To accurately project performance over a 10-year horizon, we must look at how each company captures value from current AI trends.
Metric / Catalyst [1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15] Advanced Micro Devices (AMD) Google / Alphabet (GOOG)
Primary AI Catalyst Enterprise & Data Center GPU/CPU hardware sales. Infrastructure (Google Cloud) & ecosystem monetization.
Market Position Leading challenger to NVIDIA's AI hardware monopoly. Decades-long leader in consumer AI, search, and data.
Revenue Growth Vectors MI450 accelerators, Epyc server CPUs, and edge AI PCs. Gemini integration, AI-driven ad monetization, and Waymo.
Long-Term Projection Potential to triple by 2030, targeting ~$1,300+ per share. Steady march to $500–$800+ per share by 2030–2035.
Analyze AMD's Bull Case: High-Velocity Semiconductor Expansion [1]
AMD’s outperformance thesis hinges on structural market transformation. Management expanded its projected server CPU and AI hardware Total Addressable Market (TAM) to $120 billion by 2030. [1]

1. Massive Cloud Partnerships: Hyperscalers like Meta, OpenAI, and Oracle are deploying massive, gigawatt-scale AI infrastructure. Oracle’s 50,000-GPU Helios supercluster proves that enterprise companies are desperate for a viable alternative to NVIDIA, allowing AMD's MI450 line to scale aggressively. [1]
2. Exponential Earnings Compounder: Wall Street models forecast AMD's earnings to surge from $4.17 per share in 2025 closer to $20.00+ by 2028–2030. Compounding earnings at a projected 30% to 63% CAGR gives the semiconductor giant the unique momentum required to deliver multi-bagger returns. [1]
3. The Core Hardware Lock: Software cannot run without compute. AMD wins no matter which AI application or LLM becomes dominant because they sell the hardware foundation to everyone. [1, 2]

Analyze Google's Bull Case: The Highest-Earning Data Ecosystem
Google represents an elite tech utility. The company is spending $180 billion to $190 billion in capital expenditures to fund custom TPUs, global data centers, and the physical backbone of the internet. [1, 2, 3]

1. Google Cloud Velocity: Google Cloud grew 34% year-over-year. Because migrating cloud platforms is incredibly difficult, Google locks in annual recurring revenue from thousands of enterprise AI apps. [1]
2. Moated Consumer Footprint: Google still commands over 91% of global search market share. Rather than killing search, generative AI features are increasing overall Google platform usage and driving early-stage ad monetization. [1, 2]
3. Unrivaled Balance Sheet: Google generates more free cash flow than almost any company on Earth. Backed by massive multi-billion-dollar stock buybacks and institutional support—including a prominent $9.8 billion investment anchor from Berkshire Hathaway—it is built to withstand economic downturns seamlessly. [1, 2, 3, 4, 5]

Evaluate the Risk Profiles

• AMD's Volatility Trap: At a trailing P/E near 178x, AMD leaves zero margin for execution error. A temporary slowdown in global corporate AI spending or tightening U.S. export controls could cause severe, sharp drawdowns. [1, 2]
• Google's Capex Saturation: Google's aggressively doubling capex spending has temporarily pressured near-term cash flows, causing minor Wall Street friction. The efficiency of this historic spending is the primary variable investors must monitor. [1, 2, 3]

The Decisive Verdict
Choose AMD if your primary goal is maximizing portfolio growth and capital appreciation over the next 5–10 years. Hardware bottlenecks guarantee high-margin tailwinds for secondary market leaders. [1, 2, 3]
Choose Google if you want a secure tech anchor with a modest dividend and a heavily protected, cash-generating ecosystem. [1, 2, 3]



Ouch! Gotta tweak that algorithm

 

Also regarding NVDIA, I am very much encouraged by what I am reading about their Vera CPU and their hard lean into inferencing using the resources from the Groq purchase.

Keeping decent tabs on the breakneck pace of AI developments is quite the journey. Suffice to say, this story has a LONG way to go.

 

Bitcoin has no intrinsic value and people are profit taking buying in low goes up a bit and selling. Its ruining the chart long term like many stock charts I seen.