TSMC and Intel Exploring Diverging Technology For Future Compute Capacity

[worldstarcountry]

This is a very in depth video on micro chip production, the half billion dollar warehouse sized machines that often are needed to manufacture them, and the current walls they have hit with shrinking down the scale. The gist is that despite shrinking down to 2nM , they were only able to add 6% more transistors. This is clearly a problem, so TSMC is approaching moving to what is being termed , MEGA chips. They  going to begin stacking and integrating multiple chips for compounded compute power as well as etching the tiny something in multiple passes. Very technical but very well explained. First half covers TSMC, the second half covers Intel's breakthroughs and path going forward.

Intel is taking a slightly different route which is also explained in the video. When I see how these machines work, it is pretty crazy. Light being aimed at silicone and then just summoning patterns that are integrated with tiny wires .... feels like the power of God. He aimed light at dirt and carbon and suddenly we were summoned up into existence.

For those interested in such technological type topics and where the near future is taking these processes.

[worldstarcountry]

Posted premature, but Intel's future seems to be focused on changing to LIGHT for communication, or what we would know as fiber optics. There was a video on this as well in which the theory and research is switching away from copper wires in computing towards fiber optical communication between compute communication (chips to boards and all that).

This video was from last month.



Invest accordingly 

[andy06shake]

The main concern with silicon is at very small scales(10nm & below).

Quantum tunneling becomes significant.

Electrons leak through barriers even when they shouldn't.

Which causes errors to occur and power loss...

It's an engineering/material science problem that becomes exponentially worse as the size shrinks.

[quintessentone]

Posted premature, but Intel's future seems to be focused on changing to LIGHT for communication, or what we would know as fiber optics. There was a video on this as well in which the theory and research is switching away from copper wires in computing towards fiber optical communication between compute communication (chips to boards and all that).

This video was from last month.
[Video: https://youtu.be/dytkmA61Fmg]

Invest accordingly 

I will invest accordingly but only after I, if possible, come to understand it all and come to understand exactly where or whom to invest in.

----

"NVIDIA has advanced light-based technologies across three key areas, though specific "new" public statements from NVIDIA physicists are most prominently tied to its physics-aware AI systems and optical interconnect investments. 

• Physics-Aware AI (Cosmos): CEO Jensen Huang highlighted the importance of physics in technology, noting that NVIDIA’s Cosmos platform uses world foundation models to simulate real-world physics for robotics and autonomous driving.  This shift emphasizes that future AI innovation relies on physics and materials science rather than just software. 
  
• Optical Interconnects for AI Data Centers: NVIDIA is heavily investing in silicon photonics and co-packaged optics (CPO) to move data at the speed of light.  This includes a $3.2 billion investment in Corning to build optical fiber factories and a $4 billion investment in Lumentum and Coherent to develop state-of-the-art optics technology, aiming to replace copper cables with optical connections for greater bandwidth and energy efficiency. 
  
• Path Tracing and Rendering: In graphics, NVIDIA is pushing real-time path tracing and RTX Mega Geometry to achieve photorealistic lighting.  While this is a rendering technology rather than a fundamental physics discovery, it represents a significant evolution in how light is calculated and simulated in real-time environments, moving beyond traditional ray tracing to a unified physics-based equation for lighting. 
  

While NVIDIA partners with research institutions like Northwestern University to accelerate quantum research using NVIDIA technology, and companies like Sandia National Laboratories use NVIDIA GPUs for AI-driven LED light control, the primary "new light tech" narrative from NVIDIA centers on optical data transmission and physics-simulated AI." (LLM)
-----

"Amid the AI Craze, NVIDIA CEO Says He’d Choose Physics Over Software?
 
NVIDIA CEO Jensen Huang says the future of AI isn't just code, it's physics. Find out which skills will truly matter for the next tech revolution."

"The leaders of tomorrow will be those who can bridge the digital and physical realms."

"The goal is not just to create a smarter internet, but to build a smarter, more responsive physical world. To get there, the next wave of pioneers will need a toolkit built on the fundamental laws of reality. Understanding the universe might just be the most important skill in an age dominated by artificial intelligence."
 
https://www.rekadia.co.id/news/amid-the-...r-software

[ArMaP]

The main concern with silicon is at very small scales(10nm & below).

10 nm is just around 100 silicon atoms, so there's not much room to work.

[quintessentone]

Should I be investing in the graphene market?

"Current research indicates that physicists and material scientists are exploring alternatives to silicon, such as two-dimensional transition-metal dichalcogenides (TMDs) and graphene, to create faster and more efficient chips by harnessing atomic defects and higher electron mobility.  However, NVIDIA itself does not manufacture chips or conduct this fundamental materials research; instead, it designs GPUs that are manufactured by TSMC using current silicon-based processes, with TSMC and equipment suppliers like ASML handling the physical production.  While NVIDIA focuses on software ecosystems like CUDA and AI models, the industry is looking beyond silicon due to silicon's diminishing electrical performance and thermal limitations, but no commercial replacement has yet fully superseded silicon in mass production." (LLM)

---

"The graphene stock market is characterized by high volatility and a transition from experimental R&D to early-stage commercialization, with a projected compound annual growth rate (CAGR) exceeding 30% through the early 2030s.  Investors typically access this sector through pure-play producers like NanoXplore (GRA/TSX-V) and Graphene Manufacturing Group (GMG/TSX-V), or equipment suppliers such as CVD Equipment (CVV/NASDAQ)." (LLM)

[andy06shake]

10 nm is just around 100 silicon atoms, so there's not much room to work.

Indeed.

But they are still operating at a scale where individual atomic effects become a thing...

[andy06shake]

Should I be investing in the graphene market?

"Current research indicates that physicists and material scientists are exploring alternatives to silicon, such as two-dimensional transition-metal dichalcogenides (TMDs) and graphene, to create faster and more efficient chips by harnessing atomic defects and higher electron mobility.  However, NVIDIA itself does not manufacture chips or conduct this fundamental materials research; instead, it designs GPUs that are manufactured by TSMC using current silicon-based processes, with TSMC and equipment suppliers like ASML handling the physical production.  While NVIDIA focuses on software ecosystems like CUDA and AI models, the industry is looking beyond silicon due to silicon's diminishing electrical performance and thermal limitations, but no commercial replacement has yet fully superseded silicon in mass production." (LLM)

---

"The graphene stock market is characterized by high volatility and a transition from experimental R&D to early-stage commercialization, with a projected compound annual growth rate (CAGR) exceeding 30% through the early 2030s.  Investors typically access this sector through pure-play producers like NanoXplore (GRA/TSX-V) and Graphene Manufacturing Group (GMG/TSX-V), or equipment suppliers such as CVD Equipment (CVV/NASDAQ)." (LLM)

Graphene runs into many of the same problems silicone does, just in different ways.

[quintessentone]

Graphene runs into many of the same problems silicone does, just in different ways.

I am just trying to get a handle on all of this and to know where to invest wisely. I digress: This sandwiching/layers of metals reminds me of the metal supposedly retrieved from an alien craft, where it too was layered, but that has seemingly been debunked.

https://www.metabunk.org/threads/aaro-re...ial.13548/

This is way too complicated to figure out for investment strategies. Critical minerals is it? Tungsten? Titanium? Graphite is needed here as well. Well at least I have a starting point.

https://www.americanelements.com/periodi...rence.html

-----

"Two-dimensional transition metal dichalcogenides (2D TMDs) are atomically thin semiconductors with the general formula MX₂, where M is a transition metal (such as Mo, W, Ti, or V) and X is a chalcogen atom (such as S, Se, or Te). 

• Structure: They consist of a single layer of metal atoms sandwiched between two layers of chalcogen atoms (X-M-X), held together by weak van der Waals forces that allow them to be exfoliated into monolayers. 
  
• Key Properties: Unlike their bulk forms which often have indirect bandgaps, monolayer TMDs exhibit a direct bandgap, strong spin-orbit coupling, and high optical absorption (up to 20% in a single layer). 
  
• Common Examples: The most studied materials include MoS₂ (molybdenum disulfide), WS₂ (tungsten disulfide), MoSe₂, and WSe₂. 
  
• Applications: Due to their tunable electronic properties and mechanical flexibility, they are used in nanoelectronics (transistors), optoelectronics (LEDs, photodetectors, solar cells), energy storage, and biomedical applications (biosensing, drug delivery)." (LLM)
  

[andy06shake]

I am just trying to get a handle on all of this and to know where to invest wisely. I digress: This sandwiching/layers of metals reminds me of the metal supposedly retrieved from an alien craft, where it too was layered, but that has seemingly been debunked.

https://www.metabunk.org/threads/aaro-re...ial.13548/

This is way too complicated to figure out for investment strategies. Critical minerals is it? Tungsten? Titanium? Graphite is needed here as well. Well at least I have a starting point.

https://www.americanelements.com/periodi...rence.html

-----

"Two-dimensional transition metal dichalcogenides (2D TMDs) are atomically thin semiconductors with the general formula MX₂, where M is a transition metal (such as Mo, W, Ti, or V) and X is a chalcogen atom (such as S, Se, or Te). 

• Structure: They consist of a single layer of metal atoms sandwiched between two layers of chalcogen atoms (X-M-X), held together by weak van der Waals forces that allow them to be exfoliated into monolayers. 
  
• Key Properties: Unlike their bulk forms which often have indirect bandgaps, monolayer TMDs exhibit a direct bandgap, strong spin-orbit coupling, and high optical absorption (up to 20% in a single layer). 
  
• Common Examples: The most studied materials include MoS₂ (molybdenum disulfide), WS₂ (tungsten disulfide), MoSe₂, and WSe₂. 
  
• Applications: Due to their tunable electronic properties and mechanical flexibility, they are used in nanoelectronics (transistors), optoelectronics (LEDs, photodetectors, solar cells), energy storage, and biomedical applications (biosensing, drug delivery)." (LLM)
  

Investment-wise i would not care to speculate, Quint.

Because it's just not something i have even done or know much about.

It's certainly a real and promising material...

And not just where the likes of possible IC replacement is concerned.