THE NEXT "GAME CHANGER" ... light on a computer chip.
- Greenmark 101
- 12 hours ago
- 9 min read
Updated: 20 minutes ago
Presented by Christopher Netelkos, co-author of the Subjective Probability Model algorithm known as G-101 SPM AI.
Using G-101 SPM AI to obtain a reliable answer to any question is possible by applying predictive analytics, which uses data and machine learning to forecast future trends and behaviors, and scenario planning, which involves creating potential future scenarios to inform decision-making. Its method called "Smart data analysis" is more reliable than traditional or qualitative forecasting, especially when dealing with complex systems.Â
Question: When will AI equal the capacity of the human brain?
Answer: AI may achieve near human-level-cognition around 2027 to 203o, providing such replication of the human brain can overcame significant technological hurdles, including the need for massive accounts of computational power and energy that current AI systems use. Moreover, progress in computer chip design and hardware, especially in  photonic integrated circuitry (PIC) and a better understanding of the human brain's neural connections, often called the "neural code."
How are we able to even ask such a question?
TEN GREAT INVENTIONS from the last 200 years include the telephone, which revolutionized long-distance communication; the automobile, which transformed transportation; the airplane, which enabled fast travel; and the computer, which became the foundation for the digital age. Other groundbreaking inventions are penicillin, which began modern medicine; electricity, which powers countless devices; the internet, which connects the world; and television, which changed entertainment and news dissemination. Innovations like the smartphone have further integrated computing and communication into daily life, while the lightbulb made it possible to have light on demand.Â
Of them all, the greatest idea is the computer:
As the foundation of digital age, the computer allowed our collective thinking to connect disparate pieces of information to form insights, which are them refined through a practical, iterative process of prototyping and testing. Known as associative thinking, the brain forms new ideas by connecting existing concepts in novel ways, drawing links between unrelated items, much like how information is stored and retrieved in memory. Within this framework, the G-101 SPM AI algorithm noticed patters, problems and inefficiencies that spark additional idea. At the heart of this evolving mental process is the computer chips, which communicate by sending and receiving electrical signals to control complex communication protocols. The more complex the computer chip the heater it gets due to the constant movement of electricity. The function makes the chip less reliable as high temperatures interfere with their electrical properties and can cause physical damage. As components get smaller and faster, the heat problem becomes more pronounced. Example: AI accelerators and high-performance computing systems generate massive amounts of data that need to move between processors in real time. Electrical interconnects consume significant power and generate heat, limiting how fast data can flow.
OVERVIEW: A computer chip is packed with billions of tiny transistors, each acting as a switch that turns on or off to represent the 1s and 0s of binary code. Power is consumed whenever a transistor switches state, and this process is not 100% efficient.
Electrical resistance: Even when a transistor is fully "on," it still has some electrical resistance. As electrons move through the chip's intricate network of wires, they collide with silicon atoms and release kinetic energy as heat. This is known as Joule heating.
Switching power: Each time a transistor switches states (from on to off and vice-versa), it causes a tiny capacitor to charge and discharge. This process releases energy as heat and is a significant source of power consumption in a CPU.
Leakage current: Transistors are not perfect switches. A small amount of current, known as "leakage current," can escape even when they are in the "off" state. This leakage increases with temperature, which creates a feedback loop of more heat and more leakage.
Heat reduces reliability: Overheating causes problems for computer chips in several ways, from temporary performance issues to permanent hardware damage.
Performance degradation (Thermal throttling): A chip's most immediate self-preservation mechanism is thermal throttling. When internal sensors detect that the chip is getting too hot, the system automatically reduces its clock speed to lower power consumption and generate less heat. This is why performance drops dramatically during intense tasks if a computer's cooling system is inadequate.
Increased electrical resistance: For many of the materials in a chip, increased temperature leads to higher electrical resistance. This can cause timing issues, where electrical signals don't arrive on time to perform a task before the next clock cycle begins. This leads to computational errors and system crashes.
Signal-to-noise ratio: Higher temperatures increase electrical noise within the chip, which can interfere with the signals that represent data. This can cause bit errors, where a 1 is mistakenly read as a 0, leading to commands being misinterpreted and programs crashing.
Accelerated aging and permanent damage: The physical components of a chip are not immune to heat. i.e. (1) Accelerated wear out: Constant heat exposure accelerates the aging of materials, such as the gate oxide in transistors, which can lead to permanent failure. According to one rule of thumb, every 10°C increase in temperature can cut a chip's lifespan by 50%. (2) Thermal stress: Repeated heating and cooling cycles cause the different materials in a chip (like silicon and the metal interconnects) to expand and contract at different rates. This puts stress on the chip's internal components and can cause microscopic cracks or lead to the separation of different material layers (delamination), resulting in a permanent failure. (3) Transistor failure: In severe cases of overheating, the thermal energy can become high enough to cause electrons to jump the energy bandgap of the semiconductor. This can cause transistors to stop functioning as reliable switches, leading to the chip's failure.Â
THE NEXT "GAME CHANGER" ... light on a computer chip.
FACT: Computer chips can communicate with light, a technology known as silicon photonics. Researchers have developed integrated circuits that use light pulses, or photons, to transfer data instead of relying solely on electrical signals carried by copper wires. This approach can enable higher speeds and lower power consumption, which is especially important for high-performance computing tasks like artificial intelligence.Â
Light-based communication on chips works
Engineers integrate optical components directly onto or within a silicon chip to create a photonic integrated circuit (PIC). For the process to work, electrical data is first converted to optical signals, which are then transmitted across the chip and converted back to electrical signals. The core components are:Â
Modulator: This device encodes electrical data onto a light signal by varying the light's properties, such as its phase or intensity.
Waveguides: These are the "wires" of the photonic circuit, which are tiny channels built from silicon to guide the light across the chip.
Light source: Lasers, often made from materials like indium phosphide, generate the light. These are not yet reliably made from silicon and are typically integrated into the chip using advanced packaging techniques.
Photodetector: At the receiving end, this device absorbs the photons and converts the optical data back into an electrical signal.
Advantages over electrical interconnects.
Traditional electronic circuits that use copper wires face communication bottlenecks as chips become more powerful and densely packed. Moving data with light offers several key advantages:Â
Higher bandwidth: Photons can transmit data at much higher rates than electrons. This is due to their speed and the ability to send multiple signals at once using different wavelengths (colors) of light through a technique called Wavelength Division Multiplexing (WDM).
Lower power consumption: Moving data is power-intensive, but optical interconnects consume far less energy than electrical wires, where a lot of power is lost to resistance. This is a major benefit for large, AI-driven data centers.
Reduced heat generation: Less power consumption means less heat is produced, which makes cooling large systems easier and more cost-effective.
Immunity to interference: Unlike electrical signals, photons are not affected by electromagnetic interference, which ensures cleaner and more reliable data transmission.Â
Current state and future outlook
While silicon photonics is still a developing field, it has already advanced significantly:
Specialized processors: Development of AI-specific processors that perform computations using light, offering greater efficiency for certain workloads like matrix multiplication.
Hybrid systems: Most current technology is "optoelectronic," combining optical interconnects for high-speed communication with conventional electronics for computation.
Practical applications: Light-based communication is already common in long-distance fiber optic cables and is increasingly being integrated into data centers for chip-to-chip and rack-to-rack communication. Intel, for instance, has demonstrated an integrated optical I/O chiplet for next-generation AI infrastructure.Â
Frontrunner in  silicon photonics.
POET Technologies Inc.* (NASDAQ: POET) together with its subsidiaries, designs, develops, manufactures, and sells semiconductor products and services for commercial applications in the United States, Canada, Singapore, and China. The company offers photonic integrated packaging solutions based on the POET Optical Interposer, a novel platform that allows the seamless integration of electronic and photonic devices onto a single chip using advanced wafer-level semiconductor manufacturing techniques. It also designs and develops photonic integrated circuits and optical engines based on the POET Optical Interposer platform. The company serves the data center; tele-communications; 5G interconnect markets, such as PON and GPON, and edge computing for machine-to-machine communications; and sensing markets, including LIDAR, Optical Coherence Tomography for medical devices, and virtual reality systems, as well as light source markets. POET Technologies Inc. was formerly known as Opel Technologies Inc. and changed its name to POET Technologies Inc. in June 2013. POET Technologies Inc. was incorporated in 1972 and is headquartered in Toronto, Canada. * Holding a long position.
Poet's core technology is the Optical Interposer, a platform that integrates electronic and photonic components onto a single chip using wafer-level manufacturing. The company designs optical engines for 800G and 1.6T data transmission speeds -- the backbone of next-generation AI clusters and hyperscale data centers.
The technology addresses the "heat" problem. AI accelerators and high-performance computing systems generate massive amounts of data that need to move between processors in real time. Electrical interconnects consume significant power and generate heat, limiting how fast data can flow. Optical interconnects use light to transmit information, offering higher bandwidth, lower power consumption, and less heat generation.
Poet has secured partnerships with major players, validating its approach.
In May 2024, Foxconn Interconnect Technology selected Poet's optical engines for its 800G and 1.6T optical transceiver modules.
On Sept. 30, 2025, Poet and Semtech launched 1.6T optical receivers for AI networks. The same week, Poet announced a collaboration with Sivers Semiconductors on external light sources for co-packaged optics targeting the AI market.
Using light instead of electricity to power computer chips—known as photonic or optical computing—means data is processed and transferred using photons, or light particles, instead of electrons. This approach offers significant performance and energy efficiency improvements over traditional electronic chips, which are approaching fundamental physical limits.
How optical computing works
In electronic chips, transistors switch electrical currents on and off to represent the binary "1s" and "0s" of digital data. In a photonic chip, this information is encoded in beams of light, which are manipulated by optical components.
Light manipulation: Instead of electrical circuits, tiny waveguides, lenses, and interferometers are etched onto the chip. They guide and modulate light signals to perform mathematical computations and logical operations.
Data processing: Light can perform complex calculations, such as the matrix multiplications crucial for artificial intelligence and machine learning, by interfering with itself. By splitting light beams and recombining them, the resulting interference patterns reveal the outcome of the calculation.
Hybrid systems: Many current research projects are developing "opto-electronic hybrid" chips. These chips use light for high-speed data transfer between components, while still relying on electronics for memory and certain logic functions. This allows for immediate performance gains while full optical computing matures.
Advantages of photonic computing
Higher speed: Light can travel much faster than electrons, allowing data to be transmitted at near light speed. This dramatically reduces latency and increases overall processing speed.
Lower energy consumption: Electrons lose energy as heat as they move through the copper wires of traditional chips. Photons, however, generate far less heat and can travel long distances with minimal signal loss. This lowers the need for intensive cooling systems, making the entire system more energy efficient.
Increased bandwidth and parallelism: Optical signals have a much higher bandwidth, meaning more data can be transferred at once. Multiple light beams of different wavelengths can also pass through the same channel without interfering with each other, allowing for massive parallel processing.
Improved security: Intercepting optical signals is significantly more difficult than intercepting electronic ones, as the act of eavesdropping can be detected by changes in the light signal. This adds a new layer of security to data transmission.
Challenges for future development
While promising, photonic computing is still in the early stages of development and faces several challenges.
Integration with electronics: The biggest hurdle is integrating optical components with existing electronic systems. Achieving seamless compatibility is complex and can be costly.
Memory and storage: Today's electronic memory is not easily replicated with light. Photonic systems currently have to convert light signals back to electronic ones to store data, which can slow down performance.
Manufacturing complexity: Creating and aligning the precise optical components on a photonic chip is a complex and demanding manufacturing process.
Accuracy issues: The analog nature of optical signals means they can be more susceptible to noise and errors, especially in complex, large-scale systems.Â
TRANSACTIONS AND NOTICES
As stated, the $20.00 projected EXIT for POET is a Tier One value. With a Legacy Holding status distinction, POET has a long term target of $80.00 per share; and notwithstanding as a primary T/O target from this day forward
Edge
G101SPM 8/29/2025, 7:51:09 AM $POET $5.64 ask. BUY/ADD TO LONG POSITION. DAC (dollar average cost includes original BUY at $3.35 (9.23.24) (2) $4.495. CHANGE EXIT to $20.00 (long term) from $10.00
$POET $6.38 bid. DAC (dollar average cost) $3.35 (9.23.24). EXIT $10.00 from $6.00. BRIEF: Completed its previously-announced acquisition of control of Super Photonics Integrated Circuit Xiamen Co., Ltd. (company jointly held by, and previously operated as a joint venture between, the Company and Quanzhou San'an Optical Communication Technology Co., Ltd. ("SAIC" or "Sanan").
$POET $3.35 ask. BUY/NEW POSITION carries SPM 85.76 tag to $6.00 in midterm.
Allow G-101 SPM AI algorithm to be your financial guide.