Accidental Innovations

 

Introduction

In 2025 some circles are debating on whether Artificial Intelligence has integrated and saturated industry (Kennedy, 2023).  Technology is making its presence known via algorithms showing up in,

ü  Virtual platforms, that use sentiment analysis and transcribe meetings such as Read.ai.

ü  Financial intelligence systems such as predictive algorithms and market research like Fortrade or Cleo ai which manages personal finance.

ü  Virtual assistants, help with scheduling, research, reminders etc… like Google Assistant or ClickUp.

ü  Chatbots, to help with general research, writing and idea generation such as ChatGPT or Grok.

ü  Health artificial intelligence tools to gather biometric data and make recommendations on foods, workouts, mindfulness sessions, coaching sessions and sleep patterns such as Calm or the Fitbit Ai or Apple’s FitnessAI.

ü  Image Generators that take user prompts and create images such as Canva, Dall-E3, Lensa and Adobe Firefly (Glover, 2025).

ü  Education intelligence systems that utilize real time data to personalize user learning experiences ie., Duolingo (language tool) Adapted Mind Math (Mathematics tool) or Grammarly (writing tool).

ü  Integrated AI are found in everyday applications such as Zia, the Zoho AI assistant, Microsoft’s Microsoft Copilot, Gemini, Google apps assistant and so much more.  

AI has touched most industries and researchers and technologists are seeking to ensure that machine learning and artificial algorithms are everywhere. This is not ambitious as it is already happening. Almost every month there is a new AI being announced (Mauran, 2025). So, “are we witnessing the zenith of AI’s golden age or are we on the precipice of a market saturated beyond capacity?  The tech landscape has always been dynamic, with innovations often outpacing the market’s ability to adapt” (Kennedy, 2023).

With all the benefits of AI, humanity has yet to explore potential innovations. A major part of influencing the scalability of AI is ensuring that the infrastructure is in place to allow growth to happen.  Immersion IQ notes that “the infrastructure supporting these workloads has evolved from a nice-to-have to a strategic differentiator” (2025).

A core feature of AI infrastructure is memory.

Immersion IQ explains that “memory bandwidth, storage throughput and network connectivity are often more critical to overall performance than raw compute power” (2025).

“We are tasking our computers with processing ever-increasing amounts of data to speed up drug discovery, improve weather and climate predictions, train artificial intelligence, and much more. To keep up with this demand, we need faster, more energy-efficient computer memory than ever before” (The Standford Report, 2024).

PCM Technology: Indium Selenide

Universal Memory & The Accidental Innovation using Indium Selenide

In December 2024 LiveScience magazine reports – “Accidental discovery creates candidate for universal memory – a weird semiconductor that consumes a billion times less power (Hughes, 2024). In the race to achieve efficient memory utilization in AI systems, Universal Memory is described as “computing memory that can replace both short-term memory like random access memory (RAM) and storage devices like solid-state drives (SSDs) or hard drives” (Hughes, 2024). 

Universal Memory can be represented by a traditionally expensive technology called the Phase-change memory (PCM) that engages a process using high levels of power- called melt quenching; specialized materials to store data, and provision of faster processing (Hughes, 2024, Orf, 2024 Shekhar, 2024, Godse, 2018).   “PCM works by switching materials between two states: crystalline, where atoms are neatly organized, and amorphous, where atoms are arranged randomly. This encodes values 1 and 0 by changing the state of the matter. However, the technology used for this involves heating and rapidly cooling materials and requires a lot of energy” (Rusano, 2024).

In 2024 scientists accidentally discovered data -storage benefits while using a semiconductor material called indium selenide (In2Se3) which uses up to one billion times less energy to bypass melt quenching. Indium selenide uses ferroelectric materials (which provide spontaneous polarization) and piezoelectric materials (these physically deform) to deliver the lowered energy requirements while increasing the data storage of capabilities for PCM (Hughes, 2024).  

Scientists explain that instead of engaging in the melt-quenching process, which is common to PCM Technology, Indium Selenide uses a steady electrical current to produce mechanical shocks which convert it from a crystalline state to a glass phase using less power (SciTechDaily 2024, Orf, 2024). This revolutionary process represents a major contribution towards Universal Memory because it reduces the energy requirements and increases data storage capabilities.

 Supporting forces

The forces supporting this new technology are technological.  Because of the efficiency-based advantages of using Indium Selenide as a PCM technology acceptance and use by scientists, technologists and engineers would be the primary stimuli which would ensure integration into cell phones, computers (SciTechDaily, 2024) and AI applications in the current and near future.

Conclusion

Phase-change memory technology is now one of the primary solutions towards Universal Memory, use of Indium Selenide to reduce the cost (financial and energy) by bypassing the melt quenching process with the added benefit of data storage makes it a forerunner in the race to create more storage, increasing processing and lowering energy requirements for AI (Texas Materials Institute, 2025, Godse, 2018). 

While it may seem that AI has saturated industry, there are advantages to using technology i.e. Health diagnosis analytical systems or financial systems, which all require sound infrastructure, such as memory.   Universal memory enables a wider array of responses, productivity, predictive capability, adaptation and accuracy in machine learning platforms (Sandil, 2025).  Innovations using PCM Technology provide efficiency, scalability and reduced energy consumption when using materials such as Indium Selenide.

REFERENCES

Glover, E. (2025, June 3). 42 top AI apps to know. Built In. https://builtin.com/artificial-intelligence/ai-apps

Godse, R., McPadden, A., Patel, V., & Yoon, J. (2018, November). Memory technology enabling the next artificial intelligence revolution. In 2018 IEEE Nanotechnology Symposium (ANTS) (pp. 1-4). IEEE.

Hughes, O. (2024, December 4). “Accidental discovery” creates candidate for universal memory — a weird semiconductor that consumes a billion times. Live Science. https://www.livescience.com/technology/computing/accidental-discovery-creates-candidate-for-universal-memory-a-weird-semiconductor-that-consumes-a-billion-times-less-power

Kennedy, V. (2023, September 30). AI tech boom: Is the artificial intelligence market already saturated? Cointelegraph. https://cointelegraph.com/news/ai-market-saturated-investment/

Immersion IQ. (2025, June 10). Understanding AI infrastructure requirements .https://immersioniq.io/understanding-ai-infrastructure-requirements/

Mauran, C. (2025, June 29). The biggest AI announcements and high drama of 2025 (so far). Mashable. https://mashable.com/article/biggest-ai-announcements-2025

 Orf, D. (2024, December 6). A seismic new semiconductor could lead to the holy grail of data storage. Popular Mechanics. https://www.popularmechanics.com/technology/a63095038/semiconductor-phase-change-memory/

Rusanov, A. (2024, December 6). Accidental discovery: Scientists reduce power consumption of PCM memory by 1 billion times. ITC.ua. https://itc.ua/en/technologies/accidental-discovery-scientists-reduce-power-consumption-of-pcm-memory-by-1-billion-times/

Sandil, R. (2025, January 29). The powerful role of memory in AI - Rahul Sandil - Medium. Medium. https://medium.com/@rahulsandil/the-powerful-role-of-memory-in-ai-8dd59662fe29

SciTechDaily. (2024, November 4) Shocking New Memory Tech: crystal-to-glass transformation using a billion times less energy. Indian Institute of Science (IISE). https://scitechdaily.com/shocking-new-memory-tech-crystal-to-glass-transformation-using-a-billion-times-less-energy/

Shekhar, S., Bogaerts, W., Chrostowski, L., Bowers, J., Hochbery, M., Soref, R., Shastri, B.  (2024) Road mapping the next generation of silicon photonics. Nat Commun 15, 751. https://doi.org/10.1038/s41467-024-44750-0

Texas Materials Institute. (2025, January 6). IN2SE3: a new material for enhancing AI performance. https://tmi.utexas.edu/news-events/364-in2se3-a-new-material-for-enhancing-ai-performance-led-by-dr-yurim-jeon-and-dr-deji-akinwande

The Standford Report. (2024, January 22.).  Closing in on universal memory for large data processing Stanford University. https://news.stanford.edu/stories/2024/01/closing-universal-memory-large-data-processing