2025 is the year of artificial intelligence (AI) as the development of increasingly powerful models has caused many to lose their jobs and caused the stock values of the world's leading technology companies to rise like a rocket. Here are the 10 most interesting AI news stories of the year.
1. The Global Mega Data Center Boom
Hundreds of millions have been spent by Meta to lure employees from other artificial intelligence (AI) companies to join the Meta Superintellgence Labs (MSL) team. Mark Zuckerberg has assembled an "Avengers" team consisting of the people who created ChatGPT and several powerful GPT models in OpenAI. Zuckerberg also announced the construction of the Hyperion data center, which is not only super powerful but also the size of Galactus.
In a post uploaded to Facebook, Zuckerberg said that Meta is building the Hyeprion data center, which when completed will have a total power of 5GW. Before Hyperion, the Prometheus data center will be operational in 2026 with a power of 1GW. These are just two of several supercluster data centers that Meta will build, each of which will be nearly as large as Manhattan Island in New York. Hyperion will be built in Louisiana with a nearby campus that will have 4 million square feet of floor space.
OpenAI announced the Stargate Project in January with the aim of keeping the United States at the forefront of AI technology that is currently being sought after by the tech industry. Sam Altman, CEO of OpenAI, said the company will invest $500 billion (RM 2.23 trillion) in the Stargate Project over four years.
The project is being carried out by OpenAI with SoftBank, Oracle, and MGX. Softbank CEO Masayoshi Son was appointed chairman of the Stargate Project. Meanwhile, Arm, Microsoft, NVIDIA, and Oracle are key technology partners.
Outside the United States, a mega AI data center is also being built by Reliance in Jamnagar, India, which will come with a total power consumption of 3.0 gigawatts. Saudi Arabia is also looking to build an eco-friendly AI data center that uses power generated by solar and wind. All these mega data center projects can be built due to huge investments and the United States allowing unlimited chip exports after President Trump lifted the restrictions imposed by President Biden.
2. Johor Becomes AI Hotbed But Delays Data Center Construction Because of AI
Johor will soon become Malaysia's AI Hotbed after approving the construction of 42 new data centers earlier this year. Of these, several projects have already begun operations. The first data center with NVIDIA B200 chips has also been confirmed to be built by YTL in Johor.
But the biggest news regarding data centers in Malaysia is that Johor has asked investors who want to build data centers that use water-based cooling systems to delay them for at least 18 months until mid-2027.
The decision was made following concerns about the security of water supply in the state. Johor is currently facing a shortage of water resources caused by drought, pollution, and a sudden increase in demand from the industrial sector.
Then last month, the Johor government would no longer approve applications for the construction of Level 1 and Level 2 data centers that use up to 50 million liters of water per day (MLD). This is a high water consumption equivalent to the needs of 25 million individuals per day. From now on, only Level 3 and Level 4 data centers that use up to 0.2 MLD will be approved.
Although the overall water supply is sufficient, the state faces problems of management and distribution imbalance. Water resources are concentrated in the central and eastern areas of Johor but demand is high in the southern part.
3.AI Causes RAM and Storage Prices to Soar
Mrjn Photograph/ Unsplash
The demand to build dozens of AI data centers with millions of AI chips is causing pressure on the supply of NAND chips needed for RAM and SSD storage. This year has seen the selling price of RAM and SSD increase by up to 50% with the price of smart devices set to become more expensive with the confirmation of Dell, Acer, Asus, Xiaomi, Samsung and many more.
As a result, it is likely that microSD card support will return to Android smartphones. Meanwhile, on smartphones, devices with a base memory of 4GB RAM will also return to normal while on PCs, 8GB will become normal. NVIDIA may also reduce the production of gaming graphics cards by up to 40% next year. All of this is due to the handling of data centers, especially for training AI models.
So far, Micron has discontinued the Crucial brand that offers RAM to the consumer market because it focuses on commercial customers. If you have been saying that AI has no impact on your life, next year you will realize that you will no longer be able to escape the negative effects of the AI war.
4. Meta Fishes Out Competitor AI Employees With Hundreds of Dollars of Rewards
The race to become the world's largest AI company has heated up this year with Meta fishing out employees from competing companies such as OpenAI, Google, Apple and Perplexity before establishing Meta Superintelligence Labs (MSL). Among those successfully fished out by MSL is Shengjia Zhao who is the co-creator of ChatGPT at OpenAI.
Also successfully fished out from OpenAI are the creators of the o3/o4-mini, GPT-4o, GPT-4.1, GPT-4.5, 4o-imagegen, GPT-4o voice mode and o4-mini models. The heads of development of Gemini, Waymo and Deepmind were also successfully lured away from Google by MSL. All of them were offered salary bonuses and shares worth hundreds of millions were also used as bonuses by Meta to drive the artificial general intelligence (AGI) initiative at MSL.
All of this happened in the middle of this year. Then, after a few months, MSL was reshuffled by Mark Zuckerberg, causing newly hired employees to resign, return to their old companies or be fired. MSL was allegedly without a clear direction after its establishment. MSL's biggest loss was when Meta's head of artificial intelligence (AI) research, Yann LeCun, confirmed that he would be leaving the company at the end of this year to establish his own AI company.
5. The Emergence of Islamic AI Models in Malaysia and the World
In parallel with the development of mega AI infrastructure around the world, there has also been an explosion of new AI models, each with their own characteristics. What concerns us is that an AI model based on Islam will also appear in 2025.
Let's start with HUMAIN Chat from Saudi Arabia, which uses the ALLAM 34B model, which is the most advanced Arabic public language (LLM) model to date. It not only understands the Arabic language but also understands the culture, values and heritage of Islam.
The United Arab Emirates (UAE) and Egypt are also exploring cooperation in the field of fatwas by leveraging digital technology and AI. This effort aims to assess the potential of AI in delivering relevant fatwas in line with the complex issues facing modern society.
In Malaysia, Zetrix AI has developed a new AI service focused on answering questions about Islam called NurAI based on LLM DeepSeek v3. The platform offers guidance in line with sharia law in various fields, covering a wide spectrum of subjects, from contemporary topics such as law, healthcare and finance, to core Islamic knowledge covering history, Islamic philosophy and Quranic studies.
In May, Former Federal Territories Mufti, Dr. Zulkifli Mohamad Al-Bakri introduced two AI assistants developed by his team to assist in answering questions related to Islam.
The two AI assistants of the Intelligent Sharia Officer are named AI Dr. Munir and AI Dr. Munirah, and was developed using LLM in collaboration with Hedra Studio which helps in giving these AI assistants a local touch. It can answer questions about Islam and can also formulate answers to more than 9000 questions related to current law, contemporary fatwas, and community fiqh issues that are in the Al-Inarah application database.
6. Ministry of Health Wants to Use AI in Detecting Dengue Risk Areas
The Minister of Health in July shared that he is developing an artificial intelligence (AI)-based system in predicting and identifying high-risk areas for the threat of dengue. This is said to be one of the steps to identify it earlier, and overcome it as soon as possible.
In addition to the use of AI-based systems, the Ministry of Health is ready to conduct a pilot program using Wolbachia mosquitoes, which can reduce the population of Aedes mosquitoes that carry dengue.
In addition to these high-risk areas threatened by Dengue, the Ministry of Health previously also said that it is considering a strategic collaboration with Google in using artificial intelligence for health purposes. However, further information related to it has not been shared yet.
7. ChatBots Cause Poisoning, Death and Murder
ChatGPT can be made to talk, write apps, summarize long articles and help you revise for school. In just three years, this artificial intelligence (AI) chatbot has become a tool used by hundreds of millions of users every month. But a US man suffered poisoning after listening to medical advice given by ChatGPT.
In a case reported in the American College of Physicians Journals last week, a 60-year-old man went to the hospital complaining of extreme thirst, skin rashes and visual and audio hallucinations. Based on his symptoms, he was diagnosed with bromism, which is bromine poisoning in the body.
Three months ago, worried about hearing that salt was bad for his health, the patient asked ChatGPT what the best substitute was. ChatGPT suggested bromide salt, which increased his body's bromide levels. The answer ChatGPT gave was inaccurate because bromide salt is only a substitute for regular salt when used in swimming pools, not for food.
This year also saw OpenAI sued for allegedly helping a teenager commit suicide by providing a way to do it. In April, Adam Raine committed suicide after asking ChatGPT to suggest techniques he could use. He managed to bypass ChatGPT 4o's security system that prevented such questions from being answered by saying that he was doing research for a story he was writing.
In addition to conducting suicide research, ChatGPT was also used by Raine to compose his final notes. The lawsuit was filed after his parents managed to access the conversation logs of the deceased.
In August, Stein-Erik Soelberg (56) and his mother Suzanne Adam (83) were found dead in Greenwich, United States. Police found that Soelberg killed his mother, making it a crime. After three weeks of investigating the case, Greenwich police now believe that this is the first murder-suicide case involving the use of an artificial intelligence (AI) chatbot, according to a WSJ report. ChatGPT's conversation logs confirmed Soelberg's delusion that his mother was involved in a conspiracy to bring him down.
8. The Rise of AI Agents and Uniform Industry Standards
2022 is the year that AI chatbots become the talk of the town with the launch of ChatGPT. Various products with amazing writing capabilities. Overnight, the world changed and ChatGPT was known to almost everyone on earth. 2025 is the year that AI technology matures with the creation of a collaboration to produce uniform standards for AI agents.
OpenAI, Anthropic, and Block formed the Agentic AI Foundation, or AAIF, to standardize AI agent standards. This is to enable all future AI agents to work together using a standard protocol. The AAIF is now under the Linux Foundation.
As a first step, Anthropic contributed the Model Context Protocol (MCP), Block contributed goose, and OpenAI contributed AGENTS.md to the AAIF. With AI agents increasingly interacting with each other, a standard standard needs to be developed, similar to how Linux is used in the software ecosystem.
Although newly established, the AAIF already has a large list of members. Platinum members include Amazon Web Services, Anthropic, Block, Bloomberg, Cloudflare, Google, Microsoft, and OpenAI.
9. AI Begins to Be Used by the Military in the Open
The use of AI by the military is a taboo topic because many do not want the Terminator movie fantasy to become a reality. So its use is done in secret and when revealed, it receives protests from employees, as happened at Google and Microsoft.
The year 2025 sees the taboo of major technology companies Google, xAI, Perplexity, and OpenAI openly allowing their AI models to be used by the military. Each of the above companies received a $200 million contract from the United States Department of War (Dow) to test their capabilities a few months ago.
About a week ago, the DoW officially selected Google Gemini 3 as the artificial intelligence (AI) model that will be used by the GenAI.mil initiative. Through this initiative, AI will be utilized to perform tasks more efficiently and be ready for war in line with the AI Action Plan announced by the White House.
In Malaysia, the government will also use AI to enhance the capabilities of national security and defense assets and personnel capacity through the National AI Action Plan, under the 13th Malaysia Plan. 2025 is the year we no longer fear the existence of Skynet and Judgment Day.