MIT engineers just built a salt-sized chip that could make your phone faster and cooler

What if your next phone call was clearer, your downloads faster, and your battery lasted longer - all thanks to a material smaller than a grain of salt? MIT engineers have just unveiled a breakthrough that could make this a reality, using gallium nitride (GaN) to supercharge the chips inside our devices.
For years, GaN has been the gold standard for speed and power in electronics, outshining traditional silicon - especially in high-speed communication and demanding devices. The catch? GaN has always been too expensive and tricky to use in everyday gadgets. But MIT's new fabrication method changes the game.
Instead of building whole chips from GaN, the MIT team created thousands of tiny GaN transistors on a single wafer. These are sliced into 'dielets,' each about the size of a grain of salt, and then carefully placed onto a regular silicon chip using a special vacuum tool. The process uses copper instead of costly gold and is done at low temperatures, so it doesn't damage the chips. This makes the method cheap, scalable, and compatible with existing factories - no need for a complete production overhaul.
Why does this matter? By using only a tiny amount of GaN, the cost drops dramatically. Spacing these mini GaN transistors across the silicon chip also keeps things cool, helping prevent overheating - a common problem as gadgets get smaller and more powerful.
To prove it works, the team built a power amplifier - the part of your phone that boosts your signal. The result? A chip that outperformed today's silicon-only versions, offering stronger signals, faster speeds, and better energy efficiency. For users, this could mean clearer calls, speedier downloads, more reliable wireless, and longer battery life.
What's more, the entire process uses standard industry tools and copper, making it easy for chip makers to adopt. The team's new vacuum tool can move and align each dielet with nanometer precision, ensuring perfect placement every time.
The research was presented at the IEEE Radio Frequency Integrated Circuits (RFIC) Symposium in June 2025, a major event for semiconductor innovation. Pradyot Yadav, an MIT graduate student and lead author, explained at the conference, 'We've combined the best of what exists in silicon with the best possible gallium nitride electronics. These hybrid chips can revolutionise many commercial markets.'
Looking ahead, this technology could even help power quantum computers, since GaN works better than silicon at the extremely low temperatures those systems require.

Try Our AI Features

Explore what Daily8 AI can do for you:

Learn with AIFact CheckingText to SpeechAI Summary

Related Articles

Time of India

12 hours ago

• Time of India

ChatGPT might be making you think less: MIT study raises ‘red flags' about AI dependency

ChatGPT might be making you think less: MIT study raises 'red flags' about AI dependency As AI tools become part of our daily routines, a question is starting to bubble up: What happens when we rely on them too much? A new study from MIT's Media Lab takes a closer look at how tools like ChatGPT may be affecting our brains. And what the researchers found is worth paying attention to. The study focused on how people engage mentally when completing tasks with and without AI. It turns out that while ChatGPT can make writing easier, it may also be reducing how much we think. According to the research team, participants who used ChatGPT showed noticeably lower brain activity than those who did the same task using Google or no tech at all. The findings suggest that depending on AI for tasks that require effort, like writing, decision-making, or creative thinking, could weaken the very mental muscles we're trying to sharpen. ChatGPT users show lowest brain activity in MIT's groundbreaking study The experiment involved 54 participants between the ages of 18 and 39. They were split into three groups and asked to write essays in response to prompts similar to those on standardised tests. by Taboola by Taboola Sponsored Links Sponsored Links Promoted Links Promoted Links You May Like Buy Brass Idols - Handmade Brass Statues for Home & Gifting Luxeartisanship Buy Now Undo Group 1 used ChatGPT to generate their answers Group 2 relied on Google Search to find and compile information Group 3 worked without any tools, using only their knowledge and reasoning While they worked, each participant wore a headset that tracked electrical activity across 32 areas of the brain. The aim was to see how engaged their minds were during the process. (The research was led by Dr. Nataliya Kosmyna along with a team that included Ashly Vivian Beresnitzky, Ye Tong Yuan, Jessica Situ, Eugene Hauptmann, Xian-Hao Liao, Iris Braunstein, and Pattie Maes.) ChatGPT may be hurting your creativity, MIT researchers warn The results were pretty clear: the group that used ChatGPT showed the lowest brain activity of all three groups. In particular, areas linked to memory, creativity, and concentration were significantly less active. In contrast, those who wrote without help from AI showed the highest mental engagement. They had to organise their thoughts, build arguments, and recall information, all things that activated the brain more deeply. Even the group using Google Search showed more engagement than the AI group, possibly because the process of looking for and evaluating information keeps the brain involved. There was another telling detail. Many in the ChatGPT group simply pasted the prompts into the tool and copied the output with little to no editing. Teachers who reviewed their essays said they felt impersonal, calling them 'soulless.' Dr. Kosmyna put it bluntly: 'They weren't thinking. They were just typing.' AI dependency Short-term efficiency, long-term cost Later in the study, researchers asked participants to rewrite one of their essays, this time without using any tools. The ChatGPT users struggled. Many couldn't remember their original arguments or structure. Since they hadn't processed the material deeply the first time, it hadn't stuck. Kosmyna described this as a red flag: 'It was efficient. But nothing was integrated into their brains.' That raises a broader concern: if AI is doing the heavy lifting, are we still learning? Or are we just moving text around while our cognitive skills fade in the background? The growing concern among psychiatrists and educators Dr. Zishan Khan, a psychiatrist who works with students, says he's already seeing signs of AI overuse in younger people. 'The neural pathways responsible for thinking, remembering, and adapting—they're weakening,' he explained. The fear is that early and frequent reliance on tools like ChatGPT might lead to long-term cognitive decline, especially in developing brains. MIT's team is now expanding their research to see how AI affects people in other fields. They've already started looking at coders who use tools like GitHub Copilot. So far, Kosmyna says the early results there are 'even worse' in terms of mental engagement. A word of warning for classrooms and beyond Interestingly, the MIT researchers shared their findings before going through the full peer review process, something that's uncommon in academic research. But Kosmyna felt the potential impact was urgent enough to make an exception. 'I'm really concerned someone might say, 'Let's introduce ChatGPT into kindergarten classrooms,'' she said. 'That would be a terrible mistake. Young brains are especially vulnerable.' To prove just how easy it is to lose the depth of complex research, the team did something clever: they planted subtle factual 'traps' in the study. When readers ran the paper through ChatGPT to summarise it, many versions came back with key errors, including details the researchers never even included. What does this mean for the future of AI use Not at all. The tool isn't the enemy. It can be incredibly helpful, especially when used wisely. But this study reminds us that how we use AI matters just as much as whether we use it. Here are a few takeaways from the researchers: Use AI as a partner, not a replacement. Let it offer ideas, but make sure you're still doing the core thinking. Stay actively involved. Skipping the process of learning or writing just to get a result means you're not absorbing anything. Be cautious in education. Children need to build foundational skills before leaning on technology. Also read | Nvidia CEO Jensen Huang swears by these 6 effective management strategies to run a company like a genius AI Masterclass for Students. Upskill Young Ones Today!– Join Now

Business Standard

2 days ago

• Business Standard

IIT Madras, STL launch multi-core fibre testbed at TTDF Symposium

Indian Institute of Technology Madras and optical solutions major STL Technologies have jointly launched a new multi-core fibre (MCF) testbed, with the aim of advancing India's optical communication infrastructure. The facility was inaugurated at the Telecom Technology Development Fund (TTDF) Symposium held by the Department of Telecommunications (DoT) at the IIT Madras Research Park in Chennai. The infrastructure includes both underground (4.07 km) and aerial (1.20 km) cable installations. The cables incorporate 12 multi-core fibres (four cores each) alongside 12 standard single-mode fibres, and will be open to both academic and industry partners. Essentially a research and development (R&D) platform for testing the capabilities of optical fibres in a controlled environment, the testbed will be used to assess use cases and test component interoperability within the MCF ecosystem, the DoT said in a statement. Speaking at the inaugural session, Telecom Secretary Neeraj Mittal said India was at a critical juncture in its innovation journey and stressed the need for collaboration and optimal utilisation of the country's resources. He also announced plans for a dedicated portal to facilitate such collaboration. Launched in 2022, TTDF is a government initiative that supports R&D in next-generation telecom technologies, particularly those aimed at bridging the digital divide in rural and remote areas. It has received over 1,300 proposals for R&D funding since its inception. So far, 120 projects have been approved with over ₹500 crore in funding, and more than ₹187 crore has already been disbursed to beneficiaries, including academic institutions, micro, small and medium enterprises (MSMEs), and startups, the DoT said. Implemented by Digital Bharat Nidhi, TTDF aims to reduce reliance on telecom imports, promote indigenous manufacturing and develop intellectual property in the sector. The TTDF Symposium 2025 was organised to provide a platform for funding beneficiaries to discuss synergies and potential pathways to advance their research and development. The event brought together academics, startups and industry representatives involved in the programme.

Business Standard

2 days ago

• Business Standard

Essay aid or cognitive crutch? MIT study tests the cost of writing with AI

While LLMs reduce cognitive load, a new study warns they may also hinder critical thinking and memory retention - raising concerns about their growing role in learning and cognitive development Rahul Goreja New Delhi A new study from the Massachusetts Institute of Technology (MIT) Media Lab has raised concerns about how artificial intelligence tools like ChatGPT may impact students' cognitive engagement and learning when used to write essays. The research, led by Nataliya Kosmyna and a team from MIT and Wellesley College, examines how reliance on large language models (LLMs) such as ChatGPT compares to traditional methods like web searches or writing without any digital assistance. Using a combination of electroencephalogram (EEG) recordings, interviews, and text analysis, the study revealed distinct differences in neural activity, essay quality, and perceived ownership depending on the method used. Note: EEG is a test that measures electrical activity in the brain. Setup for cognitive engagement study 54 participants from five Boston-area universities were split into three groups: those using only ChatGPT (LLM group), those using only search engines (search group), and those writing without any tools (brain-only group). Each participant completed three writing sessions. A subset also participated in a fourth session where roles were reversed: LLM users wrote without assistance, and brain-only participants used ChatGPT. All participants wore EEG headsets to monitor brain activity during writing. Researchers also interviewed participants' post-session and assessed essays using both human markers and an AI judge. Findings on neural engagement Electroencephalogram (EEG) analysis showed that participants relying solely on their own cognitive abilities exhibited the highest levels of neural connectivity across alpha, beta, theta, and delta bands — indicating deeper cognitive engagement. In contrast, LLM users showed the weakest connectivity. The search group fell in the middle. 'The brain connectivity systematically scaled down with the amount of external support,' the authors wrote. Notably, LLM-to-Brain participants in the fourth session continued to show under-engagement, suggesting a lingering cognitive effect from prior LLM use. Essay structure, memory, and ownership When asked to quote from their essays shortly after writing, 83.3 per cent of LLM users failed to do so. In comparison, only 11.1 per cent of participants in the other two groups struggled with this task. One participant noted that they 'did not believe the essay prompt provided required AI assistance at all,' while another described ChatGPT's output as 'robotic.' Essay ownership also varied. Most brain-only participants reported full ownership, while the LLM group responses ranged widely from full ownership to explicit denial to many taking partial credit. Despite this, essay satisfaction remained relatively high across all groups, with the search group being unanimously satisfied. Interestingly, LLM users were often satisfied with the output, even when they acknowledged limited involvement in the content's creation. Brain power trumps AI aid While AI tools may improve efficiency, the study cautions against their unnecessary adoption in learning contexts. 'The use of LLM had a measurable impact on participants, and while the benefits were initially apparent, as we demonstrated over the course of four months, the LLM group's participants performed worse than their counterparts in the Brain-only group at all levels: neural, linguistic, scoring,' the authors wrote. This pattern was especially evident in session four, where brain-to-LLM participants showed stronger memory recall and more directed neural connectivity than those who moved in the opposite direction. Less effort, lower retention The study warns that although LLMs reduce cognitive load, they may diminish critical thinking and reduce long-term retention. 'The reported ownership of LLM group's essays in the interviews was low,' the authors noted. 'The LLM undeniably reduced the friction involved in answering participants' questions compared to the search engine. However, this convenience came at a cognitive cost, diminishing users' inclination to critically evaluate the LLM's output or 'opinions' (probabilistic answers based on the training datasets),' it concluded.