Artificial Intelligence

US deep-tech AI startup Tiiny AI has just unveiled the world's smallest personal AI supercomputer, with the new Tiiny AI Pocket Lab, which has been officially verified by the Guinness World Record under "The Smallest MiniPC (100B LLM Locally)".

This is the first global unveiling of the new Tiiny AI Pocket Lab, which will fit in your hands -- or your pocket, duh -- and is capable of running up to a full 120-billion-parameter LLM (Large Language Model) entirely on-device, without the need of cloud connectivity, servers, or high-end GPUs.

Tiiny has developed its super-small AI supercomputer for energy-efficient personal intelligence, and the Tiiny AI Pocket Lab runs within a 65W power envelope. The new Tiiny AI Pocket Lab enables massive AI model performance at a fraction of the energy and carbon footprint of traditional GPU-based systems.

NVIDIA has built a new location verification technology that would help the company know where its AI GPUs are being operated, in a bid to help prevent its AI chips from being smuggled into countries where US exports apply.

In a new report from Reuters, the outlet said that NVIDIA has demonstrated its new location verification feature privately over the last few months, but hasn't released it just yet. When it's released, it'll act as a software option that customers and data center operators can install, so they can keep a tighter eye on the AI chips in their fleets.

The newly-developed software from NVIDIA is an opt-in, customer-installed service that keeps an eye on GPU usage, configuration, and errors. It has a pretty decent feature set, providing data center operators with the following abilities:

US authorities have busted an AI chip trafficking network that was attempting to send $160 million worth of NVIDIA H100 and H200 AI GPUs to China, as the smugglers were changing the shipments' final destination.

In a press release issued by the U.S. Department of Justice, authorities reported a trafficking network in Houston, Texas, that has been convicted of smuggling NVIDIA AI chips to China using a "complex scheme". Court documents reveal two individuals -- Alan Hao Hsu, and those who worked for his company, Hao Global LLC -- attempted to export NVIDIA H100 and H200 AI GPUs worth $160 million by manipulating official paperwork and hiding the "ultimate destination of the GPUs".

The network itself was busted by the discovery of a wire transfer that began in the People's Republic of China (PRC), with the NVIDIA AI GPUs shipped to US warehouses and then rebranded as "SANDKYAN", allowing the group to misclassify the AI GPUs and then export them.

NVIDIA CEO Jensen Huang sat down with Joe Rogan last week, where he provided some thoughts on his opinion of LLMs of today, turning into Terminator, tomorrow. Check it out:

Joe asked Jensen what he thought about AI and LLMs (Large Language Models) of today, which include generative AI, edge AI, agentic workflows, and more AI workloads. LLMs are super advanced right now, where they will replace many humans in their roles in the coming years, but some worry AI-boosted capabilities will go "too far", where they replace human beings as the "apex species".

Joe Rogan asked: "Well, I don't assume that it would do harm to us, but the fear would be that we would no longer have control and that we would no longer be the apex species on the planet. This thing that we created would now be. Is that funny?"

With the release of NVIDIA CUDA 13.1, the company is introducing the "largest and most comprehensive update to the CUDA platform since it was invented two decades ago." Alongside new features and performance improvements, the arrival of NVIDIA CUDA Tile is set to be a game-changer for AI programming.

The initial release is limited to the current Blackwell generation of GPU hardware (future versions will support more architectures), with CUDA Tile programming allowing users to bring their code up a layer with specific chunks of data called tiles. From there, the compiler and runtime determine "the best way to launch that work onto individual threads," including using hardware such as tensor cores.

With the new CUDA Tile programming, removing the need to define each thread's "path of execution," it reduces the effort required to write code that performs well across various GPU architectures.

Samsung Electronics' GDDR7 memory dies have won presidential recognition this week, as Samsung's continued technological competitiveness gains industry respect after its huge turnaround earlier this year.

In a new report from the Korea Times, we're hearing that at the recent 2025 Korea Tech Festival in Seoul, which was hosted by the Ministry of Trade, Industry, and Resources, the South Korean government awarded a presidential honor to Samsung's world-first 12nm-class, 40Gbps, 24Gb GDDR7 memory.

This isn't the first time for Samsung either, as it's the 12th time that the memory giant has received the presidential commendation, which is the highest number awarded to a single company. It previously received presidential recognition with its 14nm-class DDR5 memory in 2022, and the 64-layer 3D V-NAND flash back in 2017.

NVIDIA is reportedly cutting its B40 AI GPU production from an estimated 1.5-2M units to 900,000 units as Chinese AI companies are reportedly using RTX 5090 gaming graphics cards, as well as Hopper and locally-sourced AI chips instead.

In a new market report posted by insider @Jukan on X, we're hearing that NVIDIA's new B40 AI GPU, which complies with US export restrictions and is ready for China, is seeing its production cut, with an expected 1.5M to 2M units in 2H 2025, reduced to just 900,000 units.

The reason behind this is that the analysts' last "field trip" throughout Asia suggests that China's AI inference compute is still "largely dependent" on RTX 5090 gaming graphics cards (thanks to 32GB of speedy GDDR7 memory) and "previously improved" Hopper AI GPUs, and "some local chips".

NVIDIA will be the first customer for TSMC on its next-generation A16 process (1.6nm) with production taking off in 2027 at its Taiwan fab, while Apple will reportedly skip A16 and use the A4 (1.4nm) process.

We've been hearing rumors that NVIDIA would be the first customer for TSMC on its new A16 process node in 2026, but it looks like that will now take place in 2027, while other rumors said Apple was "not yet in talks" with TSMC to use its A16, with this new report suggesting the company is going directly to A14.

The new DigiTimes report also confirms that TSMC's plans to build 3 more 2nm semiconductor fabs in Taiwan in order to keep up with the ever-growing demand, with industry insiders saying TSMC's 2026 capex could reach $48-$50 billion, up from $40-42 billion this year. TSMC's advanced capacity at its Arizona fabs will also be going through some changes:

Microsoft has proclaimed on multiple occasions that Windows 11 and Windows in general are transforming into an 'Agentic OS,' and the latest 'Experimental Agentic Features' included in a recent Windows 11 preview build offer a first honest look at a Windows PC becoming an AI PC. The quick summary is that AI Agents will have their own accounts and privileges and run in the background while you're using your PC, leading to a situation where multiple users are logged in to your PC, with you being the only human.

Basically, you'll be able to interact with your PC using natural language. At the same time, these AI Agents will handle everything from launching office apps and creating charts to browsing, finding a deal, buying a new appliance, and searching through images to find something specific. These agents will run in the background, with Copilot as the primary interface.

Microsoft notes that you'll be able to monitor AI Agents like you can apps, while also confirming that these agents are prone to hallucinating and can even be tricked into installing malware or sending sensitive data and files to bad actors, which makes you wonder why anyone would enable these 'Experimental Agentic Features' when Microsoft is adamant that they pose a real security risk.

AMD is reportedly working on a couple of new Radeon AI PRO R9000 "RDNA 4" series consumer-grade AI cards, with a tease of the R9700S and R9600D leaking out.

AMD has officially listed its new consumer-grade Radeon AI PRO R9700S and Radeon AI PRO R9600D cards on its own support page, spotted by @KOMACHI_ENSAKA and @RubyRapids on X. We don't have any specifications on the new cards, and nothing official was posted on AMD's support page apart from the names of the two new Radeon AI PRO R9000 series "RDNA 4" cards.

Currently, AMD has its Radeon AI PRO R9700 workstation GPU on the market for $1299, featuring the full Navi 48 GPU core and up to 32GB of VRAM for the desktop. On the laptop side however, RDNA 4 hasn't graced any gaming laptops so far... so the new Radeon AI PRO R9700S could be a flagship RDNA 4-based offering for laptops, with the same Navi 48 GPU + 32GB of GDDR6 memory as the desktop R9700 variant.

The path towards combining both GPUs and HBM memory chips together is here, with NVIDIA and Meta "reviewing plans" to mount GPU cores on HBM.

The general concept is that placing GPU cores into the base die at the bottom of the HBM stack, with Meta and NVIDIA working with HBM leaders SK hynix and Samsung. In a new report from Korean outlet SEDaily and multiple industry insiders "familiar with the matter" that "next-generation 'custom HBM' architectures are being discussed, and among them, a structure that directly integrates GPU cores into the HBM base die is being pursued".

HBM (High Bandwidth Memory) is widely used on AI GPUs from both NVIDIA and AMD, with next-gen HBM4 right around the corner, and HBM4E not far behind it, and it's built perfectly for AI applications and mass amounts of data through its incredibly high memory bandwidth.

The IEEE ISSCC 2026 conference kicks off in February 2026, where SK hynix and Samsung will be showing off their very latest GDDR7, LPDDR6, and HBM4 memory products.

In the list of sessions at ISSCC 2026 is "DRAM, SRAM, and Non-Volatile Memories" and just that section alone -- outside of the hundreds of topics and sessions across the 5 days, we have SK hynix presenting its 1cnm SRAM 16Gb LPDDR6 memory with 14.4Gbps per pin of bandwidth, and its new 48Gbps 24Gb modules of GDDR7 memory for "mid-range" inference AI performance.

Samsung on the other hand will be showing off its new 36GB HBM4 modules with 3.3TB/sec of bandwidth per channel, and its new 16Gb LPDDR6 memory with 12.8Gbps of bandwidth.

NVIDIA is cooking up a new China-exclusive GPU with the new RTX 6000D Blackwell GPU, recently tested on Geekbench and teases what is happening under the hood.

The company has sliced the CUDA core count, VRAM capacity, and GPU clock speeds of its new RTX 6000D "Blackwell Pro for China" GPU. NVIDIA created the RTX 6000D after ever-increasing US export restrictions for AI hardware were imposed, with the RTX 6000D Blackwell GPU being a different SKU than the one available globally.

Inside, NVIDIA's new RTX 6000D GPU for China features 156 SMs or 19,968 CUDA cores, representing 17% less CUDA cores than the RTX PRO 6000, as well as less VRAM with 84GB of GDDR7 memory compared to 96GB GDDR7 on the full RTX PRO 6000. We should expect to see NVIDIA using 3GB GDDR7 memory modules on the RTX 6000D for China, indicating a 448-bit memory bus compared to the 512-bit memory bus on the RTX PRO 6000.

The unfortunate user with an expensive, dead NVIDIA RTX PRO 6000 card worth $10,000 has just had his card replaced by the company, as used PCIe boards enter the Chinese market.

Just a few days ago we reported that a user had sent NorthridgeFix a damaged NVIDIA RTX PRO 6000 card, which had been damaged during transport as the user kept the $10,000 card installed into the PCIe slot in his motherboard, and the PCIe board snapped. The board isn't replaceable by NVIDIA and the company doesn't sell any spare boards, so $10K was flushed down the drain... until now.

That was just a few days ago, with NVIDIA reaching out to Northridge Fix saying: "we recently saw your video featuring the NVIDIA RTX PRO 6000 GPU that a customer sent to you. We would like to provide a replacement unit to the customer. Also, if the original RTX PRO 6000 Blackwell Server Edition GPU can be returned to us, we'd like our technical team to help inspect and troubleshoot".

AMD has just unveiled one of the first models in its next-gen Instinct MI400 AI GPU family with the introduction of the Instinct MI430X, which was specifically created for HPC system buildouts, based on the new CDNA 5 architecture and using next-gen HBM4 memory.

In a new blog post, the company teased its new Instinct MI430X AI chip which was designed for large-scale AI environments, with some of the details unveiled on the new MI430X accelerator. AMD's new Instinct MI430X will feature the next-generation CDNA 5 architecture, and a huge 432GB of next-gen HBM4 memory with up to 19.6TB/sec of memory bandwidth.

AMD says its new MI430X is a "true successor" to the MI300A, with improvements on paper between both the chips -- the new CDNA 5 architecture, more VRAM and next-gen HBM4 with huge amounts more memory bandwidth -- making the MI430X ready for next-gen AI workloads.

Elon Musk has recently said that Tesla will need between 100 billion and 200 billion AI chips per year, and that both TSMC and Samsung can't meet that demand in the timeline he wants. Elon doubled-down on his plans for Tesla to build its own massive semiconductor fab.

The comments from the SpaceX and Tesla founder were made during the Baron Investment Conference, where he said that chip giant partners like TSMC and Samsung Foundry he has "tremendous respect for" but they need 5 years to get a new chip fab plant up and running, and he added that those 5 years are "an eternity".

Elon continued, explaining: "The production speed doesn't seem fast enough. When I asked them how long it would take from groundbreaking to completion of a new chip factory, they told me it would take five years to start production. I felt that five years was an endless wait for me".

Samsung has officially announced its mass production results for its new 2nm GAA process node, saying that it's 5% faster, 8% more efficient, and has 5% more area than its current 3nm GAA process.

Samsung Electronics announced the update on its new 2nm GAA process node mass production during its Q3 2025 earnings report, stating: "the 2nm first-generation gate-all-around (GAA) process has improved performance by 5%, power efficiency by 8%, and area by 5% compared to the 3nm second-generation process".

TSMC absolutely dominates semiconductor manufacturing, where according to its Q2 2025 sales, it owns 70.2% global market share, while Samsung Foundry only holds 7.3% of the global silicon manufacturing market. TSMC has enjoyed mega-success in the AI market, with big tech customers like Apple and NVIDIA using its 3nm process node for their latest silicon.

AMD confirms its next-gen Instinct MI400 AI accelerators for 2026 based on the new CDNA 5 architecture, with a huge 432GB of HBM4 memory, where it will compete against NVIDIA's next-gen Vera Rubin AI platform.

During its Financial Analysts Day event, the company confirmed its new MI400 series will come in multiple variants: MI455X is for training and inference, while the MI430X will be available in HPC variants.

AMD's new Instinct MI400 series feature the new CDNA 5 architecture, huge amounts of ultra-fast HBM4 memory, more AI formats and higher throughput, and standard-based rack-scale networking (UALoE, UAL, and UEC).

During a recent 'The Minds of Modern AI' roundtable discussion and interview with the Financial Times, editor Madhumita Murgia brought up the question or idea that we're currently living in an 'AI bubble' similar to the dotcom bubble of the 1990s, and that sooner or later it's going to burst and the market will "correct" itself.

For those who need a refresher or were too young to remember the 'dot-com bubble' of the late 1990s, this refers to a time when the stock market and investors saw the rise of the internet and the concept of the "World Wide Web" and decided to invest heavily in dot-com startups and companies. The bubble eventually burst, leaving only a few companies, such as Amazon, intact.

The reason people are comparing the AI boom to the dot-com boom is that they see parallels in the rapid growth and the lack of signs that AI investments are leading to actual profits. In response to the comparison and question posed to NVIDIA CEO Jensen Huang, his initial brief answer is 'no'; the AI boom is not like the dot-com boom because back then, the majority of the fiber was "dark," meaning unused.

NVIDIA's next-generation Rubin GPUs have reportedly entered production, with the company also having all major DRAM manufacturers provide it with HBM4 memory samples.

NVIDIA CEO Jensen Huang has been in Taiwan over the last few days, visiting TSMC and having extensive discussions with the semiconductor giant, with NVIDIA seeking more 3nm production capacity for its Blackwell and Rubin AI GPUs, seeing TSMC increase 3nm production by 50%. In even more recent news, UDN reports that NVIDIA's next-gen Rubin AI GPUs have now entered production, and that the company has received HBM4 memory samples, too.

NVIDIA only just received its first Rubin GPU at its labs a few days ago, and now we're at the production stage in just a few days, which is pretty crazy. UDN reports that regarding product and supply chain timing, Jensen said that demand for Blackwell is strong, and that it's not just for GPUs. Jensen explained: "NVIDIA is also manufacturing CPUs, network chips, switches, and many other chips related to Blackwell".