"SK HYNIX" content on TweakTown - Page 1

SK hynix plans to complete the development of next-gen HBM4E memory as early as 2026, preparing for the wave of next-generation AI GPUs.

Intel has placed orders to secure the rest of ASML's High-NA EUV lithography machines for 2024, placing orders for ALL machines built this year.

Meta's AI boss, Yann LeCun, said that the company has invested $30 billion in NVIDIA AI GPUs for AI training, new Llama 3 models coming soon.

Samsung has established a team for billions of dollars worth of HBM memory chip supply deals with NVIDIA, to beat HBM leader SK hynix.

The unstoppable demand of AI and one of the keys apart from the GPU and HBM is advanced packaging production, something TSMC is boosting right now.

SK hynix and Micron are sold out of 2024 + 2025 HBM supply, with TrendForce estimating a 5-10% increase in HBM memory pricing in 2025.

SK hynix says it's working on SSDs of up to 300TB capacity, pre-announced years ahead of release thanks to the unstoppable AI demand.

SK hynix expects to hit $10+ billion in revenue from HBM sales by the end of 2024, as most of its HBM 2024 and even HBM 2025 supply is already sold out.

SK hynix says that it's almost sold out of its 2025 HBM volume as AI demand reaches new highs, 12-Hi HBM3E production next quarter, 16-Hi HBM4 in 2028.

NVIDIA CEO Jensen Huang gives SK Group Chairman Chey Tae-won hand-written note: 'to our partnership and creating the future of AI and humanity together'.

SK hynix announces $14.6 billion will be spent on expanding production capacity of its next-gen HBM for future AI GPUs, noting 'soaring demand' for HBM.

Samsung confirms next-gen HBM4 memory is in development, should include 16-Hi stacks that will offer up to 256GB of HBM4 memory for future AI GPUs.

SK hynix partners with TSMC to strengthen HBM technological leadership, will work together on next-gen HBM4 and next-gen advanced packaging technology.

The Biden administration reaches an agreement to provide $6.1 billion in US government funding for Micron to make memory chips in New York and Idaho.

The US government is awarding up to $6.4 billion in CHIPS Act grants for Samsung's expansion of semiconductor operations in central Texas, USA.

Samsung Electronics co-CEO Qing Guixian visited Taiwan in a 'low key manner' to promote Samsung HBM, seeks Taiwan ODMs to help manufacture them.

SK hynix has high hopes for its US operations, with next-generation memory products will be delivered with next-gen advanced packaging technologies.

Samsung and SK hynix concentrating on US-based operations might push back South Korea construction schedules, with billions of dollars at stake.

Samsung announces it has manufactured a sample of 16-stack HBM, talks about next-gen HBM4 which will sample in 2025, go into mass production in 2026.

The US government is expected to award $6.6 billion in subsidies from CHIPS Act to Samsung to expand its chip output in Taylor, Texas, USA.

South Korea names 6 overseas partner institutions, including MIT and Yale University in the United States, as global technology cooperation centers.

SK hynix confirms advanced packaging facility for next-gen HBM, collaborating with Purdue University for R&D, with $3.87 billion facility in Indiana, USA.

SK hynix pushes past $100 billion market capetlization, becomes South Korea's second biggest company by market cap to Samsung, thanks to AI GPUs.

Samsung is working on its own 3D DRAM technology, calls it the 'industry-first disruptive tech adoption' that should arrive sometime after 2030.

Samsung Electronics CEO teases second-generation of its in-house Mach-2 AI accelerator, saying some clients want Mach-1 badly, so Mach-2 needs to happen.

Samsung forms dedicated HBM team for next-gen AI chip production yields, as it makes HBM4 memory and next-gen AI Mach-1 accelerator.

The South Korean government is considering subsidies for the semiconductor industry, as competition from rival countries is 'intensifying'.

Chinese president Xi Jinping told Dutch Prime Minister Mark Rutte that attempts to restrict China's access to technology will not stop the country.

Introspect Technology has shipped the world's first GDDR7 memory test system: massively parallel, 72-channel, 40Gbps PAM3 ATE-on-Bench test system.

NVIDIA's beefed-up Hopper H200 AI GPU delivers huge 3x performance gains in MLPerf 4.0 AI benchmarks with new TensorRT-LLM optimizations.

LG plans to invest $74.3 billion in South Korea through 2028 to work on AI, biotechnology, batteries, clean tech, displays, and more.

SK hynix will invest around $4 billion in building an advanced chip packaging facility in West Lafayette, a western city of Indiana, USA.

Samsung officially lists its next-gen GDDR7 memory modules at 28Gbps and 32Gbps, now sampling to partners to work on their next-gen GPUs.

SK hynix will begin construction of its new mega fab complex called Yongin Semiconductor Cluster in March 2025, completion date of 2046.

Micron says that all of its HBM memory is sold out for 2024, and most of its HBM supply is disappearing for 2025... insatiable AI GPU demand driving force.

SK hynix teased its next-gen GDDR7 memory at NVIDIA GTC 2024 this week: 16-24Gb, up to 40Gbps, and up to 160GB/sec memory bandwidth.

NVIDIA CEO Jensen Huang says the company will get HBM memory chips from Samsung to use with its best-in-class AI GPU hardware in the future.

HBM supply tightens with order volumes 'rising continuously' in 2024 says TrendForce, orders for new HBM memory are 'non-cancellable'.

SK hynix announces volume production of the industry's first HBM3E memory, will help the company cement its position as the total AI memory provider.

Samsung will have its HBM inside of NVIDIA's upcoming H200 AI GPU, also receiving certification for AMD's new Instinct MI300 AI GPU.

Samsung to address HBM yield issues by using MUF technique, which SK hynix pioneered, moving away from its NCF technology for HBM products.

JEDEC has reportedly eased HBM4 memory configurations to help HBM manufacturers, reducing thickness of HBM4 to 775 micrometers for 12-layer, 16-layer HBM4.

Samsung, LG, and SK are 'concerned' over their investments in the US, with spikes in construction costs and subsidy uncertainties.

Micron partnering with NVIDIA for its upcoming H200 AI GPU with upgraded HBM3E memory has HBM competitors like SK hynix and other 'shocked'.

SK hynix says the first 50 years of the semiconductor industry was about the front-end, while the next 50 years are about the back end aka packaging.

The first generation of GPUs based on GDDR7 memory will use 16Gb dies and 2GB VRAM capacities, at 32Gbps bandwidth, says hardware leaker.

Samsung is expected to make 'groundbreaking reveal' at NVIDIA GTC 2024, where it will show off physical HBM3E 12-Hi memory for next-gen AI GPUs.

SK hynix reportedly provided initial samples of its new 12-layer HBM3E memory to NVIDIA last month, offering 36GB HBM3E per stack.

JEDEC publishes the next-gen GDDR7 graphics memory standard: double the bandwidth of GDDR6, for the future of gaming and AI GPU hardware.

TSMC has launched its 2024 campus recruitment campaign, with 6000 jobs on offer with engineers and technicians highty sought after in Taiwan.