#HyperAccel
HyperAccel CEO Kim Joo-young said Nov 12 the firm asked Samsung Foundry for 20,000 Bertha LPUs on 4nm SF4X and initially got 10,000. It also asked SK hynix for about 100,000 LPDDR5X chips for 192GB, eight-channel cards and got a partial allocation, The Elec reported.
November 16, 2026 at 8:04 AM
The HyperAccel squeeze is about volume, not technology, TheElec reports. SK hynix covered only part of a 100,000-chip memory request, roughly what 10,000 Bertha cards need. Mass production is planned for the second quarter of next year. #semiconductors
HyperAccel Reportedly Gets 10,000 of 20,000 Bertha Chips From Samsung
Korean AI chip startup HyperAccel asked Samsung Foundry for 20,000 Bertha accelerators and was allocated 10,000, TheElec reported, and SK hynix covered only part of the memory it needs.
voltgridnews.com
November 14, 2026 at 12:13 PM
HyperAccel Unveils LPU-Powered AI Infrastructure for Cost-Effective Inference at SC26#South_Korea#Seoul#AI_Inference#HyperAccel#SC26
HyperAccel Unveils LPU-Powered AI Infrastructure for Cost-Effective Inference at SC26
HyperAccel will present its innovative LPU technology at SC26, showcasing high-efficiency AI inference solutions that promise to reduce costs and power usage significantly.
third-news.com
November 12, 2026 at 2:24 PM
Semifive starts shipping AI chips built on Samsung’s 4nm process

Read more:
https://quantumzeitgeist.com/samsung-4nm-process-semifive-starts/
Semifive Starts Shipping AI Chips Built On Samsung’s 4nm Process
SEMIFIVE announced today the mass production of a data center AI inference accelerator for HyperAccel, marking its first large-scale project utilizing Samsung Foundry’s 4nm process node. Following an initial production contract, mass production volumes are projected to expand with follow-on purchase orders aligned with HyperAccel’s rollout.
quantumzeitgeist.com
September 8, 2026 at 12:10 PM
SEMIFIVE Launches Mass Production of Revolutionary AI Inference Accelerator 'Bertha' on Advanced 4nm Technology#South_Korea#Seoul#SEMIFIVE#HyperAccel#Bertha
SEMIFIVE Launches Mass Production of Revolutionary AI Inference Accelerator 'Bertha' on Advanced 4nm Technology
SEMI FIVE has kicked off the mass production of HyperAccel's AI inference accelerator, 'Bertha,' utilizing Samsung's cutting-edge 4nm process technology, signaling immense growth.
third-news.com
September 8, 2026 at 5:31 AM
00:00 Introduction and Contextual Awareness
01:01 Nvidia-Meta (Grace)
03:57 AMD-Meta
06:04 Amazon and OpenAI Collaboration
07:28 Nvidia-OpenAI
09:40 Nvidia-Groq
11:32 AI-Chip Startup Funding
11:46 MatX
14:52 Taalas
16:54 HyperAccel
18:24 SambaNova
22:07 Ayar Labs
25:40 Nvidia Earnings
29:20 Broadco…
We can't believe how much was invested in this AI-chip company: BWR 10
Subscribe to the Byrne-Wheeler Report on YouTube for more content like this.
www.youtube.com
March 16, 2026 at 10:49 AM
HyperAccel seeks to displace Nvidia by delivering a lower-cost AI accelerator. Its Bertha 500 promises to double performance while raising power efficiency 12× compared with a Hopper-generation Nvidia GPU. Based on the company’s latency-processing unit (LPU) architecture, the startup’s NPU achieves…
HyperAccel Promises Power Efficiency Gains
To read the full post, click through. Once at http://xpu.news, you can also comment and submit your own posts to appear in this feed.
xpu-news.hn.plus
March 8, 2026 at 12:21 PM
Feed: "EE Times"
By: Sally Ward-Foxton on Wednesday, February 18, 2026
Korean Startup Takes On Cost and Latency With LLM-Specific Chip
HyperAccel is about to launch its Bertha 500 chip, an LLM inference accelerator designed for economical token generation in the data center.
www.eetimes.com
February 19, 2026 at 2:53 AM
## Automated Design Verification and Optimization of Micro-Electro-Mechanical System (MEMS) Accelerometers via HyperScore-Guided Reinforcement Learning

**Abstract:** This paper introduces a novel framework for automated design verification and optimization of Micro-Electro-Mechanical System (MEMS)…
## Automated Design Verification and Optimization of Micro-Electro-Mechanical System (MEMS) Accelerometers via HyperScore-Guided Reinforcement Learning
**Abstract:** This paper introduces a novel framework for automated design verification and optimization of Micro-Electro-Mechanical System (MEMS) accelerometers leveraging a hyper-dimensional evaluation pipeline and reinforcement learning (RL). Current accelerometer design relies heavily on iterative finite element analysis (FEA) simulations and manual optimization, which is time-consuming and computationally expensive. We propose a system, "HyperAccel," that integrates numerical simulation, logical consistency verification, original concept detection, and impact forecasting to provide a hyper-score representing the potential of a given accelerometer design.
freederia.com
November 24, 2025 at 6:14 AM
Пет пъти по-енергоефективен от H100: HyperAccel разработва икономичния чип Bertha 500 за ИИ-инференс

Южнокорейският стартъп HyperAccel, според EETimes, се готви да пусне Bertha 500 — специализиран чип, предназначен за ИИ-инференс. Благодарение на уникалната си архитектура, се твърди, че чипът…
Пет пъти по-енергоефективен от H100: HyperAccel разработва икономичния чип Bertha 500 за ИИ-инференс
Южнокорейският стартъп HyperAccel, според EETimes, се готви да пусне Bertha 500 — специализиран чип, предназначен за ИИ-инференс. Благодарение на уникалната си архитектура, се твърди, че чипът генерира пет пъти повече токени в секунда от решенията, базирани на графични процесори, при същото ниво на TOPS. Bertha 500 поставя силен акцент върху ефективността на разходите. За тази цел той използва LPDDR памет вместо скъпата HBM. Чрез елиминиране на традиционната йерархия на паметта, той постига 90% използване на честотната лента на LPDDR. Допълнителни подобрения в ефективността се постигат чрез оптимизиране на архитектурата специално за задачите за инференс.
www.kaldata.com
February 19, 2026 at 10:17 AM
Нов корейски NPU ускорява изкуствения интелект с 60%, спестява 44% енергия

Корейски изследователи от Училището по компютърни науки KAIST, в сътрудничество с HyperAccel Inc, са разработили нов невронен процесор, който подобрява производителността на генеративните модели с изкуствен интелект средно…
Нов корейски NPU ускорява изкуствения интелект с 60%, спестява 44% енергия
Корейски изследователи от Училището по компютърни науки KAIST, в сътрудничество с HyperAccel Inc, са разработили нов невронен процесор, който подобрява производителността на генеративните модели с изкуствен интелект средно с 60%. Съвременните модели на генеративен изкуствен интелект, като ChatGPT-4 и Gemini 2.5, изискват не само висока пропускателна способност на паметта, но и значително повече памет. В резултат на това компаниите, които използват облачен генеративен изкуствен интелект, включително Microsoft и Google, купуват стотици хиляди графични процесори от NVIDIA. Нов невронен процесор, създаден от корейски изследователи, подобрява производителността на генеративните модели с изкуствен интелект с 60% и намалява консумацията на енергия с 44%.
www.kaldata.com
July 8, 2025 at 10:26 AM
SEMIFIVE Concluded Mass Production Contract for AI Chip with HyperAccel

www.viv-media.com/infozx-34683...
SEMIFIVE Concluded Mass Production Contract for AI Chip with HyperAccel
www.viv-media.com
October 11, 2024 at 1:23 PM