Enfabrica
Founded Year
2019Stage
Acq - Talent | AcquiredTotal Raised
$260MAbout Enfabrica
Enfabrica focuses on artificial intelligence (AI) networking within the tech industry. The company provides network design and infrastructure solutions for computing, which aim to facilitate data distribution and performance. Enfabrica's technology is intended for sectors that need AI infrastructure. It was founded in 2019 and is based in Mountain View, California.
Loading...
Enfabrica's Products & Differentiators
Accelerated Compute Fabric SuperNIC (ACF-S) chip
The Industry's First 3.2 Tbps Multi-GPU SuperNIC Chip
Loading...
Research containing Enfabrica
Get data-driven expert analysis from the CB Insights Intelligence Unit.
CB Insights Intelligence Analysts have mentioned Enfabrica in 1 CB Insights research brief, most recently on Jun 20, 2024.
Expert Collections containing Enfabrica
Expert Collections are analyst-curated lists that highlight the companies you need to know in the most important technology spaces.
Enfabrica is included in 2 Expert Collections, including Semiconductors, Chips, and Advanced Electronics.
Semiconductors, Chips, and Advanced Electronics
7,494 items
Companies in the semiconductors & HPC space, including integrated device manufacturers (IDMs), fabless firms, semiconductor production equipment manufacturers, electronic design automation (EDA), advanced semiconductor material companies, and more
Artificial Intelligence (AI)
20,894 items
Enfabrica Patents
Enfabrica has filed 19 patents.
The 3 most popular patent topics include:
- network protocols
- parallel computing
- computer networking
Application Date | Grant Date | Title | Related Topics | Status |
|---|---|---|---|---|
6/9/2022 | 4/8/2025 | Computer memory, Parallel computing, Memory management, Virtual memory, Network protocols | Grant |
Application Date | 6/9/2022 |
|---|---|
Grant Date | 4/8/2025 |
Title | |
Related Topics | Computer memory, Parallel computing, Memory management, Virtual memory, Network protocols |
Status | Grant |
Latest Enfabrica News
Nov 4, 2025
Now Translating... 韓國 KAIST 教授、HBM（高頻寬記憶體）之父 Kim Jung-Ho 則在 Youtube 節目上直言，「AI 時代的主導權，正從 GPU 轉向記憶體！」有鑒於記憶體越來越重要，NVIDIA 可能會併購記憶體公司，如美光（Micron）或者 SanDisk。他也分享 HBF（高頻寬快閃記憶體，High Bandwidth Flash）的重要性，預期明年一、二月會有新進度，2027 至 2028 年將正式登場。 先前美系外資也提到，隨著傳統硬碟（HDD）供應可能自 2026 年底至 2027 年初出現瓶頸，近線儲存（Nearline SSD）方案有望加速導入，成為更具成本效益的替代選項。同時， HBF 被視為解決 AI 叢集記憶體容量瓶頸的關鍵技術。 在上一篇文章中提到，AI 推論時代下記憶體容量變得更加重要，大廠如何有效運用記憶體，成為更關鍵的焦點。其中，KV 快取（key-value cache）作為 AI 模型的短期記憶，也掌握著 AI 回答的反應速度。也因此，Kim Jung-Ho 更認為 HBF 有望成為下一代 AI 時代的重要記憶體技術，將與 HBM 並行發展，共同推動晶片大廠的業績成長。 HBF 是什麼？韓國記憶體雙雄如何佈局？ HBF 設計概念與 HBM 相似，透過矽穿孔（TSV）將多層晶片堆疊連接，HBF 採用 NAND 快閃記憶體進行堆疊，具備「容量更大、成本更具優勢」等特點。 Kim Joung-ho 指出，雖然 NAND 的速度不及 DRAM，但容量往往高出 10 倍以上，若以數百層乃至數千層堆疊方式建構，能有效滿足 AI 模型對龐大儲存的需求，可成為 NAND 版本的 HBM。在節目中，他更是預言「HBM 時代即將結束，HBF 時代要來了！」 Kim Joung-ho 預期，未來 AI 的記憶體架構會變成多層式設計，像一整座智慧圖書館，其中 GPU 內的 SRAM 如同桌上筆記本，最快、但容量最小；HBM 則是桌邊的書架，可以用於快速運算；HBF 彷彿是地下圖書館，儲存 AI 的深層知識，隨時補給 HBM；至於雲端儲存則是公共圖書館，透過光纖連接資料中心。他也認為，未來 GPU 將同時搭載 HBM 與 HBF，形成互補架構。 在 HBF 進度部分，SanDisk 與 SK 海力士正合作推動 HBF，並推進 HBF 的全球標準，目標 2026 年供應樣品、2027 年量產；三星近期也加入戰局，著手展開 HBF 產品的概念設計與初期開發。 受限記憶體供應短缺，華為從軟體下手釋放空間 由於中國很難獲得 HBM 等關鍵資源，因此華為開發新軟體工具「統一快取管理器」（UCM），無需使用 HBM 即可加速大型語言模型（LLM）的訓練與推理。值得注意的是，這款軟體也同樣以「KV 快取」（KV Cache）為中心的推理加速套件，主要是分級管理推論過程中產生的 KV 快取記憶數據，擴大推理上下文視窗，實現高吞吐、低時延的推理體驗，並降低每 Token 推理成本。 華為介紹，這款軟體會根據不同記憶體類型的延遲特性、各類 AI 應用的延遲需求，將 AI 資料分配在 HBM、標準 DRAM 與 SSD 之間。 UCM 分為三部分，最上層是透過「連接生態」（Connector），靈活對接業界的多樣引擎與多元算力，如華為昇騰、NVIDIA 等；再來透過中層「記憶管理」（Accelerator），透過 KV 快取動態多級管理，將演算法拆成適合快速運算的方式，使運算更高效；最後是「存儲協同」（Adapter），與專業共享儲存相結合的存取介面卡，以更高效的方式讀寫存儲資料，減少等待時間。 目前 NVIDIA 支持的晶片新創公司 Enfabrica 同樣從軟體下手，透過創新架構來降低記憶體成本， Enfabrica 利用自研的專用軟體，在 AI 晶片與大量低成本記憶體之間進行數據傳輸，進而在保證資料中心性能的同時，有效控制了成本。 若為了解決記憶體瓶頸，NVIDIA 不排除進一步布局硬體領域，入股記憶體公司也在可能範圍內。尤其當前 SanDisk 正積極投入 HBF 研發與佈局，HBM 之父的觀點不無道理。 （首圖來源：shutterstock）
Enfabrica Frequently Asked Questions (FAQ)
When was Enfabrica founded?
Enfabrica was founded in 2019.
Where is Enfabrica's headquarters?
Enfabrica's headquarters is located at 295 North Bernardo Avenue, Mountain View.
What is Enfabrica's latest funding round?
Enfabrica's latest funding round is Acq - Talent.
How much did Enfabrica raise?
Enfabrica raised a total of $260M.
Who are the investors of Enfabrica?
Investors of Enfabrica include NVIDIA, Sutter Hill Ventures, Alumni Ventures, Aramco Ventures, Atreides Management and 12 more.
Who are Enfabrica's competitors?
Competitors of Enfabrica include Celestial AI, UniFabriX, Tenstorrent, NeuReality, Hedgehog and 7 more.
What products does Enfabrica offer?
Enfabrica's products include Accelerated Compute Fabric SuperNIC (ACF-S) chip.
Loading...
Compare Enfabrica to Competitors
Mythic is an analog computing company that specializes in artificial intelligence (AI) acceleration technology. Its products include the M1076 Analog Matrix Processor and M.2 key cards, which provide power-efficient AI inference for edge devices and servers. Mythic primarily serves sectors that require real-time analytics and data throughput, such as smarter cities and spaces, drones and aerospace, and augmented reality (AR) or virtual reality (VR) applications. Mythic was formerly known as Isocline Engineering. It was founded in 2012 and is based in Austin, Texas.
Lightmatter works in silicon photonics within the computing and semiconductor industries. It develops photonic computing solutions, aims to support artificial intelligence infrastructure, and promotes industry collaboration through standardization efforts. The company's products serve to improve digital data processing and interconnectivity in artificial intelligence (AI) applications. It was founded in 2017 and is based in Mountain View, California.
DriveNets is a software company that focuses on network infrastructure for service providers and hyperscalers. The company provides routing solutions and networking fabric designed to support workloads and scalable networking. DriveNets serves the telecommunications and technology sectors with its networking solutions. It was founded in 2015 and is based in Raanana, Israel.
Wave Computing focuses on AI-native dataflow technology and processor architecture in the technology sector. The company provides products and solutions for deep learning and artificial intelligence applications, applicable to devices from edge to datacenter environments. Wave Computing's MIPS division offers licensing for processor architecture and core intellectual property. It is based in Campbell, California.
T-Head is a company specializing in artificial intelligence chips, storage system solutions, and intelligent terminal products within the tech industry. The company provides products and services that include AI chips, cloud computing, and data processing technologies for businesses that require computing and data storage solutions. It is based in Hangzhou, Zhejiang . T-Head operates as a subsidiary of Alibaba.
Tenstorrent is a computing company specializing in hardware focused on artificial intelligence (AI) within the technology sector. The company offers computing systems for the development and testing of AI models, including desktop workstations and rack-mounted servers powered by its Wormhole processors. Tenstorrent also provides an open-source software platform, TT-Metalium, for customers to customize and run AI models. It was founded in 2016 and is based in Toronto, Canada.
Loading...