BROADCOM

About Broadcom

Technology company

Recent News about Broadcom

• OpenAI, Anthropic, Big Tech urge cyber ‘defensive surge’ on AI hacks

  On August 27, more than 100 companies including OpenAI, Anthropic, Microsoft, Alphabet and Amazon signed a joint letter calling for a “society‑wide defensive surge” against AI‑driven cyberattacks. The letter warns that there is only a limited window to significantly strengthen digital defenses before increasingly capable AI tools fuel a wave of automated hacking.

  Source: Reuters (via Investing.com)2026-08-27T17:02:00.000Z

• OpenAI Jalapeño chip posts up to 4x faster AI inference

  OpenAI published the first benchmarks for its custom Jalapeño inference chip on August 25, 2026, claiming 1.5 to 1.9 times more AI work per watt and up to 4.1 times higher interactive performance than leading systems. The company said Jalapeño outperforms Nvidia’s latest GB300 and other accelerators on the InferenceX benchmark across several large models and is already being integrated into its production stack.

  Source: OpenAI2026-08-25T21:08:59.911Z

• Samsung and SK sign $950B AI chip and data center deals

  On July 25, 2026, South Korea’s presidential Blue House detailed about $950 billion in AI chip and infrastructure cooperation over five years between Korean conglomerates and US tech firms. SK Group will pursue around $750 billion in advanced memory supply deals, while Samsung Electronics signed a memorandum of understanding with Broadcom worth up to $200 billion, alongside large AI data center and physical AI projects with Nvidia, Anthropic, NAVER, SK Telecom, Brookfield and Hyundai Motor Group.

  Source: Aju Press2026-07-25T06:20:00.000Z

• Super Micro stock sinks 60% as AI server bet strains margins

  On 20 July 2026, Iber Empresa reported that Super Micro Computer’s share price has fallen about 60% from its 52‑week high and 21% in the past four weeks, leaving investors in its June equity raise under water. The article cites razor‑thin 9.9% gross margins, $8.8 billion in debt, heavy inventories tied to $39 billion of non‑binding AI server orders, and potential supply risk from an ITC probe into Samsung DRAM as key pressures.

  Source: Iber Empresa2026-07-20T12:50:00.000Z

• Meta plans Iris AI chip rollout to double data center compute by 2027

  Internal memos seen by Reuters show Meta will begin manufacturing its in‑house "Iris" AI data center chip in September 2026 and aims to double AI computing capacity to 14GW by 2027. The MTIA project, built with partners including Broadcom and TSMC, is intended to reduce dependence on Nvidia and AMD as Meta spends up to $145 billion on AI infrastructure this year.

  Source: CyberNews (Reuters syndication)2026-07-09T21:06:58.332Z

• DeepSeek develops in-house AI inference chip to cut Nvidia dependence

  Reuters reports that Chinese startup DeepSeek is secretly developing its own AI inference chip to reduce reliance on Nvidia and Huawei hardware. The effort is in early stages, with the company recruiting chip designers and sounding out foundries and memory suppliers.

  Source: Reuters (via WHTC)2026-07-07T10:34:00.000Z

• Apple reportedly skips M6 Pro/Max to fast‑track AI‑focused M7 chip

  On June 26, 2026, The Eastern Herald reported that Apple plans to forgo M6 Pro and M6 Max Mac chips and instead accelerate development of an AI-optimized M7 for future MacBook models. The report situates the move alongside broader industry shifts toward on-device AI, including Apple’s work with OpenAI and Broadcom on the Jalapeño AI inference chip. Apple has not formally confirmed the specific M7 roadmap described.

  Source: The Eastern Herald2026-06-26T05:08:28.311Z

• OpenAI Jalapeño chip targets cheaper, faster AI inference

  OpenAI on June 25, 2026 publicly detailed Jalapeño, its first custom AI inference chip co-developed with Broadcom, via multiple tech outlets. The processor is designed to cut reliance on Nvidia GPUs and lower the cost of running services like ChatGPT and Codex, with early tests showing higher performance per watt than current accelerators.

  Source: iTechPost2026-06-25T13:30:00.000Z

• OpenAI Jalapeño chip aims to slash AI inference costs

  On June 25, 2026, Fortune revealed that OpenAI has officially named its first in‑house accelerator chip 'Jalapeño,' built with Broadcom and optimized for large language model inference. OpenAI plans to deploy Jalapeño across its data centers starting in late 2026 to reduce reliance on third‑party GPUs and lower serving costs for models like GPT‑5.5.

  Source: Fortune2026-06-25T09:48:00.000Z

• OpenAI Jalapeño AI chip targets cheaper, faster LLM inference

  OpenAI and Broadcom on June 24, 2026 unveiled Jalapeño, OpenAI’s first custom “Intelligence Processor” built specifically for large language model inference and manufactured by TSMC. Engineering samples are already running workloads such as GPT‑5.3‑Codex‑Spark in OpenAI’s labs, with large‑scale data center deployment planned beginning in 2026. ([openai.com](https://openai.com/index/openai-broadcom-jalapeno-inference-chip/))

  Source: DIGITIMES2026-06-25T01:03:00.000Z