Marvell

v1.0.0

Provides in-depth insights on Marvell's evolution from storage controllers to a leading custom AI ASIC supplier for data centers.

0· 39·0 current·0 all-time

Install

OpenClaw Prompt Flow

Install with OpenClaw

Best for remote or guided setup. Copy the exact prompt, then paste it into OpenClaw for hanxueyuan/marvell.

Previewing Install & Setup.
Prompt PreviewInstall & Setup
Install the skill "Marvell" (hanxueyuan/marvell) from ClawHub.
Skill page: https://clawhub.ai/hanxueyuan/marvell
Keep the work scoped to this skill only.
After install, inspect the skill metadata and help me finish setup.
Use only the metadata you can verify from ClawHub; do not invent missing requirements.
Ask before making any broader environment changes.

Command Line

CLI Commands

Use the direct CLI path if you want to install manually and keep every step visible.

OpenClaw CLI

Bare skill slug

openclaw skills install marvell

ClawHub CLI

Package manager switcher

npx clawhub@latest install marvell
Security Scan
VirusTotalVirusTotal
Benign
View report →
OpenClawOpenClaw
Benign
high confidence
Purpose & Capability
The skill name/description promise (background and analysis of Marvell's transition to AI ASICs) matches the SKILL.md content. There are no unrelated requirements (no env vars, binaries, or installs) that would be inconsistent with an informational skill.
Instruction Scope
SKILL.md is static content with a small 'read_when' trigger list and an article-style briefing. It does not instruct the agent to access files, environment variables, network endpoints, or external services beyond presenting the written content.
Install Mechanism
No install spec and no code files are present, so nothing is written to disk or executed during install. This is low-risk and expected for a documentation-style skill.
Credentials
The skill declares no required environment variables, credentials, or config paths. There is no indication of disproportionate or unexplained secret access.
Persistence & Privilege
always is false and the skill does not request elevated or persistent privileges. Autonomous invocation is allowed by platform default, which is normal; the skill itself does not request additional privileges.
Scan Findings in Context
[no_regex_findings] expected: The static scanner found nothing because this is an instruction-only skill with a single SKILL.md text file. That is expected for a read-only informational skill.
Assessment
This skill appears low-risk from a security and permissions standpoint because it contains only an informational article and requests no credentials or installs. However, the source and homepage are unknown — verify any factual claims (financial figures, dates, market rankings) against reputable sources before using this content for decisions. If you require provenance or citations, ask the publisher for sources or prefer skills that link to primary references.

Like a lobster shell, security has layers — review code before you run it.

latestvk97f05w3fgz89ks3rs8eqs8pth85j6yy
39downloads
0stars
1versions
Updated 1d ago
v1.0.0
MIT-0

Marvell:被低估的 AI 芯片幕后推手

当所有人都在谈论 NVIDIA 的 GPU 时,另一家公司正在为科技巨头们悄悄设计 AI 推理芯片。这家公司就是 Marvell Technology。

起源:一条街的名字

1995年,印尼华裔工程师 Sehat Sutardja 在加州圣克拉拉创立了 Marvell。公司名字取自他家乡印度尼西亚的一条街道——"Marvell"。创业三人组带着一个简单信念:芯片设计可以做得更集成、更优雅。四年后,他们推出了业界首款单芯片 Gigabit 以太网收发器,一炮而红。

过山车般的转型史

Marvell 的故事是一部半导体行业的教科书级转型案例:

2000年互联网泡沫时期,公司市值一度突破 300 亿美元,随后暴跌 90%。但管理层没有躺平——2006年以 60 亿美元吃下 Intel 的 XScale 处理器业务,2016年又用同样的 60 亿美元收购了 Cavium(网络和基础设施芯片),2018年收购 Avera Custom ASIC 正式踏入 AI 定制芯片领域,紧接着 2019 年拿下 Inphi(高速互连芯片)。每一次收购都在为最终的战略转型蓄力。

为什么 AI 芯片巨头需要 Marvell?

NVIDIA 卖的是通用 GPU,但 Google、Amazon、Microsoft 这些云厂商需要的不是通用方案——他们需要为自家工作负载量身定制的 ASIC。Marvell 的核心竞争力恰恰在于此:

  • 高速 SerDes 技术:芯片间互连的速度直接决定 AI 集群效率,Marvell 在这一领域积累了超过 20 年的经验
  • 定制 ASIC 设计能力:从架构到流片,设计周期 2-3 年,客户一旦选定就很难更换供应商(切换成本极高)
  • TSMC 先进制程的优先合作权:在 3nm/5nm 节点上,Marvell 是少数能拿到充足产能的芯片设计公司之一

数字说话

2024 财年,Marvell 营收约 57 亿美元。其中数据中心收入占比已突破 50%,AI 相关收入年增速超过 100%。在 AI ASIC 市场,它是仅次于 Broadcom 的全球第二大供应商。

竞争格局

在定制 AI 芯片赛道上,Marvell 的主要对手是 Broadcom(博通)和 Google 自研的 TPU 团队。但有趣的是,Broadcom 的客户集中度极高,而 Marvell 正在通过差异化服务争取第二梯队的云厂商和大型企业客户。这是一场关于"谁是 AI 时代最佳芯片代工设计伙伴"的竞赛。

Comments

Loading comments...