
又一家值得关注的早期智能体管理初创公司:Mintive。之所以值得关注,是因为其创始人的背景,以及它已经吸引到的众多知名投资方。
Here’s another early-stage, agent-management startup to watch because of the backgrounds of the founders and the sheer number of big names it has already attracted as investors: Mintive.
Mintive 累计融资 1100 万美元,其中包括最新宣布的 700 万美元种子轮,由 Cowboy Ventures 领投。其他投资方包括 Coatue、Operator Collective、Hustle Fund,天使投资人包括 Okta 联合创始人弗雷德里克·克斯滕、Discovery Loop 联合创始人、前谷歌首席科学家杰夫·迪恩,以及 OpenAI 联合创始人安德烈·卡帕西等。
Mintive has raised $11 million total, including a newly announced $7 million seed led by Cowboy Ventures. Additional investors include Coatue, Operator Collective, Hustle Fund, and angels include Okta co-founder Frederic Kerrest, Discovery Loop co-founder and former Google Chief Scientist Jeff Dean, and OpenAI co-founder Andrej Karpathy, among others.
Mintive 由梁继全(Jiquan Ngiam,上图最右)和维贾伊·瓦苏德万(Vijay Vasudevan,上图最左)创立。梁继全是 Mintive 的首席执行官,曾是 Coursera 的创始工程师(1 号员工,由吴恩达聘用)。梁继全后来在谷歌大脑从事神经网络研究。瓦苏德万是公司首席技术官,最为人熟知的身份是他在谷歌大脑期间参与创建了 TensorFlow。那是早于 PyTorch 的元老级 AI 训练平台。
Mintive was founded by Jiquan Ngiam (pictured above, far right) and Vijay Vasudevan (pictured above far left). Ngiam, Mintive’s CEO, was the founding engineer of Coursera (employee No. 1, hired by Andrew Ng). Ngiam later worked on neural networks at Google Brain. Vasudevan, the company’s CTO, is best known as one of the co-creators of TensorFlow while he was at Google Brain. That was the OG AI training platform that predated PyTorch.
Mintive 正进入一个拥挤的领域,许多产品都承诺为企业 AI 智能体提供治理、配置和身份管理。其平台包括一个 MCP 网关,让 IT 人员可以授权智能体能够访问哪些云服务,以及可以对它们执行哪些操作。它还提供一个 LLM 网关,可将请求路由到不同模型并限制成本;处理可能与其所服务的员工不同的智能体身份和授权;以及一个智能体监控器,追踪公司内所有智能体。该公司还允许企业创建能够主动执行长时间运行任务的智能体。
Mintive is joining a crowded field of products that promise to govern, provision, and offer identity management to enterprise AI agents. Its platform includes an MCP gateway that allows IT professionals to authorize what the cloud services agents may access and what actions they can perform on them. It also offers an LLM gateway that routes to different models and can cap costs; handles agent identities and authorizations that may differ from the employees that they serve; and an agent monitor that tracks all agents across the company. The company also allows enterprises to create agents that proactively perform long-running tasks.
尽管如此,很难想象还有哪个领域竞争比这更激烈。亚马逊、微软、OpenAI、Anthropic 都提供智能体管理系统,此外还有大量单点产品(来自 Langchain、New Relic、Rippling 等各类公司,仅举几例)。还有无数初创公司正在研究这一想法(仅 Y Combinator 就有 40 多家)。
Still, it’s hard to imagine a field with more competition. Agent management systems are offered by Amazon, Microsoft, OpenAI, Anthropic, plus as point products galore (from companies as diverse as Langchain, New Relic, Rippling, to name a few). Then there are countless startups working on the idea (40+ from Y Combinator alone).
那么,Mintive的秘诀是什么?Ngiam表示,鉴于创始人在为企业构建AI方面拥有悠久历史,他们设计的系统特别易于使用。这家初创公司承诺,设置过程只需点击几个按钮即可完成。
So, what’s Mintive’s secret sauce? Given the founders’ long history in building AI for enterprises, they have designed their system to be especially easy to use, Ngiam says. Set-up amounts to clicking a few buttons, the startup promises.
“我认为我们的客户看到后感到有些震惊的是,部署速度之快、操作之简便,”Ngiam对TechCrunch表示。“我们确实已经想透了所有可以实现的方式,所有的访问控制、权限层,让他们能够如此轻松地部署,所有复杂性都被隐藏起来。你不需要考虑沙箱,不需要考虑测试框架,不需要考虑记忆,因为我们会为你管理好这些。”Mintive表示,自该初创公司2月份推出产品以来,这一卖点已经赢得了50多家企业客户,其中包括Ro、Modern Treasury、密歇根大学、Coursera、Apax和Stability AI。
“I think what our customers look at and they are a bit shocked at, is how quick it is and easy to deploy,” Ngiam tells TechCrunch. “We have really thought through all the ways this can be done, all the access control, permissioning layer, to make it so easy for them to deploy it, that all the complexities are hidden away. You don’t think about sandboxes, you don’t think about harnesses, you don’t think about memory because we manage that for you.”That pitch has won over 50 enterprise customers so far, Mintive says, since the startup launched its product in February. These include Ro, Modern Treasury, the University of Michigan, Coursera, Apax, and Stability AI.
但真正打动Cowboy Ventures管理合伙人Aileen Lee的,是创始人的行业经验。
But it was the founders’ industry experience that won over Cowboy Ventures managing partner Aileen Lee.
“AI将能够驱动不可思议的事情,但我们需要更多的基础设施来为我们提供治理、可见性和控制力,”Lee对TechCrunch表示。
“AI is going to be able to power incredible things, but we need more infrastructure that’s going to give us governance and visibility and control,” Lee told TechCrunch.
Lee还表示,她尤其尊重Ngiam的谨慎态度——他只融资了1100万美元。相比之下,其他构建智能体的初创公司在独角兽估值下筹集了巨额资金。仅举一例:Prime Intellect在10亿美元估值下完成了1.3亿美元的A轮融资。
Lee also says she particularly respects that Ngiam has been careful by only raising $11 million. This compares to other agentic-building startups raising vast amounts at unicorn valuations. As just one example: Prime Intellect raised $130 million for its Series A at $1 billion valuation.
“我真的觉得很幸运能赢得与他合作的机会,”Lee说。“他是个成熟的人,所以他不是那种追求‘虚荣’轮融资的人。”
“I feel really lucky to have won the chance to work with him,” said Lee. “He is a grown-up and so he was not a person chasing a ‘vanity’ round.”
Lee补充道:“我认为很多公司因为追求融资规模而注定失败,而不是思考‘我到底需要多少资金才能真正推动公司实现突破?’所以,是的,在人们筹集前所未有金额的资金的时候,我们只是在用一砖一瓦的、老派的方式做事。”
Lee added: “I think a lot of companies are being set up for failure by chasing round size, rather than, ‘how much do I need to actually really move the needle for the company?’ And so, yes, in a time when people are raising unprecedented amounts of money, we’re just kind of doing it the brick by brick, old-fashioned way.”