在二月的报导中，我提到风险投资公司正将人工智慧人才纳入内部，以提升对这项技术的使用能力；Eclipse 将这股趋势推向新阶段。该公司是以火箭等「moonshot」技术为核心的 VC，在完成 13 亿美元融资后，聘请前 Meta 高阶主管 Amir Frenkel 担任首位首席人工智慧长。Eclipse 希望借此强化自身 AI 能力，并支援其投资组合公司将 AI 真正嵌入产品与路线图。
这次任命之所以特别，不在于一般投资后支援团队已有，而在于 Frenkel 以纯 AI 实作为核心，且背景高度技术导向。当前他不直接负责投资决策，而是主要协助 Eclipse 旗下创办人与 CTO，含晶片厂 Cerebras，进行 AI 成熟度、架构与导入规划。这不只是建议付费使用高阶 ChatGPT、Claude，或试用 Granola 之类生产力工具；Frenkel 强调 AI 只是一个术语而非解法，核心是配对需求、风险承受度与技术条件，帮助企业「rightsize」方案。
Frenkel 有二十余年科技产业经验。最近曾任 Meta 的生成式 AI 副总裁；在此之前近十年主导 Meta 的 AR 与 VR 方向，并曾担任 Google 的 AR 与可穿戴装置工程主管，参与过 Google Glass 计划。更广泛地，OpenAI 与 Anthropic 也在加码支持其他企业导入 AI，显示整个生态系从「投资兴奋」转向「如何不盲目铺陈」的执行化阶段；Frenkel 的看法是，AI 应用应聚焦解决具体问题，而非把 AI 无差别全面洒布。
In February, I wrote that venture capital firms were moving AI talent in-house to improve their own use of the technology; this is now becoming a new phase. Eclipse, a VC focused on moonshot technologies, followed that trajectory after raising $1.3 billion by hiring former Meta VP Amir Frenkel as its first-ever chief AI officer. The goal is twofold: strengthen internal AI capability and help portfolio companies integrate AI into products and roadmaps.
The hire is notable because it is not a standard founder-support operations model but a role centered on technical AI implementation. Frenkel is not focused on writing checks for now; he will advise founders and CTOs, including teams at chipmaker Cerebras, on AI setup and rollout. The work is more than suggesting upgraded subscriptions to ChatGPT or Claude, or tools such as Granola. Frenkel stresses that AI is a term, not a solution, and that success comes from right-sizing the solution to the business problem, risk appetite, and technology feasibility.
Frenkel brings over two decades in major technology firms. He was most recently Meta’s VP of generative AI after nearly a decade in the company’s AR and VR efforts, and he previously led Google’s AR and wearables engineering, including work linked to Google Glass. The shift is broader than one VC: OpenAI and Anthropic are also investing in helping other companies adopt AI. As more venture firms and operators pour capital into AI consulting, the tone in boardrooms is moving from excitement toward operational panic control: deploy AI selectively where it solves real problems, not indiscriminately.