Q1 was an active quarter for us. We deployed $1.75m across two new seed investments: a team based in Greater China building the agentic operating system for the one-person company economy and a team based in the US building AI infrastructure for embedded systems. This note summarizes what we observed in the market and how it is shaping our activity.

Q1 Activity

We spent time with founders and allocators across the Bay Area, NYC, Boston, Austin, Tokyo, Hong Kong, Beijing, and Shenzhen, and continued deploying across our core areas of focus. Across our conversations, a consistent message emerged from many funds: they are not deploying at pace.

The reasons are not uniform, but they point in the same direction:

• Public market volatility has made exit timelines harder to underwrite.
• There is uncertainty around how AI will reshape software margins and where durable value will sit across the stack.
• Hardware, robotics, and physical AI introduce longer feedback cycles and unclear capital intensity.
• Internally, many firms are also managing pacing, reserve allocation, and DPI pressure from prior vintages.

Each factor is rational on its own, but together they have slowed decision-making.

Market Observations

What we observed was a split in the market. At the early stage, fewer processes are competitive and rounds are taking longer to come together. Founders are often engaging with a smaller set of investors, and pricing is more dependent on specific conviction than on broad momentum. In parallel, capital has concentrated into a narrow set of categories. AI infrastructure, agent-driven systems, and physical compute continue to absorb a disproportionate share of funding, while most other sectors see limited activity. Large funds are still writing large checks in these areas, but outside of them, many are effectively in observation mode.

This combination changes the structure of opportunity. When large pools of capital slow down, access improves for investors who are willing to engage early and underwrite independently. Fewer competitive processes mean a higher likelihood of building direct relationships with founders and shaping rounds rather than reacting to them. Entry prices become less anchored to recent comps and more to the specific risk and progress of each company. This environment rewards speed and clarity of thought, particularly at the seed and Series A stages.

The current environment allows for earlier access, less competitive processes, and more variability in pricing. Historically, these conditions have led to stronger entry points for early-stage investors, particularly when capital is concentrated and decision-making slows at larger firms.

Where We’re Spending Time

Our approach this year has been to remain active within digital assets, where the dislocation is more pronounced. The focus has been on market infrastructure, agent-driven financial systems, and applications that abstract complexity for end users. On the infrastructure side, we are spending time on primitives that improve how value moves onchain, including settlement, liquidity formation, and the coordination of capital across fragmented environments. We are also seeing early signs of agent participation in financial activity, which introduces new requirements around execution, pricing, and risk, effectively forming early agent economies onchain. At the application layer, the most relevant teams are building products that package these systems into something usable, whether for traders, developers, or mainstream users.

Alongside this, we are tracking adjacent shifts in AI infrastructure and systems that bridge software with the physical world. These areas increasingly intersect with digital assets, particularly as compute, coordination, and capital allocation become more programmatic. The common thread across both digital assets and these adjacent systems is exposure to real usage and capital flow, rather than isolated technical progress.

The founders operating in these areas tend to share a consistent profile. They iterate quickly, often with incomplete information, and are comfortable building in environments where there is no established playbook. Internally, we refer to this group as “Squirrley Operators.” The term is informal, but it captures a real pattern. These teams are building infrastructure that is used continuously, often by other systems, and their success is measured in performance rather than visibility.

In Q1, we continued deploying across these categories with a global sourcing approach. Being physically present across multiple ecosystems has been important for both access and filtering. Many of the most relevant teams are not running broad fundraising processes, and early engagement has been necessary to build conviction. Our check sizes and stage focus remain consistent with our strategy, but the environment has allowed for more direct access and, in some cases, more favorable entry points than would have been available in a more competitive market.

Outlook

Periods like this have historically favored investors who are willing to act while uncertainty is still high. The advantage does not come from predicting macro outcomes with precision, but from identifying where real work is happening and underwriting it before it becomes consensus. That is the position we are taking as we continue deploying through 2026.

We are continuing to deploy within these areas with a focus on early-stage opportunities where we have direct access and clear conviction. The market remains selective, but we are seeing consistent signals in specific segments, and we plan to stay active as the year progresses.

If you’re looking for an active partner operating in the above markets, please reach out – we are actively deploying.

来自一线的一手观察：数字资产、AI 与全球市场

第一季度对我们来说是忙碌的一季。我们完成了两笔新的种子轮投资，合计部署资金 175 万美元：一家团队位于大中华区，正在为"一人公司"经济打造代理型操作系统；另一家团队位于美国，正在为嵌入式系统构建 AI 基础设施。本文总结了我们在市场上的观察，以及这些观察如何指导我们的行动。

第一季度动态

我们与湾区、纽约、波士顿、奥斯汀、东京、香港、北京和深圳的创始人与出资人进行了交流，并持续在我们聚焦的核心领域进行部署。在与众多基金的对话中，一个共同的信息浮现出来：许多基金并没有在加速布局。

背后的原因不尽相同，但指向了同一个大方向：

• 公开市场的波动使得退出时间表更难评估；
• AI 将如何重塑软件利润率、以及在整个技术栈中持久的价值将落在何处，仍存在不确定性；
• 硬件、机器人及物理 AI 带来了更长的反馈周期，资本密集度尚不明朗；
• 许多机构内部也在管理投资节奏、储备金分配以及来自过往年份的 DPI 压力。

每一个因素单独来看都合乎逻辑，但叠加在一起，却拖慢了决策的速度。

市场观察

我们观察到的是市场的分化。在早期阶段，竞争性流程减少，融资轮次的推进变得更为缓慢。创始人往往只与少数投资者接触，定价更多地取决于特定信念而非广泛的市场势头。与此同时，资本集中涌入少数几个类别：AI 基础设施、代理驱动系统以及物理计算继续吸收着不成比例的融资份额，而大多数其他领域的活动则相当有限。大型基金仍在这些领域开出大额支票，但在此之外，许多基金实际上处于观望模式。

这种组合改变了机会的结构。当大量资本放缓时，愿意早期介入并独立做出判断的投资者，反而获得了更好的进入机会。竞争性流程减少，意味着更有可能与创始人建立直接关系、共同塑造融资轮次，而非被动回应。入场的价格也不再锚定于最近的同类交易，而更多地取决于每家公司的特定风险与实际进展。这种环境奖励的是思路清晰、行动迅速的人——尤其是在种子轮和 A 轮阶段。

当前的环境允许更早期的接触、更少的竞争压力、以及更多样的定价空间。从历史经验看，这样的条件往往为早期投资者带来更好的入场点位——尤其是当资本高度集中、大型机构决策放缓的时候。

我们正在花时间深耕的领域

今年我们的策略是在数字资产领域保持活跃——那里的错位更为明显。我们关注的重点是市场基础设施、代理驱动的金融系统，以及为终端用户抽象复杂性的应用层。在基础设施一侧，我们花时间研究那些能改善链上价值流转的底层元件，包括结算、流动性形成，以及在碎片化环境中的资本协同。我们也看到了代理开始参与金融活动的早期迹象，这带来了围绕执行、定价和风险的新要求——实际上正在链上形成早期的代理经济体。在应用层，最值得关注的团队正在构建产品，将这些系统封装为可用的形态，无论是面向交易者、开发者还是主流用户。

与此同时，我们也在追踪 AI 基础设施以及连接软件与物理世界的相关系统的邻近变化。这些领域与数字资产的交集日益增多——尤其是当算力、协同和资本配置变得越来越可编程时。贯穿数字资产及这些邻近系统的共同主线，是对真实使用场景和资金流动的触及，而非孤立的技术进步。

活跃在这些领域的创始人往往具备相似的特质：他们迭代速度极快，常常在信息不完整的情况下行动，并且乐于在没有现成剧本的环境中构建。我们内部称这类创始人为"松鼠型实干家"。这个说法虽不正式，却捕捉到了一种真实的模式。这些团队构建的是被持续使用的基础设施——往往是被其他系统所使用，他们的成功更多体现在运行表现而非市场声量上。

第一季度，我们继续以全球化的触角在这些类别中进行部署。在多个生态系统中保持实地的物理存在，对于获取接触机会和进行有效筛选都至关重要。许多最具潜力的团队并未启动大规模的融资流程，早期的接触对于建立信念必不可少。我们的单笔投资规模与阶段重点始终与策略保持一致，但当前的环境提供了更直接的接触通道，并且在某些案例中，入场点位比在一个竞争更激烈的市场中要更为有利。

展望

历史上，这类时期往往更有利于那些在不确定性仍高时便愿意行动的投资人。优势并不来自精确预测宏观结果，而在于识别出真实的工作正在哪里发生，并在它成为市场共识之前就做出判断。这正是我们在 2026 年持续部署时所采取的立场。

我们将继续在这些领域进行布局，重点关注我们有直接接触渠道且信念明确的早期机会。市场仍将保持选择性，但我们在特定细分领域看到了持续的信号，并计划随着年度的推进保持活跃。

如果您正在上述市场中寻找一位真正活跃的合作伙伴，欢迎与我们联系——我们正在积极部署。