This monthly report covers fund performance, core holdings, and the investment manager’s review and outlook, with a focus on market divergence, AI infrastructure, personal AI assistants, and structural opportunities in memory and connectivity.

In this 2026 Mid-Autumn letter, BEDROCK Investment reflects on the long-term opportunity created by AI, its consistent approach to selecting businesses, and the use of AI to improve investment research.

When old playbooks are no longer enough to explain the present, how should we understand change? In this episode, we talk with Trader Jiu about new signals in Japan's economy, AI-era opportunities and organizational change, and the tension between technical edge, passion, and real-world returns.

Silicon Valley or Wall Street: Who Is Getting AI Wrong?

Over the past month, the portfolio's net asset value rose modestly while exposure and investment structure remained broadly unchanged. We continued to focus on AI-related investments and on companies with durable fundamentals, strong competitive barriers and reasonable valuations.

Today's market consists of a broad but still shallow majority, alongside a small leading group whose AI use is already deep.

The market should not both judge that Warsh may succeed and continue to price assets under the old Fed Put. Nor should it assume in advance that he will fail and make a major crisis the base case while he has yet to commit a clear policy error and is pursuing a steadier path. What must be priced in ahead of time is the shift in who bears tail risk, not the inevitability of a broad crisis.

It is, of course, a $1.5 billion film. More importantly, it places a 28-year capability curve before us: from a $6,000 weekend film to the ability to deliver global events consistently across studios, genres, and team configurations.

Over the past month, the portfolio experienced a notable drawdown. We continue to focus on fundamental evidence, portfolio concentration, and overall risk exposure.

How we moved from manual execution and process checks to systems that run the workflow: middle-office workflow automation, from document intake to controlled publication, so repetitive work keeps moving automatically.

In this episode, we continue our in-depth conversation with clients. We discuss how AI is reshaping the foundational layers of the world, from computing power and connectivity to healthcare and life extension. The conversation then expands into the industrial evolution of AI infrastructure, including chips, networks, and system architecture, and how to apply first-principles thinking amid this structural transformation. This episode is for research discussion and personal vie

BEDROCK Growth Monthly Report for June 2026.

Bedrock Growth Fund SPC ranked No. 1 by net return for May 2026 in BarclayHedge Equity Long-Bias.

This episode discusses how the investment framework is being rewritten in the AI era, from research workflows and industry opportunities to portfolio construction.

Global markets continued to recover in May 2026, with clear divergence across regions. We maintained high equity exposure, focusing on semiconductors, advanced manufacturing, internet, fintech, and consumer-related opportunities.

As the first half of 2026 approaches its close, this episode reviews BEDROCK’s portfolio operations and discusses how AI is reshaping investment research.

Excluding non-market factors and colors, Huawei’s τ law can be understood as the same thing as Lao Huang’s extreme co-design.

From 48V to 800V, and then to solid-state transformers, the bottlenecks in AI data centers are continuing to move from chips, networks, and liquid cooling to more upstream power electronics and energy supply.

This episode is a collaboration between BEDROCK and Trader. Topics include: Starting with DCI, the driving logic behind it; The structure and bottlenecks of AI infrastructure; The rationality and sustainability of AI capital expenditure; The diffusion of opportunities in the industry chain, technological iteration, etc.; The coexistence of opportunities for both old and new AI.

The role of DCI has changed. Once computing power is no longer crammed into a data center, demand will not only follow the number of users, but will be amplified by the number of nodes, single connection bandwidth, and topology level.