13F Disclosures Reveal Q2 Institutional Repositioning: AI Enters 'Survival of the Fittest' Phase, SpaceX and Power Infrastructure Become New Capital Destinations

13F Filings Reveal Q2 Institutional Repositioning: AI Enters "Survival of the Fittest," SpaceX and Power Infrastructure Become New Capital Destinations

As US institutional investors' 13F filings for the second quarter of 2026 are gradually disclosed, the repositioning paths of global capital across AI, space, and growth technology are coming into sharper focus. More than 6,300 filings show that institutions have not withdrawn from AI, but the era of "indiscriminate buying" has ended. Capital is now shifting toward companies that can truly deliver profits, cash flow, and long-term growth potential, with SpaceX and power infrastructure simultaneously becoming new focal points for positioning.

Market Background: Interest Rate Regime Reset and Valuation Pressure

Over the past few years, major global economies experienced a sharp burst of inflation, and major central banks raised policy rates at the fastest pace in decades. Although inflation has fallen from its peak, policy rates are likely to remain structurally higher than pre-pandemic levels. This places greater demands on equity valuation models: when the risk-free rate is no longer near zero, long-term earnings growth must be discounted more prudently. Meanwhile, the AI-led technology sector has continued to gain weight in indices, and large technology companies now wield market influence beyond that of traditional blue chips. This high concentration not only facilitates index gains but also amplifies the risk of capital crowding into a handful of stocks. The quarterly 13F disclosures happen to provide a real-world sample of how institutional investors are responding to this changed valuation and liquidity environment.

Current Capital Flows: AI Divergence Widens, Semiconductors Remain Sought After, SpaceX and Physical Infrastructure Heat Up

Reuters analyzed more than 6,300 13F holdings filings and found that in the second quarter of 2026, roughly 44% of institutional investors reduced exposure to the "Magnificent 7" (the seven major tech giants), while about 42% chose to add to or newly build positions. The two camps are almost evenly split, showing that institutional capital has diverged significantly over core AI assets.

A more structural trend is that the proportion of institutions net-buying semiconductor stocks still exceeds the proportion net-selling. This indicates that although the pace of commercialization on the AI application side is under scrutiny, demand for infrastructure such as computing power and storage has not yet been invalidated. The market appears to be shifting from "buying the AI story" to "buying certainty in the AI supply chain."

Equally noteworthy are the moves of certain representative institutions. Berkshire Hathaway, Baillie Gifford, Bridgewater Associates, and ARK have followed very different repositioning paths, yet together they reveal three themes:

• First, the major AI players remain long-term core holdings but are no longer being added to unconditionally;
• Second, SpaceX has quickly become a new core of growth capital, with private space infrastructure drawing increasing attention from long-term investors;
• Third, capital is spilling over into "physical infrastructure," especially power companies and energy infrastructure that provide energy support for AI.In particular, the expansion of SpaceX's valuation and investor list shows that global institutions' interest in frontier fields such as space traffic, satellite internet, and orbital manufacturing has extended from venture capital to cross-asset-class asset allocation. Although 13F filings do not directly cover private companies, some institutions indirectly reflect their exposure by holding funds or special vehicles that invest in SpaceX. Changes in such "peripheral holdings" often carry leading significance.

Investment Logic Analysis: From "Broad Gains" to "Selective Picking," Capital Searches for a Second Growth Curve

Why has capital flow changed in this way? Behind it lies the inevitable evolution of the AI investment lifecycle. Over the past few years, the defining feature of artificial intelligence investment has been an "arms race": large-model companies, cloud vendors, and chip companies simultaneously expanded capital expenditure, and the market paid extremely high premiums for the future. Over time, investors' focus has gradually shifted from the "model parameter contest" to "real revenue, cash flow, and competitive moats." The second phase of AI investment is more about answering one question: among the many models and products, who can become a true winner that is continuously used by enterprises and consumers?

This logic resembles the investment path following the spread of electricity in the early 20th century: the earliest capital mainly went to power stations and grid equipment, and only later gradually flowed toward electricity-driven industrial manufacturing, urban transportation, and household consumption. Today, the power demand of data centers is becoming a major branch of this logic. AI workloads have far higher density than traditional applications, and new data centers face bottlenecks such as long grid-connection lead times and difficulty securing electricity. Against this backdrop, power companies, transformer suppliers, and natural gas infrastructure have become "growth-stock-like" utilities, attracting attention from traditional value funds to macro hedge funds.

Meanwhile, the space economy brought by SpaceX represents another long-term theme. Satellite communications can directly serve millions of unconnected users worldwide, and low-orbit satellite constellations feature high capital barriers and network effects. For long-term investors, this constitutes a new "platform-level" opportunity with ecosystem evolution potential similar to that of AI and cloud computing. As reusable rockets further reduce launch costs, the space economy may gradually move from the infrastructure stage into an application boom.

It is precisely these drivers that are shifting institutional investors' asset allocation from a single growth pole toward a multipolar structure. A common portfolio approach is: on one end, allocate to high-beta AI- and space-related growth companies; on the other end, allocate to power and energy infrastructure that provides cash flow, inflation hedging, and physical scarcity. This "barbell allocation" preserves the offensive power of technological innovation while increasing portfolio resilience under macroeconomic uncertainty.

Risk Factors: Valuation, Policy, and Crowded Trades

Even though the long-term logic is clear, the risks in the current market cannot be ignored.AI-related stock valuations are at historical highs. Even after the volatility, the market capitalization of leading tech companies remains built on expectations of sustained high growth. If earnings growth slows significantly over the next few quarters, the gap between valuations and fundamentals could trigger a systemic correction. The fact that nearly half of institutions trimmed their Mag 7 positions in 13F data is itself "a rebalancing of concentration risk."

Regulatory and policy risks also exist. Around the world, governments are intensifying intervention in semiconductor export controls, AI safety, and cross-border data flows. Spectrum allocation, space debris governance, and military applications in the space domain are also facing growing international competition. Any new regulation could change the profit distribution structure of the industry.

In addition, primary-market projects such as SpaceX still face uncertainty in technical execution, while power infrastructure projects face approval-cycle and interest-rate sensitivity issues. At the macroeconomic level, the normalization of US fiscal deficits, combined with a higher interest-rate center, could compress the overall valuation level of risk assets. When too much capital gathers around a few hot themes, the rise in volatility caused by crowded trades should not be ignored.

Long-Term Outlook: Structural Trends Persist, Institutional Allocations Diversify

Extending the horizon to the next 3-10 years, we have reason to believe that AI and the space economy remain the most important long-term investment narratives, but the investment toolkit will become more diverse: shifting from passive beta to selective alpha, and expanding from secondary-market equities to private equity, real assets, and thematic ETFs.

In the Long-Term Outlook, global institutions will face a "multipolar growth" world: US AI and technology innovation, China's new-energy industrial chain, European defense and infrastructure reconstruction, and demographic and digital dividends in emerging markets will all have a place in portfolios. This requires institutional investors to have more systematic macro-trend tracking and more flexible asset allocation capabilities.

At the same time, we expect pension funds, sovereign wealth funds, and family offices to continue increasing their allocations to alternative investments in order to capture the illiquidity premium. The development of ETF vehicles, meanwhile, allows individual investors to share in the long-term growth of global innovation themes at lower cost.

In essence, 13F disclosures are only the "footprints" of institutional asset managers over a single quarter. What is truly worth studying is the logic behind their portfolio diversification adjustments based on long-term risk-return considerations. In an uncertain geopolitical and interest-rate environment, maintaining a panoramic view of global markets and understanding the interplay between capital flows and macroeconomic policy will be key to long-horizon managers staying competitive.

---

Source: This article was written based on reference content published in the moomoo community. See the original article.