Analysis of Institutional Q2 13F Holdings: AI Investment Enters the 'Survival of the Fittest' Stage, SpaceX and Real Assets Draw Attention

Analysis of Institutional Q2 13F Holdings: AI Investment Enters the "Survival of the Fittest" Phase, SpaceX and Real Assets Draw Attention

The Q2 13F filing season for US institutional investors is nearing completion, and the data reveals a trend of capital reallocation centered on artificial intelligence, space, and growth technology stocks. Although institutions have not exited the AI track, the days of "unconditionally buying" tech stocks are over. A Reuters analysis of more than 6,300 13F filings shows that approximately 44% of institutional investors trimmed their "Magnificent 7" positions in Q2, while about 42% increased or initiated new positions, presenting a broadly balanced picture overall. At the same time, the semiconductor sector continued to see net buying, with the proportion of institutions making net purchases higher than those making net sales. This signals that capital is shifting from broad, concentrated bets on AI toward the "winners" that can truly deliver earnings, cash flow, and long-term growth. In the latest rebalancing moves, the activities of Berkshire Hathaway, Baillie Gifford, Bridgewater Associates, and ARK deserve particular attention. Although these four institutions have vastly different styles, three common keywords can be distilled: "AI giants remain favored," "SpaceX is rapidly emerging as a new core of growth capital," and "capital is spilling over into physical infrastructure such as power and energy."

Market Background

Currently, the global interest rate cycle is gradually retreating from elevated levels, inflationary pressures are easing, but economic growth prospects remain uncertain. In this environment, institutional investors need to reassess the pricing logic of growth assets. Over the past few years, the AI theme has been a major driver of equity markets, with technology leaders represented by the "Magnificent 7" contributing the majority of index gains. However, as the interest rate environment changes and base effects emerge, the market's focus on earnings quality and valuation reasonableness has risen noticeably. The Q2 13F data unfolds precisely against this macroeconomic backdrop, reflecting the transition of institutions from "buying Beta" to "selecting Alpha."

Current Capital Flows

The 13F filings show that institutions' attitudes toward tech stocks became more divided in Q2. About 44% of institutions trimmed their "Magnificent 7" holdings, but 42% chose to increase or initiate positions, indicating not a unanimous bearish view but rather a structural rebalancing. Semiconductors, as the core of AI's underlying computing power, remain among the sectors with the highest proportion of institutional net buying. In addition, a noteworthy phenomenon is that some growth-oriented institutions have begun to turn their attention to primary market targets such as SpaceX. Although 13F filings only cover publicly traded securities, institutions have expressed strong interest in space economy infrastructure, satellite internet, and launch services through channels such as venture capital and private equity. At the same time, capital has also clearly flowed into physical infrastructure areas such as power and energy, in response to the energy demand brought by the expansion of AI computing power. These industries are regarded as the "shovel sellers" of the AI era.

Investment Logic AnalysisWhy are institutions simultaneously reducing and increasing their positions in AI? The core logic lies in the fact that AI is transitioning from a "concept-driven" phase to an "earnings-driven" phase. As technology penetration enters a plateau period, the market is no longer willing to pay a premium for every loosely related concept; instead, it requires companies to demonstrate a clear commercialization path, order visibility, and free cash flow. This "survival of the fittest" elimination round is an inevitable stage of the technology cycle. SpaceX has become the new focus of growth capital because it represents another technological curve beyond artificial intelligence—the space economy—whose potential market size and strategic significance are no less than those of AI. Meanwhile, the favor shown to power and energy infrastructure is the "spillover effect" of AI capital expenditure. The energy consumption required to train large models and run data centers is growing exponentially, generating a long-term demand for stable, low-cost electricity. This structural force has brought traditional utility sectors back into the core allocation view of institutions.

In terms of structural factors, intensifying global geopolitical competition and the emphasis placed by various countries on technological autonomy and supply chain security have also driven capital to concentrate on critical infrastructure and strategic emerging industries. When constructing investment portfolios, institutional investors are increasingly placing equal weight on "safety" and "growth"—a shift that will profoundly affect asset allocation directions in the coming years.

Risk Factors

Despite the clarity of the above trends, institutions also face multiple risks. First is valuation risk: after substantial gains in AI-related stocks, some targets have already priced in several years of future growth expectations. Once earnings fall short of expectations, the pullback risk cannot be ignored. Second is policy risk: the continued tightening of AI regulations, cross-border data flows, and export controls by various countries may affect the global business layouts of tech giants. Third is geopolitical risk: trade policies after the U.S. election and supply chain restructuring could all impact the technology industry. In addition, private targets such as SpaceX carry relatively high liquidity risk; their valuations depend on primary market pricing and lack continuous price discovery in secondary markets, so institutions need to fully consider the investment horizon when allocating. Finally, physical infrastructure investments are sensitive to interest rate changes. If global rates resume an upward trend, the relative attractiveness of these assets will be compressed.

Long-Term Outlook

Looking ahead three to ten years, AI investment will no longer be a rising tide that lifts all boats, but will move toward deep differentiation. Companies that truly possess moats in data, computing power, application scenarios, and ecosystem barriers will continue to attract capital, while concept companies lacking substantial progress will be marginalized. At the same time, the space economy is expected to become another long-term growth pole after AI. As launch costs decline and satellite internet commercialization materializes, the capital value of the space industry chain will be gradually unlocked. Power and energy infrastructure will also undergo a structural revaluation driven by the demands of the digital economy and AI computing power, becoming an indispensable "ballast stone" in institutional portfolios.For institutional investors, the core challenge ahead lies in how to establish a dynamic balance among traditional assets, AI assets, the space economy, and physical assets. This requires investment teams to cross the boundaries between primary and secondary markets, deeply understand the industrial logic of technological evolution, and possess the resilience to cope with macroeconomic fluctuations. It is foreseeable that the landscape of global capital allocation is being redrawn, and the Q2 13F filings are merely an early footnote to this long-term trend.