Morgan Stanley 13F Portfolio

The biggest incremental capital went to the same place the narrative has moved: AI infrastructure. NVDA, already the largest single name at 3.77% of the book, still saw a fresh $2.69B add, even with the position sitting roughly +475.4% above the fund’s average cost. That is not averaging down; that is pressing a winner the manager believes is structurally under‑owned relative to its role in the AI economy.

The more interesting story is how aggressively Morgan Stanley is rounding out the rest of the hardware stack:

• Micron (MU) was boosted +9.7%, adding about $1.88B, a clear call that high‑bandwidth memory is a long‑duration bottleneck, not a cyclical swing factor.
• Applied Materials (AMAT) jumped +14.2% in shares and roughly $1.21B in value, a decisive tilt toward the tools side of the AI capex boom.
• Intel (INTC) was lifted +13.9%, with about $1.05B more at work despite already being roughly +181.7% above cost, suggesting a belief in its catch‑up roadmap and sovereign/US foundry optionality.
• Broadcom (AVGO) and Microsoft (MSFT) each absorbed around $1.2B and $1.04B respectively, reinforcing a thesis around networking, custom accelerators and cloud AI distribution, not just GPUs.
• Amazon (AMZN) saw a further $1.30B added, tying the hardware bet back to hyperscaler demand and commerce‑plus‑cloud earnings power.

Even in the ETF book, the adds are selective rather than blanket. IVV grew +5.5% (roughly +$965.3M), while mega‑growth vehicles like VUG and style funds like VTV were nudged up. That says the manager is comfortable with US large‑cap growth as a factor, but prefers to customize sector risk on top via hand‑picked AI winners.

Outside tech, Eli Lilly (LLY), Merck (MRK), AbbVie (ABBV), and large banks like Bank of America (BAC) and Goldman Sachs (GS) all saw incremental adds. These moves look like purposeful ballast: durable cash-flow machines to keep the portfolio’s risk budget stable while the AI book gets more aggressive.

The primary funding source this quarter is clearly the passive book. SPY is the standout, with shares cut -20.3% and an estimated -$6.14B coming out. RSP (equal‑weight S&P) and IWD (large-cap value) also shrank, down -4.8% and -3.7% in shares, respectively, with a combined roughly -$731M released. The message: if you want more AI infrastructure, you do not get it from equal-weight value.

The most notable single‑stock reduction in tech is Palo Alto Networks (PANW). Shares fell -8.9%, freeing up about -$978.3M despite the position still sitting roughly +344.7% vs cost. That looks less like a thesis break and more like disciplined profit‑harvesting to recycle into higher‑conviction semis.

Large, mature growth platforms that have rallied hard were dialed back at the margin. Apple (AAPL) was lightly reduced (-0.8% in shares, about -$538.3M), and one of the Alphabet lines (GOOGL) was trimmed -0.4%, even as the other (GOOG) was nudged up. This suggests sizing discipline around advertising-heavy platforms while still keeping meaningful AI exposure.

Away from tech, the fund shaved Procter & Gamble (PG) by -4.2% and Coca-Cola (KO) by -2.3% in shares. Both had relatively modest gains vs cost (around +13.9% for PG and +78.2% for KO), so using them as funding signals a willingness to give up some low‑beta staples in exchange for more cyclical, higher‑upside AI and hardware exposure.

On the surface, sector moves look incremental. Technology’s weight ticked up from 47.0% to 47.77%, but underneath that, the composition is shifting toward semis and equipment. The manager is deliberately clustering risk in the physical infrastructure of AI — memory (MU), foundry and design (NVDA, AMD, INTC, TSM), and tools (AMAT, LRCX) — not just in software and ad‑driven platforms.

Unclassified holdings, dominated by ETFs and Berkshire Hathaway, fell from 22.04% to 21.14%. That one‑point move is the macro story: less generic index exposure, more idiosyncratic stock risk where Morgan Stanley thinks it has an edge. The trims in SPY, RSP, and IWD, alongside the adds in IVV and style/growth funds, show a rotation even inside the passive sleeve toward plain-vanilla S&P and growth, away from equal‑weight and value.

Consumer Discretionary nudged up from 10.21% to 10.34% as the fund added to Amazon, Walmart, Costco, and Home Depot. This is a very specific flavor of discretionary: scale retailers and platforms with pricing power, not speculative consumer names. Health Care rose slightly from 6.4% to 6.5% on added LLY, ABBV, MRK, and JNJ, supporting the view that pharma and managed care are the ballast that makes the high‑beta tech bet tolerable.

Finance held roughly steady (4.39% to 4.37%), with modest trims in JPMorgan and small adds in BAC and Goldman Sachs. Energy, Industrials, and Consumer Staples all slipped a bit as the fund pared Chevron, Procter & Gamble, Coke, and a portion of GE Vernova, indicating less enthusiasm for classic late‑cycle defensives versus AI‑linked capital spending.

Taken together, Q2’s moves say Morgan Stanley believes we are still early in the AI capex super‑cycle. The choice to add billions to NVDA, MU, AMAT, AVGO, and INTC after large gains versus cost shows they see structural earnings power, not a fad that should be traded around the edges. The ETF cuts, especially in SPY and equal‑weight/value products, underline their view that generic beta is now less attractive than owning the specific companies building the new compute and memory stack.

At the same time, the book is not an undisciplined growth binge. Incremental capital into LLY, ABBV, MRK, JNJ, and big banks suggests an intent to keep a spine of steady cash flows and dividends behind the AI trade. Slight reductions in consumer staples and mega‑cap ad platforms show a willingness to shrink lower‑conviction safety trades and crowd names to fund higher‑octane ideas.

Looking ahead, expect more of the same pattern if the macro backdrop holds: marginal dollars flowing into semis, tools, and cloud‑tied platforms, financed by trimming passive sleeves and lower‑growth defensives. If AI hardware demand or capex guidance wobbles, the holdings in healthcare and financials are already in place as shock absorbers. For now, though, the 13F makes the call explicit: Morgan Stanley would rather live or die with the AI infrastructure build‑out than with another unit of S&P beta.