Billionaire Ole Andreas Halvorsen Dumped His $39 Billion Fund's Stakes in Nvidia and Amazon for Another Trillion-Dollar Superstar

Viking Global Investors (roughly $39 billion AUM) materially reshaped its Q3 equity positioning, fully exiting 3,681,935 shares of Nvidia and 3,897,092 shares of Amazon while initiating a large Microsoft position of 2,429,412 shares (~$1.26 billion), making MSFT the fund’s fifth-largest holding at ~3.2% of invested assets. The article highlights valuation-driven motives and AI-bubble concerns—citing Nvidia’s P/S briefly above 30 and Amazon trading at a forward P/E near 34—while noting Microsoft’s strong cash profile ($102 billion cash and equivalents, $45 billion operating cash in the fiscal quarter) and 39% Azure constant-currency growth as rationale for the new stake.

Analysis

Market structure: Halvorsen’s exits from NVDA and AMZN and large MSFT buys signal rotation from AI hardware/high-P/S momentum into cash-generative platform software. Winners: MSFT, Azure-dependent SaaS vendors and legacy cash-flow compounders that can finance AI spend; losers: pure-play AI hardware (NVDA) and stretched growth names with P/S >20–30. This repricing implies short-term demand shock for semis and reallocation of institutional flows into mega-cap software over the next 1–3 quarters.

Risk assessment: Tail risks include an AI bubble unwind that knocks NVDA down >40% in a month, or regulatory/antitrust moves against MSFT that compress multiples by 20–30% over 12–24 months. Immediate (days) effects are volatility spikes and higher implied vol for NVDA/AMZN; short-term (weeks–months) depends on Q4 guidance and data-center capex; long-term (quarters–years) remains structural: cloud + AI adoption favors software margins. Hidden deps: data-center capex cadence (Intel/TSMC cycles), enterprise software spending elasticity, and cloud price competition that could compress Azure growth if AWS/Google respond.

Trade implications: Favor modest overweight in MSFT (platform exposure) and tactical underweight/hedge for NVDA exposure; implement options to manage skew rather than outright large short. Sector rotation: increase weight to Software/Cloud, reduce Semiconductors and Consumer Discretionary exposure by 2–5% each over next 30–90 days. Entry should be staggered: scale into MSFT over 4–6 weeks, harvest alpha after next two earnings beats or if forward P/E re-rates above 30.

Contrarian angles: Consensus treats NVDA as fragile and MSFT as safe — that may under-price NVDA’s durable moat (software stacks, CUDA lock-in) and over-rate MSFT’s immunity to multiple compression. Historical parallel: 2010s cloud winners gained share while chip vendors oscillated with cycles — patience in NVDA could pay if earnings/capex resumes. Unintended consequence: crowded short/put exposure into NVDA could trigger sharp squeezes; size hedges small and defined-risk rather than outright naked shorts.

Key Decisions for Investors

• Establish a 2–3% long position in MSFT over 4–6 weeks, add another 1% if Azure constant-currency growth stays >30% in next two fiscal quarters; fund by trimming 1–2% from semiconductor exposure.
• Implement a defined-risk hedge against NVDA: buy a 3-month put spread (long 10% OTM put / short 20% OTM put) sizing max loss to 0.5–1% of portfolio to protect vs a >30% drawdown event.
• Run a relative-value pair: go long MSFT and short NVDA with 1:1 dollar notional (net delta‑neutral) sized 1–2% of portfolio for a 3–9 month horizon to capture rotation from hardware to software.
• Trim AMZN exposure to underweight if forward P/E >34 or if next-quarter free-cash-flow guidance falls >10% y/y; consider redeploying proceeds into high-quality cash generators (MSFT, selected SaaS) or buy financed covered calls on MSFT (3–6 month, 5–10% OTM) to reduce cost basis.