3 Stocks Billionaires Are Piling Into Right Now -- and 1 Might Surprise You

Billionaire hedge fund 13-F filings show David Tepper added about 1 million shares of Micron, while Ken Griffin added nearly 11 million shares of Amazon and 12 million shares of Nvidia. David Einhorn also initiated a new position in recently spun-off Versant Media Group, which went public in January and is down 10% since launch but up 48% over the past month. The article is primarily a positioning roundup with no new operating results, so the likely market impact is limited.

Analysis

The signal here is less about the named stocks and more about where sophisticated capital sees asymmetry after the Q1 dislocation: AI capex beneficiaries on one side and underappreciated cash-flow durability on the other. In memory, the setup is not just a cyclical rebound; it is a supply discipline story meeting accelerating compute/storage intensity, which tends to support multiple expansion as much as earnings. If that holds, the more important second-order effect is that the market may keep rerating the whole AI hardware stack, but with the most torque in the names that still screen as “cheap” on next-twelve-month earnings.

For AMZN and NVDA, the relevant edge is not that they are good businesses, but that they have become increasingly self-funding platforms: each incremental AI dollar spent by customers reinforces ecosystem stickiness, cloud attach rates, and pricing power. That tends to compress the window for skeptics to bet on valuation mean reversion, because every quarter of execution reduces the probability that these names trade like ordinary mega-caps. The risk is that consensus is now assuming a clean second-half reacceleration; if enterprise spend pauses or hyperscaler capex growth normalizes faster than expected, the market could quickly rotate away from duration and into defensives.

The more interesting contrarian angle is the media spin-off. The market still discounts cable-adjacent assets as secular decline, but the bundle of live/news and non-PayTV revenue creates a narrower-than-assumed decay path, which can matter a lot in a newly public security where positioning is sparse and float is limited. That combination often produces sharp squeezes, but it also means the path dependency is ugly: if the first few quarters do not show stabilization, the multiple can compress quickly because there is no legacy shareholder base to defend it. CMCSA may be a quiet relative winner if the market starts to ascribe value to the asset separation rather than the consolidated drag.

Key Decisions for Investors

• Long MU on a 3-6 month horizon into the next earnings cycle; thesis is continued AI/server memory tightening with upside if ASPs hold. Use a staggered entry on pullbacks, and size for volatility because this can re-rate fast but also retrace 15-20% on any supply scare.
• Buy NVDA pullbacks vs short a basket of slower AI infrastructure laggards; the cleaner trade is relative share gain, not outright beta. Timeframe 1-2 quarters, with risk that broad AI capex sentiment de-risks if hyperscaler commentary softens.
• Initiate a small starter long in AMZN for 6-12 months, favoring call spreads over stock if volatility is elevated. Reward is multiple expansion plus continued operating leverage; risk is a capex-heavy quarter that delays margin inflection.
• Watch VSNT as a tactical event-driven long only if borrow is tight and liquidity remains thin; best expressed via options or a small cash position. The setup is attractive for a 1-3 month squeeze, but exit quickly if post-spin fundamentals fail to show stabilization.
• Avoid shorting CMCSA outright; if bearish, express it as a hedge against long media/platform exposure rather than a standalone short. The risk/reward is poor if the market starts valuing separation optionality and cash generation independently.