Billionaire Steve Cohen Sold Amazon and Nvidia but Loaded Up on This Beaten-Down Stock

Steve Cohen’s Point72 trimmed Amazon by 6% and cut Nvidia by 24% in Q4 2025–Q1 2026 as both stocks are up only ~5% YTD in 2026 versus ~10% for the S&P 500. At the same time, Cohen increased Boston Scientific exposure by 50%, even though the stock is down >50% YTD, highlighting a deep-value healthcare bet amid a weaker outlook: FY adjusted earnings guidance was cut to $3.34–$3.41 from $3.43–$3.49 and Q2 organic growth is guided down to as low as ~5%. The valuation backdrop (P/S ~3.2x vs ~6x avg; P/E 19x vs ~64x) could support a long-term rebound, but near-term guidance deterioration tempers sentiment.

Analysis

Point72’s trim in AMZN/NVDA is more useful as a crowding signal than as a fundamental call. When the AI complex stops leading on price, the market usually starts rewarding the bottlenecks around it: power, cooling, grid interconnects, and other enablement layers. That creates a narrower but more durable trade than owning the headline platforms alone, though names like BE still trade on execution and financing risk rather than clean secular visibility.

BSX looks like a classic post-reset value setup, but the opportunity is time-dependent. In the next 1-2 quarters, the key variable is whether procedure growth and mix normalize enough to restore confidence in the full-year guide; if not, the low multiple can stay low for months. The upside case is a reacceleration back to high-single-digit organic growth, which would force systematic and value investors to re-rate the stock, but a second guidance cut would overwhelm any cheapness argument.

The contrarian risk is that 13F lag makes this look more “smart money rotation” than “new thesis.” Cohen may simply be de-risking mega-cap AI after a long run while recycling into a beaten-up medtech compounder; that does not mean NVDA/AMZN fundamentals are deteriorating. If AI capex and cloud demand re-accelerate, shorting the leaders here is a bad trade because earnings revisions can outrun narrative rotation.

Key Decisions for Investors

• Do not short NVDA or AMZN outright on a stale 13F signal; if already long, hedge into strength with 1-2 month call spreads or collars ahead of the next earnings window. Falsifier: another upside guide from hyperscale peers or capex commentary that re-ignites AI revisions.
• Watch BE as a leveraged beneficiary of AI infrastructure bottlenecks, but prefer entry only on a pullback or after evidence of order conversion. Risk/reward is attractive only if data-center power constraints stay binding for 6-12 months; if utility interconnect timelines improve, the multiple can compress quickly.
• Initiate a tactical watchlist on BSX for a post-earnings re-entry, not a blind dip-buy. Best setup is a pair versus XLV or IHI once management shows organic growth stabilization; invalidation is another guide-down or continued deceleration below mid-single-digit growth.
• If forced into a pair, prefer long BE / short a broad AI hardware basket only as a relative-value trade, not a directional AI bear. Time horizon: 1-3 months, with stop-loss if NVDA earnings or hyperscaler capex commentary reasserts the old leaders.