Cathie Wood’s ARK buys Broadcom stock, sells AMD

Nvidia Q2 revenue was cited as more than doubling, reinforcing the broader AI growth narrative, while Cathie Wood’s ARK ETFs showed notable Aug 26 trades. ARK bought Broadcom worth $20.6M (57,705 shares across ETFs) and added Cerebras Systems ($12.8M) and Cloudflare ($13.3M), but trimmed AMD by $18.2M. Net-net, the activity signals active re-positioning within high-growth tech (including a buy of Generate Biomedicines for $6.2M and sales across Tempus AI, Twist Bioscience, and others) with likely modest single-stock impact rather than market-wide effect.

Analysis

ARK’s tape is less a fundamental signal than a positioning tell: capital is being moved from higher-duration “story” names into AI infrastructure with clearer monetization. That favors AVGO and, to a lesser extent, NET/CBRS, because the next leg of AI spend is likely to accrue to networking, custom silicon, and traffic-management layers rather than only headline GPU demand. AMD is the clearest relative loser if the market starts demanding evidence of share gains instead of just sector growth; its multiple is more vulnerable to any slip in launch timing or hyperscaler adoption.

NVDA’s growth confirmation matters more for ecosystem breadth than for the stock’s immediate upside. Over the next 1-3 months, the main risk is not demand collapse but digestion: if hyperscaler capex flattens, lead times normalize, or power/interconnect bottlenecks force slower deployment, the market can re-rate from “accelerating” to “sustaining.” That would pressure the broad AI complex even if revenue remains strong.

The contrarian view is that consensus may be underweight the beneficiaries one layer down the stack: switch silicon, optics, power delivery, and network security. Those are the areas where incremental AI dollars often show up with less valuation fragility than the core GPU names. By contrast, selling TEM/TWST/HOOD on ARK flow alone is weak evidence; those names need earnings confirmation, not ETF activity, to justify a bearish stance.

Key Decisions for Investors

• Long AVGO / short AMD for 1-3 months: AVGO has better AI monetization breadth and lower execution risk; target 15-20% relative outperformance, invalidated if AMD shows sustained share gains or raises AI revenue guidance.
• Add to NVDA only on 3-5% pullbacks, or use bull call spreads to limit multiple compression risk over 3-6 months; do not chase strength after a euphoric open.
• Small tactical long NET on any post-earnings confirmation of AI traffic and security attach rates; otherwise keep it on watch, as the stock needs operating leverage to justify upside.
• Do not short TEM/HOOD/CRWD solely on ARK selling; wait for independent confirmation via guide cuts, retention deterioration, or margin compression before taking a bearish position.