Prediction: Following Alphabet, This Could Be the Next Trillion-Dollar Artificial Intelligence (AI) Stock Greg Abel Adds to Berkshire's Portfolio

Berkshire Hathaway under CEO Greg Abel moved deeper into AI-linked mega-cap tech, initiating/adding to Alphabet to nearly 54 million shares (more than triple in Q1 2025) and fully exiting Amazon. The article argues Nvidia is increasingly “Berkshire-like” due to its AI CUDA software moat, expanding AI-stack investments (e.g., $1B investment in Nokia for AI-native RAN, plus data-center ecosystem support), and strong profitability with shares up over 1,000% in recent years. It also notes Nvidia has expanded its dividend and that its forward P/E has become more reasonable versus the ~17x S&P 500 long-run average.

Analysis

The market should read this less as a Berkshire-specific endorsement and more as a validation of a narrow AI quality screen: cash generative, ecosystem-led, and still capable of compounding without relying on speculative adoption curves. That framing is supportive for NVDA and GOOGL, but the bigger second-order beneficiaries are the picks-and-shovels names tied to network upgrades and optical interconnects — MRVL, COHR, and LITE — because institutional “permission” often lifts the entire infra basket before fundamentals show up in consensus estimates.

The near-term risk is that this becomes a one-day sentiment trade with little follow-through unless reinforced by actual 13F accumulation or stronger capex commentary from hyperscalers. Over 1-3 months, NVDA’s real catalyst is still earnings/guidance and data-center capex, while GOOGL needs proof that AI monetization is offsetting search disruption; if either misses, the Berkshire angle fades quickly. AMZN is the relative loser on narrative, not economics: if allocators favor clearer free-cash-flow conversion and capital returns, AWS/retail’s longer-dated AI optionality looks less compelling in a factor model.

Contrarian view: consensus may be overrating the breadth of the signal. Berkshire moving toward a few large-cap tech winners does not mean it will tolerate balance-sheet-intensive, lower-margin AI adjacent names; it likely prefers monopoly-like economics over “AI exposure” as a theme. The cleaner trade is not long everything AI, but long the franchises with durable pricing power and short the names where AI spend is still a cost center rather than a monetization engine.

Key Decisions for Investors

• Buy NVDA on 2-3% pullbacks over the next 1-2 weeks; hold 1-3 months into earnings. Reward/risk is favorable if data-center growth stays >30% y/y; thesis breaks if gross margin guidance compresses or hyperscaler capex rolls over.
• Pair trade: long GOOGL / short AMZN for 1-3 months. The market is rewarding AI monetization plus buybacks over spend-heavy AI optionality; exit if AWS reaccelerates materially or search monetization weakens less than feared.
• Initiate a basket long MRVL/COHR/LITE vs. short AMD only if industry checks confirm sustained AI networking demand. This is a higher-beta expression of the same quality screen; stop if Nvidia ecosystem share shifts toward custom silicon faster than expected.
• Use BRK.B as a defensive proxy long against QQQ only if you want exposure to the 'quality AI' factor with lower multiple risk. The signal is incremental, not a standalone catalyst, so size modestly.