Berkshire Hathaway stock is a haven again as investors flee AI bloodbath while its own portfolio takes a hit amid reshuffling

Berkshire Hathaway Class B shares are up 13.6% year to date versus 7.8% for the S&P 500, and have gained 14% since Oct. 1 as investors rotate out of AI-linked tech stocks. Third-quarter investment gains fell 46% year over year to $9.4 billion, while Berkshire repurchased $9.6 billion of stock in the quarter and about $2 billion more in the first four weeks of Q4. Railroad, energy and manufacturing profits were strong, insurance underwriting weakened, and CEO Greg Abel agreed in July to buy Taylor Morrison for $6.8 billion; the article says Democratic gubernatorial wins in data-center hot spots are expected to slow AI buildout and the overall economy.

Analysis

Berkshire is a better relative hedge against an AI-led growth reversal than a true macro hedge. Its operating businesses can benefit from grid and infrastructure investment, but also carry exposure to a broad slowdown if data-center spending or related construction is curtailed. A local political backlash is a risk to project timing and permitting, not yet evidence of a nationwide capex reversal; watch actual utility orders, project cancellations and capital-spending guidance before pricing that downside.

Near term, the rally and faster repurchases may attract defensive and buyback-oriented flows, but chasing after a sharp relative move leaves less cushion if tech rebounds. Repurchases are supportive only if Abel is buying below intrinsic value; they are not a valuation floor. Over 1–3 months, the key read-through is operating earnings: a decline in insurance underwriting or weakness in railroad, energy and manufacturing profits would undermine the “steady compounder” case. Over 6–18 months, succession execution and capital allocation matter more than quarterly investment gains, which can swing with equity marks and obscure operating performance.

Contrarian point: Berkshire’s traditional businesses are not insulated from the AI cycle—their upside partly depends on the same infrastructure buildout now facing political and permitting risk. Conversely, if that buildout slows, the market may overstate the damage to Berkshire because its earnings base is broader. With no valuation or cash-balance detail here, the evidence supports a relative-value watch, not an unconditional outright long.

Key Decisions for Investors

• Do not chase the recent outperformance. Consider a staged, beta-aware relative long in Berkshire against an AI-heavy growth basket on a pullback; the thesis is defensive relative resilience, not immunity to a growth slowdown. Reduce or exit if Berkshire’s operating earnings weaken across multiple segments while the relative-strength trend rolls over.
• Treat repurchases as a conditional catalyst, not a buy signal. Verify the next filing’s repurchase pace, share count change and cash balance; a sharp slowdown in buybacks despite continued share-price weakness would weaken the capital-return support thesis.
• Over the next 1–3 months, monitor insurance underwriting alongside railroad, energy and manufacturing operating results. Broad-based deterioration would falsify the claim that diversified operations can offset an AI-capex slowdown; investment-gain volatility alone is a weaker signal.
• Keep data-center-policy risk on alert rather than trading it as a confirmed demand shock. Look for measurable project delays, utility-order cancellations or downward revisions to grid and infrastructure spending; absent those, avoid extrapolating local political outcomes into a national Berkshire earnings hit.