Citi cuts tech stock weighting: 'Difficult to see how everyone in AI/Tech path wins'

Citi downgraded its tech weighting to market weight from overweight, citing idiosyncratic trading in the Mag 8 and rising concerns that AI-related capex returns are getting harder to justify. The note comes as chip stocks weakened, with Nvidia down about 1%, Micron off about 4%, and the VanEck Semiconductor ETF (SMH) down 3% on the day. Despite the cautious tech call, Citi still sees the S&P 500 at 8,100 year-end, implying broader equity upside and a rotation away from tech.

Analysis

The important signal is not that tech is weak, but that the AI supply chain is starting to bifurcate: beneficiaries with hard revenue today can still work while the broader ecosystem de-rates on questions about payback periods. That usually hurts the second-derivative trades first — semis, networking, power/cooling, and software names levered to future AI capex — because their multiples are most sensitive to whether incremental spending remains accretive. In that setup, NVDA can be comparatively resilient versus the more balance-sheet-sensitive and inventory-sensitive parts of the chain, but the trade is no longer “own everything AI.”

The timing matters. This is a positioning unwind risk over days to weeks, not a fundamental end-of-cycle call, and the near-term catalyst is simply that investors will use any news flow to trim crowded winners into quarter-end and early July. If tech leadership narrows, breadth can actually improve without a full market reset, which means the better expression is to reduce beta concentration rather than fight equities outright. The clearest second-order beneficiary is cyclicals and value segments that gain from lower inflation expectations and a broader risk-on tape.

The contrarian view is that the market may be overpricing the delay in AI monetization while underpricing the earnings durability of the large platforms. If capex slows, the market could briefly punish semis, but that same discipline may extend the cycle by forcing better ROI and more targeted spending. For NVDA specifically, the risk is less demand collapse than multiple compression if investors decide the next leg of AI growth will be more selective and less uniform across vendors.

Key Decisions for Investors

• Trim NVDA into strength over the next 1-2 weeks rather than exit outright; use a 5-10% position reduction to de-risk crowded AI exposure while preserving upside if the tape stabilizes.
• Short SMH vs long equal-weight S&P futures for 2-4 weeks as a relative-value expression; the risk/reward favors underperforming semis if the AI capex ROI debate stays front-and-center.
• Initiate a tactical rotation into financials via C or broader market breadth exposure for the next 1-2 months; upside improves if disinflation broadens and rate-sensitive valuation pressure eases.
• Buy short-dated NVDA downside protection, e.g. 1-2 month put spreads, to hedge a further 5-8% drawdown from positioning unwind; defined risk is preferable here versus outright shorts.
• If semis sell off another 3-5% without deterioration in earnings guidance, consider scaling back hedges and re-adding selectively to quality leaders, as the move may become technically exhausted.