Jefferies raises Nvidia stock price target to $300 on AI strength

Nvidia drew a wave of bullish analyst action after earnings, including Jefferies lifting its price target to $300 from $275 and citing a $20 billion CPU/server revenue opportunity in fiscal 2027. The company also raised inventory purchase commitments and prepaids to $145 billion, authorized an additional $80 billion in buybacks, and increased its quarterly dividend to $0.25 per share. The upbeat Nvidia readthrough helped lift Asia stocks, while Samsung-related news drove the KOSPI up 8%.

Analysis

The immediate read-through is not just “NVDA beats, Asia rallies” but that memory and rack-level content intensity is becoming the real earnings multiplier. If hyperscale buildouts are shifting from chip-only procurement to full-stack rack consumption, then the beneficiaries broaden from semis into HBM, advanced packaging, substrates, power management, and network gear; the scarce input is no longer GPUs alone, it is the integrated bill of materials. That argues for a second-order re-rating of suppliers with low current share in AI spend but high leverage to per-node content expansion.

The board repurchase authorization is more important as a signal than as a near-term EPS driver: it tells the market management is comfortable monetizing visibility into 2026-27 demand, which should dampen fears of a post-ramp air pocket. The main near-term reversal risk is not demand, but digestion — large commitments can pull forward revenue optics, then create a quarter or two of uneven order cadence as customers re-time deployments around product transitions. That makes the tape vulnerable to any hint of Blackwell-to-Rubin slippage or memory allocation bottlenecks.

The KOSPI reaction suggests the market is finally pricing supply-chain spillovers, and Samsung is likely the first domestic winner if DRAM/HBM pricing tightens alongside AI rack expansion. The underappreciated loser is anyone selling generic server components or legacy networking into enterprise IT: hyperscale capex concentration means budget share is being reallocated toward AI-native infrastructure, leaving less room for slower-growth refresh cycles. Consensus is still too focused on unit demand for accelerators and not enough on margin capture across the full rack ecosystem; that’s where the upside surprise should persist for 2-3 quarters.

Key Decisions for Investors

• Long NVDA into any post-earnings consolidation, using 30-60 day pullbacks to add; risk/reward favors a higher floor as buybacks + visibility reduce downside, but trim if commentary suggests shipment pacing into 2H is being pulled forward rather than expanding organically.
• Pair trade: long SK Hynix or Samsung exposure versus short a basket of legacy server/OEM names over 1-3 months; the thesis is that AI rack content expansion is shifting value capture upstream into memory and component bottlenecks while commoditizing standard enterprise hardware.
• Initiate a basket long in AI infrastructure suppliers (HBM, substrates, power, thermal, networking) for a 3-6 month horizon; this is a second-order trade on rack-level content intensity with better asymmetry than chasing NVDA after a strong move.
• Use NVDA upside calls instead of outright equity for the next 4-8 weeks if positioning looks crowded; defined-risk participation in the Computex/ramp narrative is preferable given event-driven volatility and the possibility of a near-term digestion phase.
• Avoid chasing broad Asia beta after the one-day gap; fade into strength in indices if semiconductor leadership narrows, because the rally is likely to mean-revert unless downstream suppliers confirm order acceleration over the next earnings cycle.