The First Half of 2026 Is Over. These 2 Spectacular Artificial Intelligence (AI) Stocks Can Soar in the Second Half.

Nvidia is framed as undervalued at 21.7x forward earnings (vs ~34x over the prior two years), with analysts projecting 96% YoY revenue growth next quarter and a potential rerating in the latter half of 2026 as the growth narrative extends beyond 2027. Meta is down ~12% YTD as markets focus less on Q1 revenue (+33% YoY) and more on elevated AI infrastructure capex, though it recently began selling excess capacity externally and the stock is described as cheap at 19.5x forward earnings. Overall, the article suggests AI-chip demand strength for Nvidia but lingering capex/ROI uncertainty for Meta.

Analysis

The setup is less about a new AI cycle and more about where the incremental economics accrue. If NVDA’s demand really extends into 2027, the market is underpricing the durability of the earnings stream, but the cleaner upside is in the supply-chain choke points: memory and packaging names like MU and SNDK should keep pricing power longer than GPU multiples stay elevated. AMD and INTC are the relative losers if buyers continue paying up for proven throughput rather than optionality; their risk is not just share, but a higher bar for margins as they spend to stay relevant.

META is a different mechanism: the stock is being punished for capital intensity, yet any credible external monetization of excess compute would improve asset turns and reduce the market’s “wasted capex” discount. The important second-order effect is that a successful rent-out model would also soften the bear case across hyperscalers by reframing AI data-center buildout as revenue-generating infrastructure, not pure expense. The key missing data is utilization and contract duration; without those, the market should treat the announcement as a signaling event rather than a real earnings inflection.

Contrarian view: consensus may be too focused on 2027 demand narratives and not enough on 1-2 quarter digestion risk. If order timing slips, the longest-duration names can de-rate faster than their fundamentals deteriorate, especially after a strong run in the AI supply chain. The thesis is falsified if NVDA guides below the implied 90%+ growth path or if META’s capex keeps rising without a measurable revenue offset within the next 2-3 quarters.

Key Decisions for Investors

• Long NVDA vs short AMD on a 2-4 month horizon; thesis is that NVDA’s growth visibility deserves the premium, while AMD’s richer multiple leaves less room for execution misses. Falsify if AMD re-accelerates materially or NVDA prints a guide-down.
• Buy MU or SNDK on weakness as the second-order beneficiaries of sustained AI buildout; 3-6 month trade on continued memory tightness. Risk/reward is better than chasing NVDA if the market keeps rotating into supply bottlenecks.
• Initiate a small META call spread or buy the stock on post-news pullbacks, looking for a 1-3 month re-rating if outside-capacity monetization gets details. Exit if capex rises again without evidence of external demand or ad growth slows materially.
• Set an alert on NVDA after the next earnings cycle: if the market still values it near the S&P multiple despite continued revenue acceleration, add on any consolidation; if the multiple compresses further on no fundamental miss, the crowd may be over-discounting duration.