These Are the AI Trades I'd Make Today if I Were Starting My Portfolio From Scratch

The article highlights Qualcomm, Dell, and ServiceNow as attractive AI investments, with Qualcomm entering AI data center inference chips and Dell citing $64 billion in AI-optimized server orders and a $43 billion backlog. ServiceNow posted 19% constant-currency revenue growth and 11% operating income growth, while analysts see 40% upside to a $140.63 consensus target. Overall tone is constructive on AI infrastructure and applications, but the piece is opinion-driven rather than a direct market-moving announcement.

Analysis

The cleanest read is that AI spend is broadening from model GPUs into the plumbing layer where enterprise adoption actually scales. That tends to favor vendors with attachment points to the data center buildout and workflow layer, while making the “pure compute” trade increasingly crowded and valuation-sensitive. The second-order effect is that every incremental dollar of AI capex creates follow-on demand for servers, networking, storage, implementation, and software orchestration, so the winners are becoming more diversified than the market’s first AI basket implied.

Qualcomm’s more interesting opportunity is not inference silicon itself, but edge deployment where latency, privacy, and bandwidth costs matter. If edge AI grows at a sustained high-teens/20%+ rate, the market is likely underpricing Qualcomm’s ability to monetize non-handset sockets over the next 24-36 months; that said, this is still a prove-it story because hyperscaler traction in accelerators can be fickle and pricing pressure will be intense. The bigger near-term upside may come from sentiment re-rating as investors recognize that edge AI is a volume market, not a peak-margin market.

Dell is the most direct beneficiary of AI capex conversion because it monetizes the unglamorous but essential deployment layer. The backlog/ordering dynamic reduces the usual hardware cyclicality, but it also means the stock can de-risk quickly if order timing slips or if customers delay delivery into later quarters. The risk is that the market is extrapolating backlog into revenue too mechanically; the reward is a multi-quarter earnings revision cycle if execution stays clean.

ServiceNow looks like a quality compounder that the market has already punished for being ‘too good too early.’ The bearish consensus seems to be over-focused on valuation reset and underweight the fact that workflow automation is one of the few software categories where AI should increase expansion rates rather than cannibalize pricing. The main risk is multiple compression if growth decelerates even modestly, but over a 6-12 month horizon the setup favors mean reversion if operating metrics remain intact.

Key Decisions for Investors

• Long QCOM on a 6-12 month horizon as an edge-AI re-rating trade; use pullbacks to build because the upside is adoption-driven rather than catalyst-driven, with downside limited by diversified handset/laptop cash flow.
• Long DELL vs. short a higher-multiple AI infrastructure basket over 3-6 months; Dell has the clearest backlog-to-earnings visibility, while the short leg captures the risk that investors are overpaying for less tangible AI exposure.
• Add NOW on weakness over the next 1-2 quarters; this is a quality-growth mean-reversion setup where the market may be discounting slower AI monetization too aggressively relative to durable workflow penetration.
• If already long NVDA/AVGO, rebalance toward QCOM/DELL/NOW to reduce pure-play semis concentration; the trade is for second-order AI beneficiaries with better valuation support and less multiple fragility.
• Use a call spread on DELL for a 3-6 month catalyst window; the backlog supports upside, but a spread limits risk if AI server shipments normalize or margin expectations reset.