2 High-Growth AI Stocks I'd Be Buying Instead of SpaceX

The article is bullish on Broadcom and AMD, arguing both have clear, visible paths to strong AI-driven revenue growth, while SpaceX’s cited $1 trillion 2030 revenue target is described as a longshot. Broadcom projected more than $100 billion in ASIC revenue in fiscal 2027 versus nearly $64 billion in total revenue in fiscal 2025, and AMD highlighted two $100 billion GPU deals plus a $120 billion CPU market opportunity tied to agentic AI. The piece is primarily an investing commentary rather than new company-specific news, so expected market impact is limited.

Analysis

The market is still underpricing how much of the next AI capex cycle is shifting from brute-force training toward inference-heavy, system-level optimization. That favors the companies that sit closest to custom silicon design, memory bandwidth, and data-center orchestration rather than the headline GPU monopoly trade. In that regime, Broadcom is the cleaner beneficiary because it monetizes both the chip design flow and the networking spend that scales with larger clusters, creating a second-order pull-through effect that extends beyond a single accelerator win.

AMD’s setup is more interesting than the usual “Nvidia alternative” framing because the upside is increasingly about memory architecture and bundled solutions, not just raw GPU share. If inference demand continues compounding over the next 12-24 months, the real earnings lever is not a one-time chip shipment but recurring platform adoption tied to software, system integration, and CPU attach rates as agent workloads proliferate. That makes AMD’s upside more convex than the market typically assumes, but also more dependent on execution and customer concentration than Broadcom.

The contrarian miss is that SpaceX-style moonshot narratives can crowd out the more monetizable AI picks-and-shovels opportunity. Investors may be extrapolating AI scarcity into any capital-intensive platform story, when in reality the near-term cash flows accrue to the vendors selling shovels, interconnect, and customization layers. The risk is that hyperscalers slow bespoke silicon spend if enterprise AI ROI normalizes faster than expected; that would hit AVGO first, then AMD, with NVDA relatively insulated on training.

Near term, the catalyst path is product-cycle, not macro: design wins, backlog conversion, and guidance updates over the next 1-3 quarters. The main reversal trigger is a capex digestion phase in which cloud customers pause after the current buildout, compressing multiples before revenue catches up. If that happens, the highest-beta names in the custom/edge AI stack will de-rate faster than the category leaders.

Key Decisions for Investors

• Go long AVGO on any 3-5% pullback; use a 6-12 month horizon. Best risk/reward is in owning the infrastructure toll-booth on custom AI and networking pull-through rather than chasing the training leader at peak consensus.
• Add AMD only on confirmation of inference-guidance momentum; prefer a staggered entry over 30-60 days. Upside is larger if agentic AI adoption accelerates, but the position should be sized below AVGO due to higher execution risk.
• Pair trade: long AVGO / short NVDA on a 3-6 month horizon if you want to express inference shift without taking full AI beta. Thesis is that custom silicon and networking spend re-accelerate while NVDA’s training premium narrows.
• For more convex exposure, buy AVGO or AMD call spreads 9-12 months out rather than outright calls. This limits downside if hyperscaler capex pauses while preserving upside if the buildout continues through next year.
• Avoid chasing space-infrastructure AI narratives until cooling/radiation/manufacturing feasibility is proven; if you want optionality, keep it as a venture-style side bet, not core public-equity exposure.