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

The article argues that SpaceX's aspirational $1 trillion revenue target by 2030 looks unlikely, while Broadcom and AMD have more visible AI-driven growth paths. Broadcom says ASIC revenue could exceed $100 billion in fiscal 2027 versus nearly $64 billion of total revenue in fiscal 2025, and AMD cites two $100 billion GPU deals plus a projected $120 billion CPU market. Overall tone is constructive on AI semiconductor beneficiaries and skeptical on SpaceX's long-term revenue ambitions.

Analysis

The market is effectively drawing a line between speculative platform narratives and monetizable infrastructure. The clearest second-order winner is not just AVGO/AMD, but the entire “picks-and-shovels” stack around custom silicon, high-speed interconnect, and memory optimization: as hyperscalers diversify away from NVDA for inference, more budget shifts from discrete GPUs into bespoke architectures, networking, and software that improves memory density. That tends to compress headline capex ROI scrutiny across the ecosystem and rewards suppliers with design-in leverage and multi-year visibility.

Broadcom’s setup is stronger than a simple AI beta trade because custom ASIC wins create switching costs that persist through several upgrade cycles. The key implication is that if hyperscalers keep bifurcating training vs inference spend, AVGO can capture share even if unit growth in AI accelerators moderates, while its networking attach should compound as cluster size expands. The hidden risk is concentration: one or two large customer ramps can create a temporary growth illusion, but they also make 2027 guidance vulnerable to customer digestion if internal TPU-like programs overdeliver and reduce incremental outsourcing.

AMD’s upside is more “catch-up plus product fit” than outright displacement. The market may be underestimating how much inference economics depend on memory bandwidth and total solution cost, which favors AMD’s packaging and system-level approach more than raw GPU specs; however, this is still a credibility trade that needs sustained software adoption and uptime proof. The contrarian point is that NVDA is not necessarily the loser of this shift—its moat migrates toward frontier training and platform software, while the real share loss may fall on smaller accelerator vendors and weaker inference-only challengers.

SpaceX is the distraction trade here: a long-dated moonshot with engineering constraints that could consume capital before monetization. For portfolio construction, the better expression is to own the businesses selling the tools that make AI compute cheaper, denser, and more power-efficient, while fading narratives that rely on breakthroughs in orbital infrastructure over the next few years.

Key Decisions for Investors

• Add AVGO on 3-6 month pullbacks; target a 12-18 month hold. Best risk/reward is continued multiple support from ASIC visibility and networking attach, with downside mainly from customer capex digestion rather than structural demand loss.
• Initiate a relative-value long AMD / short NVDA pair for a 6-9 month horizon. Thesis: incremental inference share gains and end-to-end solution wins can rerate AMD faster than NVDA’s already-premium AI multiple; stop if NVDA reaccelerates in inference software attach.
• Buy AVGO calls 6-12 months out rather than common if timing a catalyst into the next hyperscaler capex cycle. This expresses upside to custom silicon share gains while limiting drawdown if guidance proves conservative.
• Avoid chasing SpaceX-linked speculation; if anything, use any enthusiasm-driven rally in adjacent unprofitable AI infrastructure names as an opportunity to short or trim exposure. The market is assigning venture-stage optionality to a multi-year engineering problem.
• For conservative capital, build a basket long AVGO + AMD and underweight pure-play AI hardware hype names. This captures the visible monetization path while reducing exposure to breakthrough-risk narratives that can underperform for quarters before the market reprices them.