SpaceX, Micron, and Sandisk Are Getting All the Hype -- but This Stock Quietly Beats Them All for Long-Term Investors

Broadcom’s Q2 fiscal 2026 revenue rose 48% YoY to $22.2B, with AI semiconductor revenue up 143% YoY to $10.8B; management guided AI semiconductor revenue of $16B for Q3. The company generated $10.3B in free cash flow on only $231M capex, helped by outsourcing ~95% of wafer production to TSMC. Key risks include customer concentration (~45% of net revenue from top 5) and potential financing exposure up to $29B tied to a 5-year AI rack lease arrangement; despite this, the stock trades about 20x the FY2027 estimate ($19.53) as EPS is expected to grow 68% YoY.

Analysis

AVGO is the cleanest way to own AI infrastructure because it monetizes both the silicon layer and the interconnect layer while keeping its own capital intensity low. That combination matters: in an AI buildout, the company with the best cash conversion usually deserves the premium multiple, and it also has more flexibility to keep buying back stock or absorbing temporary demand noise. TSM is an indirect beneficiary as the manufacturing choke point, while SPCX is the opposite end of the spectrum: high capex, high execution risk, and more dependence on external funding conditions.

The underappreciated risk is that AVGO is starting to look less like a pure semiconductor compounder and more like a hybrid operating company with embedded credit exposure. If the customer behind the lease support weakens, the market will likely reprice AVGO before any accounting loss shows up, and the first move would come through the equity multiple, then spreads on the financing chain, then supplier sentiment. That creates a 1-3 month catalyst path where headlines matter more than fundamentals.

For MU and SNDK, the near-term trend is still favorable, but the trade is increasingly about how long pricing stays above replacement-cost economics. The contract coverage reduces the odds of an immediate air pocket, yet it does not eliminate a 6-18 month normalization if memory supply keeps catching up. Consensus may be underweighting that AVGO has the better quality of earnings, while also underpricing the possibility that one financing misstep can compress its multiple quickly.

Key Decisions for Investors

• Long AVGO on pullbacks vs. short MU or SNDK as a quality-vs-cyclical pair for the next 1-3 months; seek 2:1+ upside if AI infrastructure spend stays intact, with the main risk being a continued memory squeeze that keeps MU/SNDK elevated.
• Add TSM as a lower-beta AI supply-chain hedge to AVGO; 6-12 month thesis is steady wafer demand with less financing overhang, but trim if U.S.-China export rhetoric or foundry utilization softens.
• For accounts already long AVGO, buy 3-6 month put spreads as protection against customer-concentration/lease-financing headline risk; this is a tail hedge against a sharp multiple reset, not a view on near-term fundamentals.
• Avoid chasing SPCX strength until capex is proven self-funding; the risk/reward is poor if the AI investment cycle cools or capital markets tighten.