UBS reiterates Broadcom stock Buy rating on Google TPU deal By Investing.com

UBS reiterated a Buy on Broadcom with a $475 price target after the company expanded its Google partnership through 2031 and lifted Anthropic-related TPU commitments toward $50 billion. UBS raised its 2027 revenue estimate to $195 billion for fiscal 2027 and $212 billion for calendar 2027, with AI revenue now modeled at $145 billion versus $133 billion previously. The article is broadly positive for Broadcom, though analyst views remain mixed, with Seaport downgrading to Neutral while other firms maintained bullish targets.

Analysis

This is less a re-rating of Broadcom’s near-term AI demand than a validation that the custom silicon supply chain is becoming the bottlenecked toll road for hyperscaler spend. The second-order winner is the broader rack-level ecosystem: if TPU volumes keep scaling, Broadcom’s networking, interconnect, and component attach rates should compound faster than pure-ASIC unit growth, while incremental bargaining power shifts away from merchant GPU suppliers toward vertically integrated buyers who need end-to-end supply assurance.

The market likely underestimates how much of this is a duration story rather than a one-quarter earnings story. Extending visibility to 2031 reduces perceived cyclicality and should compress the discount rate on AVGO’s AI cash flows, but it also increases the risk that investors extrapolate a straight-line ramp and ignore capacity, execution, and customer concentration constraints if one or two hyperscaler programs slip. The real tell over the next 6-12 months will be whether Broadcom can convert promise into widening gross margins and whether non-TPU ASIC diversification keeps pace with investor expectations.

GOOGL is the subtler beneficiary: it is effectively de-risking internal AI infrastructure at a time when compute scarcity is strategic, not just operational. The hidden risk is that more committed TPU capacity may raise the bar for monetization — if model usage or enterprise AI demand does not keep up, the market could eventually treat this as capex intensity disguised as platform strength. For now, the move looks underdone in AVGO and only modestly constructive for GOOGL, but the setup is more vulnerable to sentiment reversal if AI spending broadens into a cost-control narrative or if competing accelerators prove good-enough on price/performance.

Key Decisions for Investors

• Add AVGO on any 3-5% post-headline pullback; use a 3-6 month horizon with a favorable skew toward multiple expansion if investors start capitalizing the 2031 visibility, but size down for customer concentration risk.
• Pair trade: long AVGO / short a basket of merchant AI-fabless semi names with weaker supply visibility over the next 1-2 quarters, targeting relative performance as custom silicon confidence improves.
• For more convex exposure, buy AVGO call spreads 6-9 months out; the best payoff is if the market begins to mark 2027 revenue up as a durable run-rate rather than a one-off contract step-up.
• Accumulate GOOGL only on weakness, not strength; the trade is a 12-month de-risking of AI infrastructure, but upside is capped unless TPU-driven efficiency translates into measurable cloud share gains.
• Avoid chasing the broad AI basket here; this news is idiosyncratically positive for AVGO and only secondarily supportive for semis, so beta-heavy longs are poor risk/reward versus a targeted AVGO expression.