JPMorgan reiterates Broadcom stock Overweight on TPU v9 progress

JPMorgan reiterated an Overweight rating and $580 price target on Broadcom, saying reports of TPU v9 2nm ASIC delays are incorrect and the chip remains on track for a 2028 ramp. The note highlights Broadcom’s leadership in Google TPU programs, with TPU v8i expected to ramp this quarter and the March Google agreement supporting revenue growth through 2031. Broader news includes a $2.5 billion debt tender offer and additional security updates, reinforcing active capital management and product execution.

Analysis

The key market implication is not the headline on near-term GPU/ASIC demand, but the implied extension of Broadcom’s design-win duration. If the roadmap is genuinely locked for multiple TPU generations, the equity is starting to price a quasi-annuity on AI custom silicon, which should compress the discount rate on forward revenue more than it boosts near-term estimates. That favors AVGO vs. the broader semiconductor group because the market still tends to underwrite custom silicon as cyclical capex rather than contractual platform spend.

Second-order beneficiaries are the HBM, advanced packaging, and leading-edge foundry ecosystems, but the real bottleneck is system integration and power delivery, not just wafer starts. Any delay elsewhere in Google’s internal TPU effort increases Broadcom’s strategic leverage because it effectively widens the performance gap between a turnkey external partner and a still-maturing in-house stack. That creates a subtle negative read-through for smaller AI ASIC challengers: the longer enterprise customers observe a functioning outsourced model scaling, the harder it becomes to justify multi-year internal chip programs.

The bond tender is also meaningful in a different way: it signals management is comfortable returning incremental cash to the capital structure while preserving AI investment, which reinforces the view that this growth is being funded from operating cash flow rather than balance-sheet stress. That should support AVGO credit spreads and reduce equity fear around capex intensity, but it also caps upside if the market was hoping for a more aggressive M&A or buyback posture. The main tail risk is that the AI custom-silicon narrative gets ahead of actual delivery cadence; if ramp timing slips even modestly, the multiple can compress quickly because the stock is already trading on long-dated expectations.

Key Decisions for Investors

• Long AVGO vs. long-only semiconductor basket (SOXX/SMH) for 3-6 months: AVGO should outperform if the market continues to re-rate AI custom silicon as contractual, not cyclical; downside is a broad AI multiple reset.
• Buy AVGO Jan-2026 call spreads into weakness: expresses upside from duration re-rating with defined risk; target is a 15-20% equity move, while loss is limited to premium paid.
• Pair trade: long AVGO / short weaker AI ASIC development names or semis with no durable design-win visibility for 6-12 months; thesis is that execution certainty is being rewarded more than raw AI exposure.
• Add AVGO credit via existing corporate bonds or CDS tightener bias for 1-3 months: the tender and cash generation reduce balance-sheet stress, but stop out if AI capex commentary turns more aggressive than expected.
• Avoid chasing the first post-news spike; better entry is on any 3-5% pullback over the next 2-4 weeks once sell-side revisions stabilize and before the next ramp confirmation.