AVGO
Broadcom Inc.
AI Custom Silicon & Networking
~$420
May 2026 · Mkt Cap ~$1.98T
AI revenue Q1: $8.4B (+106% YoY)
✦ STRONG ADD
Entry zone: $400–430
Hyperscaler custom AI silicon (Google TPU, Meta XPU, OpenAI XPU 2027) + AI networking. AI revenue +106% YoY; $73B AI backlog; Anthropic 1GW TPU 2026 → 3GW+ 2027. FCF margin 42%. Inflexion Capital 5-star pick. Best complement to MU+PLTR — different AI layer.
Fwd P/E
37–38×
PEG ~1.0 blended
AI Revenue Q1
$8.4B
+106% YoY
Q1 Rev Growth
+29%
$19.3B total
Q2 Guide
$22B
+47% YoY
FCF Margin
42%
Above semi median
Adj EBITDA
68%
of revenue
AI Backlog
$73B
Next Earnings
~Jun 3
Q2 FY26 · Est.
Valuation vs Growth
B+
Fwd P/E 37-38× on 29% blended growth, 42% FCF margin. AI segment PEG ~0.6 (106% growth). Q2 guide $22B (+47% YoY) signals acceleration. $73B AI backlog underpins multi-year visibility. $10B new buyback authorised.
Justified by accelerating AI mix. Blended multiple looks high; AI-segment PEG is attractive.
Fundamentals vs Hype
A
Custom ASIC with hyperscaler lock-ins (Google TPU, Meta, Anthropic, OpenAI XPU 2027). EBITDA 68% of revenue. AI revenue +106% YoY is concrete and measurable. Estimate revisions UP. FCF $8.1B Q1. Diversification from NVDA (custom vs GPU architecture).
Grounded — concrete hyperscaler lock-ins with long-cycle custom silicon programmes.
Institutional vs Retail
A
Strong institutional accumulation per Q4 2025 13Fs. Less retail noise than NVDA/AMD — professional-dominated holder base. Confirmed by Inflexion Capital 5-star pick. Short interest low. Analyst consensus strong-buy.
Institutionally accumulated.
Verdict
Strong Add. Best complement to MU/PLTR — custom silicon/networking vs memory and software platform. The ASIC moat is sticky: switching costs from a hyperscaler's own silicon programme are multi-year. Deploy after the May 20 NVDA print confirms hyperscaler capex is intact, then ahead of AVGO's own Q2 print (June 3). At $400-430, the risk/reward is asymmetric given the $73B backlog.
Entry zone: $400–430 (deploy post-NVDA May 20 print, ahead of AVGO Jun 3)
⚠ Key Risks
- 01Single-hyperscaler concentration — if any major customer (Google, Meta, OpenAI) pauses its custom silicon programme, AI revenue derates sharply
- 02VMware integration complexity — $69B acquisition adding enterprise software overhead; margin profile different from semiconductor
- 03NVDA custom ASIC competition — Jensen has signalled willingness to offer custom configurations; long-term competitive pressure
- 04AI capex digestion — same cascade risk as all hyperscaler-dependent names; any Q2 guide-down ripples through AVGO AI revenue
- 05CEO Hock Tan succession — company is CEO-dependent; any leadership uncertainty creates multiple compression