Broadcom (AVGO)

Investment Review After Q2 FY2026 Earnings (stock price as of 2026.06.03)

TL;DR

Broadcom is the custom AI accelerator (XPU) and AI networking silicon leader for six hyperscale frontier-model customers, paired with VMware infrastructure software as a high-margin recurring base. Just reported: Q2 FY2026 revenue record $22.2B (+48% YoY), operating margin a record 67%, AI semiconductor revenue $10.8B (+143% YoY) with over $30B of AI bookings — a strong beat. Q3 FY2026 guide is $29.4B revenue (+84% YoY) including AI semi $16B (+200%+ YoY), with operating margin steady at ~67% — accelerating sharply. Stock sits at $479 on 49x P/E(ttm), +55% post last quarter end, an extreme multiple pricing in flawless AI execution — shares fell ~14% after-hours despite the beat.

Summary

Finding	Assessment
5-Year Stock Return	Stock currently $479, up ~11x from $43 (5 years ago). TTM OP grew from ~$5.6B to $50.1B. P/E(ttm) ranged 12x–49x, currently 49x.
Business Model	DOMINANT dual engine. AI semiconductors (custom XPUs for 6 customers) + Infrastructure Software (VMware/VCF at ~93% GM). Fabless model with strong FCF conversion.
Competitive Moat	WIDE Custom XPU co-design lock-in, AI networking leadership (Tomahawk/Jericho/CPO), VMware enterprise lock-in. Three reinforcing moat sources.
Past 5-YR Drivers	Custom AI XPU franchise (6 customers, near-zero to $56B FY26 run-rate), AI networking share gains (Tomahawk, Jericho, CPO), VMware integration (~$7.2B/qtr, ~93% GM).
Recent Stock Move	Stock currently $479, vs. $310 (Apr 2026 close). +55% since last quarter end, but fell ~14% after-hours on the print despite the beat. P/E(ttm) re-expanded to 49x.
Revenue Trend	ACCELERATING Rev YoY: 48% (Q2 2026) → 86% (Q3 2026E).
Margin Trend	FLAT OM%: 67% (Q2 2026) → 67% (Q3 2026E).
TTM OP Trajectory	ACCELERATING TTM OP QoQ: 11% (Q2 2026) → 19% (Q3 2026E). TTM OP $50,130M.
Key Opportunities	FY2027 AI semi >$100B reiterated; AI bookings over $30B with visibility to 2028; Apollo/Blackstone 20GW XPU platform; VMware re-accelerating (+31% Q3).
Key Risks	49x P/E demands flawless AI execution. ~14% after-hours drop on the beat; gross margin compressing (77%→74%); 6-customer concentration; non-AI semi only +6%.
Catalysts	Q3 FY2026 $29.4B revenue (+84% YoY), AI semi $16B (+200%+). 200-terabit Tomahawk tape-out. FY26 AI $56B / FY27 >$100B. CFO transition (Spears→Thuener).
P/E(ttm) Valuation	EXTREME at 49x, top of the 5-year range, pricing in flawless AI execution. Supported by +48% revenue growth, record 67% OM, and a $56B→$100B+ AI semi path — but the ~14% after-hours drop shows the bar is high.

Q2 FY2026 Key Data (Feb-Apr 2026), reported 2026.06.03

Revenue record $22.2B (+48% YoY). Operating margin record 67%; operating income record $14.9B (+52%); adjusted EBITDA $15.2B (69% of revenue); FCF record $10.3B (46%).
Semiconductors record $15.0B (+79% YoY), 68% of revenue. AI semi record $10.8B (+143%), 49% of total; AI bookings over $30B; non-AI semi $4.2B (+6%). Software $7.2B (+9%), ARR +17%.
Q3 FY2026 guided $29.4B revenue (+84% YoY). Semis $20.5B (+124%) incl AI semi $16B (+200%+); software $8.9B (+31%); operating margin ~67%; gross margin ~74% on mix.

Key Takeaways from Earnings Call

AI demand is "insatiable" and the entire growth story. Q2 AI semi revenue hit a record $10.8B (+143% YoY) while bookings exceeded $30B in the quarter. Management guides 2H'26 AI semi to double 1H, FY2026 AI semi ~$56B (+~180% YoY), and reiterates FY2027 AI semi above $100B. Visibility now extends to 2028 (from 2027 three months ago) as the six core customers order years ahead to secure wafers, HBM, and power.
Six core customers anchor multi-gigawatt, multi-year commitments. Google (long-term multi-generation TPU + networking agreement, "very substantial $"), Anthropic (1GW in 2026, +5GW from 2027), OpenAI (silicon delivered, production late 2026, 1.3GW in 2027 within a 10GW-by-2029 deal), and Meta (3GW MTIA through 2028). The other two customers begin shipping late 2026; AI purchase orders already total $6B.
Margins stay strong despite mix dilution; networking ~40% of AI. Operating margin rose 200bps YoY to a record 67.3% even as gross margin fell 230bps to 77.1% on richer semiconductor mix. Q3 gross margin guides to ~74% as AI scales — "not a structural change," just TPU/ASIC mix vs high-margin connectivity. Networking was ~40% of AI revenue this quarter (Tomahawk 6, Jericho, CPO), which Hock expects to settle nearer 30%.
New AI XPU platform with Apollo and Blackstone to fund frontier-lab compute. Broadcom is creating a vehicle with top-balance-sheet investors to deploy 20+ gigawatts of compute through 2028; the first $35B tranche is being launched by Apollo. The goal is to give frontier labs (Anthropic, OpenAI) lowest-cost, lowest-power capacity. CFO Kirsten Spears retires June 12 after 12 years; Amie Thuener joins as incoming CFO.

Key Segments and Revenue Mix (FY2025)

Segment	FY25 Rev	Mix	Products	End Markets	Key Competitors
AI Semiconductors	~$20B	26%	Custom XPUs (TPU, MTIA, etc.), AI accelerators	Hyperscale AI training & inference (Google, Meta, etc.)	Marvell, NVIDIA (different approach)
AI Networking	~$10B	13%	Tomahawk 6 switches, Jericho fabric, SerDes, CPO	AI data center fabric, hyperscale networking	Cisco, Marvell, NVIDIA (Spectrum)
Non-AI Semiconductors	~$16B	21%	Broadband, server storage, wireless	Enterprise networking, broadband, wireless	Marvell, Qualcomm, Intel
Infrastructure Software (VMware)	~$25B	32%	VMware Cloud Foundation (VCF), vSphere, NSX	Enterprise private/hybrid cloud, multi-cloud	Nutanix, Microsoft (Hyper-V), Red Hat
Total FY2025	~$77B	100%	AI semiconductor + networking ~39% of total. VMware integration driving ~93% software gross margins.

Business Model

SUPPLIERS

TSMC (3nm/5nm fabrication)
HBM & DRAM memory vendors
Packaging & substrate vendors
EDA tools (Synopsys, Cadence)

Wafers / HBM →

BROADCOM

Custom XPUs for 6 hyperscalers
Tomahawk switches, Jericho, CPO
VMware Cloud Foundation (VCF)
Broadband, wireless, storage ICs
OM 67%, EBITDA 69%

Chips / Software → $ Revenue

CUSTOMERS

Google (TPU), Anthropic, OpenAI
Meta (MTIA), + 2 other customers
Enterprise (VMware/VCF)
Telcos (broadband)
Wireless OEMs (Apple)

Moat Assessment: WIDE

Custom XPU co-design moat: Broadcom embeds into each customer’s silicon roadmap. Switching costs are massive — software stacks, requalification, and multi-year design cycles create multi-generation lock-in across six core customers.
AI networking dominance: Broadcom holds a generation of interconnect leadership — the 100-terabit Tomahawk 6, Jericho 3/4, 200G/400G SerDes, and CPO. As clusters standardize on this fabric, the ecosystem deepens.
VMware enterprise lock-in: vSphere/VCF holds dominant virtualization share. Migration costs are prohibitive — years of configurations and ops knowledge. VCF 9.1 heterogeneous compute (AMD/Intel/NVIDIA) deepens the lock-in.

Stock vs. Earnings — 5-Year Performance

Stock currently $479, up roughly 11x from $43 (5 years ago), driven by the emergence of a dominant dual-engine model — custom AI silicon for hyperscalers and VMware infrastructure software.
TTM operating income grew from ~$5.6B to $50,130M, as Broadcom transformed from a diversified chipmaker into the leading custom AI XPU and networking platform for the world’s largest cloud and frontier-model companies.
P/E(ttm) ranged 12x–49x over the period, currently 49x at the top of the range. The premium reflects the market’s recognition of Broadcom’s structural position as the custom-silicon partner for six hyperscale AI customers.

Broadcom stock price vs TTM operating income, 5-year chart

Past 5-Year Key Drivers

Custom AI XPU franchise: Broadcom co-designs custom AI accelerators (XPUs) for six core customers including Google (TPU), Meta (MTIA), Anthropic, and OpenAI. AI semi revenue scaled from near-zero to a $56B FY2026 run-rate.
AI networking share gains: Tomahawk switches, Jericho fabric, SerDes, and CPO became the de facto AI data center interconnect. Networking ran ~40% of AI revenue this quarter, with the 200-terabit Tomahawk taping out.
VMware integration driving software margins: The ~$69B VMware deal (closed Nov 2023) made Broadcom dual-engine. Software now generates ~$7.2B/quarter at ~93% gross margins and 17% ARR growth — a recurring stream offsetting chip cyclicality.

Quarterly Key Metrics

	Q2'24	Q3'24	Q4'24	Q1'25	Q2'25	Q3'25	Q4'25	Q1'26	Q2'26	Q3'26(E)
Quarter End	Apr 2024	Jul 2024	Oct 2024	Jan 2025	Apr 2025	Jul 2025	Oct 2025	Jan 2026	Apr 2026	Jul 2026
Stock Price	$135	$164	$172	$232	$167	$276	$330	$346	$310	$479
Stock QoQ	+24%	+22%	+5%	+34%	-28%	+65%	+20%	+5%	-11%	+55%
Rev YoY	+43%	+47%	+51%	+25%	+20%	+22%	+28%	+29%	+48%	+86%
OM%	57%	61%	63%	66%	65%	66%	66%	66%	67%	67%
R40 (Rev YoY + OM%)	+100%	+108%	+114%	+91%	+85%	+88%	+94%	+96%	+115%	+153%
TTM OP ($M) non-GAAP	25,260	27,672	30,736	33,733	36,380	38,887	41,997	44,995	50,130	59,625
TTM OP QoQ	+7%	+10%	+11%	+10%	+8%	+7%	+8%	+7%	+11%	+19%
FCF/OP ttm	73%	67%	63%	61%	62%	64%	64%	64%	65%	65%
Debt/EBITDA	2.0x	1.7x	1.5x	1.4x	1.3x	1.2x	1.1x	1.1x	1.0x	1.0x
P/E(ttm)	32x	35x	35x	43x	28x	43x	48x	47x	37x	49x

Recent Stock Performance & Drivers

Stock currently $479, vs. $310 (Apr 2026 quarter close). The shares ran +55% since last quarter end on the AI re-rating, but fell ~14% after-hours to ~$414 on the Q2 print despite the beat — a valuation/expectations reset.
TTM OP grew to $50.1B — up +11% QoQ in the just-reported quarter, driven by the AI semiconductor ramp (record $10.8B, +143% YoY) and stable, expanding operating margins.
P/E(ttm) re-expanded to 49x from 37x at the Q2 close, the top of the 5-year 12x–49x range. R40 jumped to +115% this quarter on accelerating revenue and a record 67% operating margin.
Debt/EBITDA improved to ~1.0x — Broadcom continues deleveraging post-VMware, with FCF/OP around 65% and a record $10.3B free cash flow in the quarter.

Opportunities

FY2027 AI semiconductor revenue reiterated above $100B. After ~$56B in FY2026 (+~180% YoY), management sees FY2027 nearly doubling again, with 2028 a "substantial" further step up as six core customers deploy multi-gigawatt clusters.
AI bookings over $30B in the quarter vs $10.8B shipped. Visibility now runs to 2028. Customers order years ahead to secure wafers, HBM, and power, giving Broadcom unusual semiconductor demand visibility and $6B of AI purchase orders already in hand.
VMware re-accelerating into AI. Software guides to $8.9B in Q3 (+31% YoY) as VCF 9.1 rides strong server demand; heterogeneous compute (AMD/Intel/NVIDIA) positions it for enterprise AI inferencing on private cloud.

Challenges

P/E(ttm) at 49x is extreme and prices in flawless execution. The ~14% after-hours drop on a clear beat shows how high the expectations bar sits; any AI capex wobble or XPU customer loss could compress the multiple sharply.
Gross margin compressing on AI mix. Gross margin fell to 77.1% and guides to ~74% in Q3 as lower-margin TPU/ASIC volume scales faster than high-margin connectivity and software — structurally fine, but optically a headwind.
Extreme customer concentration. Six core AI customers — led by Google, Anthropic, and OpenAI — drive nearly half of total revenue, with lumpy bookings. Non-AI semis grew only +6%, a modest cyclical recovery that dilutes the AI story.

Next 12 Months

Q3 FY2026 guided $29.4B revenue (+84% YoY), AI semi $16B (+200%+). A beat would confirm the 2H AI doubling; soft Q4 sequencing (the FY26 $56B framing implies Q4 AI flat-to-down) is the key debate after the print.
AI networking mix. Networking ran ~40% of AI revenue this quarter; Hock expects it to settle nearer 30% as XPU volume scales. The 200-terabit Tomahawk tape-out this quarter is the next milestone.
CFO transition. Kirsten Spears retires June 12 after 12 years; Amie Thuener steps in. Continuity of capital allocation ($3.1B Q2 dividends, $0.65/share) and disclosure cadence is a watch item.

Next 12-24 Months

20+ gigawatts of compute through 2028 via the Apollo/Blackstone XPU platform. The first $35B tranche funds frontier-lab capacity — converting multi-gigawatt commitments (Anthropic 5GW, OpenAI 10GW, Meta 3GW) into durable revenue.
Rising content per gigawatt. As XPUs add SRAM, sparse cores, embedded CPU cores, multi-die, and more HBM, dollar content per gigawatt rises each generation — a structural tailwind on top of gigawatt growth.

Core Metrics Scorecard

	Q4'21	Q1'22	Q2'22	Q3'22	Q4'22	Q1'23	Q2'23	Q3'23	Q4'23	Q1'24	Q2'24	Q3'24	Q4'24	Q1'25	Q2'25	Q3'25	Q4'25	Q1'26	Q2'26	Q3'26(E)
Quarter End	Oct21	Jan22	Apr22	Jul22	Oct22	Jan23	Apr23	Jul23	Oct23	Jan24	Apr24	Jul24	Oct24	Jan25	Apr25	Jul25	Oct25	Jan26	Apr26	Jul26
R40 > 40%	✅	✅	✅	✅	✅	✅	✅	✅	✅	✅	✅	✅	✅	✅	✅	✅	✅	✅	✅	✅
R40 accelerate		✅	✅	✅						✅	✅	✅	✅			✅	✅	✅	✅	✅
TTM OP QoQ > 10%													✅						✅	✅
TTM OP QoQ accelerate		✅	✅	✅						✅	✅	✅	✅				✅		✅	✅
TTM FCF/OP > 60%	✅	✅	✅	✅	✅	✅	✅	✅	✅	✅	✅	✅	✅	✅	✅	✅	✅	✅	✅	✅
Debt/EBITDA < 3.5	✅	✅	✅	✅	✅	✅	✅	✅	✅	✅	✅	✅	✅	✅	✅	✅	✅	✅	✅	✅
Total Passes	3	5	5	5	3	3	3	3	3	5	5	5	6	3	3	4	5	4	6	6
Stock QoQ	+2%	+37%	-5%	-23%	-9%	+25%	+16%	+35%	-4%	+30%	+24%	+22%	+5%	+34%	-28%	+65%	+20%	+5%	-11%	+55%
P/E(ttm)	17x	21x	19x	13x	12x	14x	15x	20x	19x	26x	32x	35x	35x	43x	28x	43x	48x	47x	37x	49x

Q2’26 hit a clean 6/6 sweep, the strongest reading in the dataset. All six criteria fired — R40 >40% and accelerating, TTM OP QoQ >10% and accelerating, FCF/OP >60%, Debt/EBITDA <3.5x. 6-pass quarters (Q4’24, Q2’26) mark AI-ramp inflections.
R40 >40% and Debt/EBITDA <3.5 pass in all 20 quarters — a guaranteed floor of 2. FCF/OP >60% also passes in all 20. The swing factors are the acceleration metrics plus TTM OP QoQ >10%, which only fired in Q4’24, Q2’26, and Q3’26E.
TTM OP QoQ >10% is the differentiator between 4-pass and 6-pass quarters. The score sat at 3 through the FY2023 pre-VMware deceleration, rose to 5 in the 2024 AI ramp, and now peaks at 6. The Q2’25 stock blowout (+65%) came at only 4 passes.

Chart: R40 vs Stock Price (Since 2020)

Broadcom R40 vs stock price chart since 2020

Chart: Scorecard Total Passes vs Stock Price (Since 2020)

Broadcom scorecard total passes vs stock price chart since 2020

Q1 FY2026 Key Data, reported 2026.03.04

Revenue +29% YoY, OM 45%, TTM OP $29B (+131% over 6 quarters). R40 +74%, FCF/OP 100%, Debt/EBITDA 1.4x — aggressive deleveraging from 3.7x post-VMware.
AI semiconductor ramp at 6 hyperscaler XPU customers (OpenAI added 2025). AI networking targeting 40% of total AI revenue via Tomahawk 7 and 200G SerDes. CY2027 target: >$100B AI chip TAM.
Q2 FY2026 guided $22B revenue (+47% YoY), AI semi $10.7B (+140% YoY). VMware/VCF generating ~$6.8B/quarter at 93% GM. P/E(ttm) compressed from 60x to 47x as earnings caught up with valuation.

Key Takeaways from Earnings Call

AI is the entire story. 6 confirmed custom XPU customers (OpenAI added 2025), each hyperscaler deploying larger AI training clusters that require custom silicon. AI networking accelerating to 40% of AI revenue via Tomahawk 7 and 200G SerDes. CY2027 target >$100B AI chip TAM; capturing 20-25% would mean $20-25B AI semi revenue.
VMware integration exceeding expectations. VMware/VCF generating ~$6.8B/quarter at 93% gross margins. Subscription transition driving higher recurring revenue and customer lifetime value. Aggressive transition has caused customer friction; Nutanix actively targeting displaced VMware customers, so 93% GM requires sustained pricing power.
Multiple-compression risk is the bear case. P/E(ttm) at 47x requires sustained AI execution — any slowdown in hyperscaler AI capex or loss of a major XPU customer triggers compression toward 35x. Non-AI semiconductor business (~$16B, 21% of revenue: legacy broadband, wireless, server storage) growing only low single digits, diluting the AI story.

Source: SEC EDGAR annual and quarterly filings, company earnings conference calls and presentations.

Disclaimer: This report is for educational purposes only. NO investment advice.