With the new Google LTA, how confident is Broadcom on its share and is there upside, or is it a fixed amount?

Hock called it a very strong agreement reflecting the partnership's strength and multi-generation IP, describing the commitment as a very substantial dollar amount. He acknowledged that, given Google's own AI-compute growth, there will likely be some diversity of sources, but Broadcom's dollar commitment from Google remains very substantial. “we fully expect that there will be some diversity of sources for them. But our commitment from them is a very substantial dollar amount”

What is driving the sharper gross-margin decline — XPU vs networking mix, rack vs chip scale — and will it continue?

Kirsten explained ASICs/TPUs and some wireless carry lower margins, so accelerating TPUs pressure margins, partially offset by rich-margin AI networking; Hock stressed semiconductor margins are structurally stable and the dilution is mix. Both clarified Broadcom is in the chip business only, not racks. “We're in the chip business only... We are only chips”

Is dollar content per gigawatt rising materially (as a competitor suggested) as compute and networking content grows?

Hock said billions-of-dollars content per gigawatt is relatively stable because higher-power individual chips mean fewer chips at higher ASP; the real accelerator is the number of gigawatts, which is growing very fast — already running ahead of expectations of six months ago for Anthropic and OpenAI, with 2028 expected to grow substantially over 2027. “dollars per gigawatt -- billions of dollars per gigawatt is relatively stable. But the number of gigawatts will keep -- will accelerate”

AVGO

Broadcom (AVGO) Q2 FY26 earnings call: guidance and Q&A

Earnings call· May 2026 (Q2 FY26)· Information Technology· Jun 3, 2026

Source: earnings call transcript

Executive summary

Broadcom Q2 FY26 — Record $22.2B revenue on insatiable AI semiconductor demand

Broadcom is now an AI-infrastructure franchise: a small set of frontier-model customers is booking multi-gigawatt, multi-year compute far ahead of delivery, extending visibility into 2028 and prompting a co-created funding platform to underwrite the buildout. The offsetting tension is structural mix — rich-margin networking cushions but cannot fully offset TPU dilution — so management pushes investors to model semis and software margins separately as the AI mix swells.

Highlights

Record consolidated revenue of $22.2B, up 48% YoY, above guidance on AI semiconductor strength
Record AI semiconductor revenue of $10.8B, up 143% YoY, above outlook, with AI bookings of over $30B against $10.8B shipped
Record operating income of $14.9B (up 52% YoY) with record 67.3% operating margin (up 200 bps YoY) and record 69% adjusted EBITDA margin, above the 68% guide
Record free cash flow of $10.3B, 46% of revenue
Infrastructure software revenue of $7.2B, up 9% YoY, with ARR sustaining 17% YoY growth

Concerns

Consolidated gross margin fell 230 bps YoY to 77.1% on AI/semiconductor mix shift, and is guided down further to ~74% in Q3 as lower-margin TPUs/ASICs scale
Non-AI semiconductor revenue grew just 6% YoY to $4.2B — a cyclical recovery still in early innings
Days of inventory rose to 86 days from 68 days in Q1 as the company builds ahead of H2 AI demand
Google signaled likely diversification of sources on its TPU program despite a very substantial dollar commitment — a customer-concentration/share question the company acknowledged

Guidance & targets

Category	Target	Timeline	Confidence
Q3 consolidated revenue	$29.4 billion, up 84% year-on-year high materiality	Q3 · FY26	High
Q3 semiconductor revenue	approximately $20.5 billion, up 124% year-on-year high materiality	Q3 · FY26	High
Q3 AI semiconductor revenue	$16 billion, up over 200% year-on-year high materiality	Q3 · FY26	High
Q3 non-AI semiconductor revenue	approximately $4.5 billion, up 12% from a year ago medium materiality	Q3 · FY26	High
Q3 infrastructure software revenue	approximately $8.9 billion, up 31% year-on-year medium materiality	Q3 · FY26	High
Q3 consolidated gross margin	approximately 74% medium materiality	Q3 · FY26	Medium
Q3 operating margin	approximately 67% of revenue (flat quarter-on-quarter) high materiality	Q3 · FY26	High
Q3 adjusted EBITDA margin	approximately 68% of revenue medium materiality	Q3 · FY26	Medium
Non-GAAP tax rate (Q3 and full-year FY26)	approximately 16% low materiality	Q3 FY26 and FY2026	High
Q3 non-GAAP diluted share count	approximately 4.94 billion shares low materiality	Q3 · FY26	High
Second-half FY26 AI semiconductor revenue	to double from the first half high materiality	H2 FY26	High
Full-year FY26 AI semiconductor revenue	$56 billion, up approximately 180% from fiscal 2025 high materiality	FY2026 · FY26	High
Full-year FY27 AI semiconductor revenue	in excess of $100 billion high materiality	FY2027 · FY27	High
FY27 AI compute shipment (gigawatts)	approximately 10 gigawatts, back-half loaded high materiality	FY2027 · FY27	High
FY28 AI semiconductor revenue / gigawatts	substantial growth from FY27 (a lot more gigawatts) high materiality	FY2028 · FY28	Medium
Networking share of total AI revenue (normalized)	closer to around 30% medium materiality	forward / ongoing	Medium
Anthropic TPU-based compute access	over 1 gigawatt in 2026, plus another 5 gigawatts of next-generation TPU-based compute beginning in 2027 high materiality	2026-2027	High
OpenAI XPU deployment	1.3 gigawatts in 2027, as part of a larger 10 gigawatts by 2029 high materiality	2027-2029	High
Meta MTIA XPU deployment	3 gigawatts through the end of 2028; initial 1 GW order (XPUs + networking) to start delivery in H2 2027 high materiality	through end-2028	High
AI XPV compute platform capacity	more than 20 gigawatts of compute capacity through 2028 high materiality	through 2028	High

Segment performance

Segment	Revenue	YoY	QoQ	Margin
Semiconductor Solutions (total) Record semiconductor revenue with growth accelerating to 79% YoY, driven by AI. Strong operating leverage lifted operating margin 460 bps YoY despite an AI-mix-driven gross-margin dilution. Share of total revenue: 68%Operating expenses: $1.2B (8% of revenue), reflecting increased leading-edge AI R&DAI semiconductor revenue as share of total: 49%	$15 billion	79%	—	~70% gross margin; 62% operating margin (up 460 bps YoY)
AI semiconductors Record AI revenue above outlook; networking ~40% (expected to normalize to ~30%), XPU/compute the balance. Bookings dramatically outran shipments, extending visibility to 2028. Share of total revenue: 49%Networking share of AI revenue: almost 40%AI bookings in quarter: over $30B vs $10.8B shipped	$10.8 billion	143%	—	Blended lower than segment (TPUs/ASICs dilutive; networking rich-margin)
Non-AI semiconductors Broadband, server storage and enterprise networking up, partially offset by seasonal wireless decline; bookings >$6B cited as evidence of a full cyclical recovery. Bookings in quarter: exceeded $6BQ3 guide: ~$4.5B, up 12% YoY	$4.2 billion	6%	—	—
Infrastructure Software In-line revenue with strong bookings and 17% ARR growth; VMware Cloud Foundation 9.1 deployment driving accelerated Q3 growth to ~$8.9B (+31% YoY). Share of total revenue: 32%ARR growth: 17% YoYOperating expenses: $1B in the quarter	$7.2 billion	9%	—	93% gross margin; ~79% operating margin (up 310 bps YoY)

AVGO operating KPIs by quarter

AVGO operating KPIs stated on its earnings calls, by fiscal quarter
KPI	Aug 2025 Q3 FY25	Nov 2025 Q4 FY25	Feb 2026 Q1 FY26	Jul 2026 Q3 FY26	Change vs prior quarter
Backlog	$110B And our current consolidated backlog for the company hit a record of $110 billion. Source: earnings call transcript	$162B And all these components combined with XPUs, bring our total order on hand in excess of $73 billion today, which is almost half Broadcom's consolidated backlog of $162 billion. Source: earnings call transcript	—	—	+47.3%
Customers XPU	—	—	5 Now we expect this momentum to accelerate as our custom AI XPUs hit their next phase of deployment among our 5 customers. Source: earnings call transcript	6 Our 6 XPU customers are accelerating the adoption of custom accelerators. Source: earnings call transcript	+20%

Operating figures the company stated on at least two of the calls shown, checked against each call's transcript. Click a figure to read the sentence. A dash means it was not stated that quarter.

Orderbook & backlog

AI semiconductor bookings (in-quarter) over $30 billion Q2 FY26

against $10.8B AI shipped in the quarter (~2.8x coverage)

Bookings, not backlog; large orders placed ahead of delivery to secure wafers, HBM/DRAM and power. Underpins visibility now running into 2028 (vs ~2027 three months ago).

Non-AI semiconductor bookings (in-quarter) exceeded $6 billion Q2 FY26

Cited as a clear indication of a path toward a full cyclical recovery.

Purchase orders from two newer XPU customers $6 billion to date as of Q2 FY26

For two of the six core customers whose shipments begin late 2026 and accelerate into 2027; distinct from the $30B in-quarter AI bookings.

Product announcements

Product	Type	Timeline
Tomahawk 6 (100 terabit Ethernet switch)	milestone	shipping for over a year
Next-generation 200 terabit Ethernet switch	roadmap	taping out this quarter (Q2 FY26)
Co-packaged optics (1.6 terabit DSPs, CW and EML lasers)	milestone	—
VMware Cloud Foundation 9.1	launch	just released (Q2 FY26)
Jericho3 and Jericho4 fabric solutions	milestone	—

Deals & partnerships

Google Long-term supply agreement (multi-generation TPUs and AI networking) undisclosed; described as a very substantial dollar commitment multiple generations / foreseeable future

Reflects strength of the partnership and Broadcom's differentiated IP/execution vs alternatives; disclosed via an intra-quarter 8-K.

Anthropic TPU-based compute supply agreement over 1 GW of TPU compute in 2026; another 5 GW of next-gen TPU compute from 2027 beginning 2026, next-gen from 2027

Hock corrected an analyst's 'backstopped by Broadcom chips' framing — Broadcom is the direct provider of the TPU chips/compute capacity, disclosed in a recent 8-K.

OpenAI Custom AI accelerator (XPU) supply agreement 1.3 GW in 2027, part of a larger 10 GW by 2029 through 2029

10 GW-by-2029 agreement announced last year; late-2026 production start with 1.3 GW deploying in 2027.

Meta Partnership for multiple generations of MTIA XPUs 3 GW through end of 2028; initial 1 GW order (XPUs + networking) received through end 2028

Multi-generation MTIA XPU program.

Apollo and Blackstone (and other leading investors) Joint funding platform (AI XPV) for AI compute capacity First tranche valued at $35B through 2028

Strategic vehicle pairing Broadcom technology with strong-balance-sheet investors to deliver compute at lowest cost and power.

Two additional unnamed core customers XPU supply relationships $6B of purchase orders received to date

Named as the 'other 2 customers' beyond Google, Anthropic, OpenAI and Meta.

Capital programs

AI XPV compute platform (with Apollo, Blackstone and other investors) announced / first tranche launching First tranche valued at $35B; platform sized for more than 20 GW of compute

Funding: External — Apollo, Blackstone and other leading investors with strong balance sheets (not Broadcom capex)

Start: currently being launched by Apollo

Benefit:More than 20 gigawatts of compute capacity deployed through 2028

Vehicle to fund Broadcom chips for frontier labs (incl. Anthropic and OpenAI) that might otherwise struggle to finance capacity, delivering compute at lowest cost and power. Keeps Broadcom in a pure chip-supplier model; own capex remained just $231M in the quarter. 'XPV' transcribed as stated.

Risks & headwinds

Gross-margin dilution from AI/TPU mix shift Q2 FY26 and ongoing as AI mix grows

Consolidated gross margin down 230 bps YoY to 77.1%; guided down to ~74% in Q3; TPUs/ASICs and some wireless carry structurally lower margins

Mitigation:Rich-margin AI networking partially offsets; strong operating leverage keeps operating margin flat at ~67%; management stresses no structural semi-margin change and urges modeling segments separately

Customer concentration in a small set of frontier-model / hyperscaler customers FY26-FY28

AI semiconductor revenue is 49% of total; growth rests on six core customers, with Google signaling likely source diversification

Mitigation:Broadening customer base (two new customers with $6B POs); differentiated multi-generation IP; Google's dollar commitment described as very substantial despite diversification

Supply constraints — wafers, HBM/DRAM, and power 2026-2029

Not quantified; customers order well ahead because lead times require securing chips, memory and power/'power shell'

Mitigation:Supply secured for 2026-2027, working on 2028-2029; management comfortable meeting incremental customer demand and considering foundry optionality

Inventory build ahead of demand Q2 FY26, positioning for H2

Inventory $4.3B; days on hand rose to 86 from 68 in Q1

Mitigation:Framed as a deliberate strategic build to secure supply for accelerating H2 AI semiconductor growth, not a passive/air-pocket build

Non-AI semiconductor cyclicality Q2 FY26; recovery early

Non-AI revenue up only 6% YoY to $4.2B; wireless in seasonal decline

Mitigation:Bookings exceeded $6B, cited as evidence of a path toward full cyclical recovery; Q3 guided +12% YoY

Power and physical-infrastructure gating of AI deployments 2026-2028

Not quantified; power/interconnect availability cited as a key constraint customers must align before delivery

Mitigation:XPV platform and multi-year planning coordinate power and infrastructure alongside chip supply; extended lead times give Broadcom more visibility

Q&A highlights

How to square the ~$56B FY26 AI figure with 2x-H2 math (implying a Q4 sequential dip), and whether the 18-month AI backlog now sits at $200B+.

Hock confirmed H1 AI revenue was ~$19B, so 2x in H2 ties to ~$56B for FY26. He reiterated FY27 AI revenue in excess of $100B (~2x FY26) on the same trajectory as the back half of 2026, but declined to endorse the analyst's $200B backlog figure or guide '27 quarterly.

“you will easily see that 2027 will exceed very easily $100 billion in 2027... we're not trying to guide you every quarter what '27 would be like”

asked by Harlan Sur · answered by Hock Tan

3 min read 8 chapters

Detailed narrative

AI semiconductor ramp and the multi-year revenue trajectory

AI semiconductor revenue hit a record $10.8B in Q2, up 143% YoY and above outlook, and is guided to accelerate to $16B in Q3 (>200% YoY). Management expects H2 FY26 AI revenue to double H1 (~$19B), supporting a full-year FY26 figure of ~$56B (+~180% vs FY25).

Beyond that, FY27 AI revenue is reiterated at 'in excess of $100 billion' (~2x FY26) with FY28 expected to grow substantially further, underpinned by initiatives across six core customers. Hock characterized demand for XPUs and networking as 'simply insatiable.'

Six core customers and multi-gigawatt commitments

Broadcom detailed a widening AI customer roster: a long-term Google agreement for multiple TPU and AI-networking generations (April); Anthropic access to >1 GW of TPU compute in 2026 plus another 5 GW of next-gen TPU compute from 2027; OpenAI's contractual 1.3 GW in 2027 within a 10 GW-by-2029 agreement (silicon delivered, production late 2026); and a Meta MTIA partnership to deploy 3 GW through end-2028, with an initial 1 GW order (XPUs + networking) delivering H2 2027.

Two additional customers are set to ship late 2026 and have placed $6B of purchase orders to date. Management plans ~10 GW of shipments in 2027, back-half loaded.

The XPV funding platform with Apollo and Blackstone

To fund chips for frontier labs that may struggle to finance capacity themselves, Broadcom is creating an 'AI XPV platform' with Apollo, Blackstone and other large-balance-sheet investors to deploy more than 20 GW of compute through 2028. The first tranche, valued at $35B, is currently being launched by Apollo.

The vehicle is designed to deliver compute at the lowest cost and power to labs including Anthropic and OpenAI while keeping Broadcom in its pure chip-supplier model.

Networking leadership within the AI franchise

Networking represented almost 40% of Q2 AI revenue — a level Hock expects to normalize closer to 30% as XPU compute scales. Broadcom claims at least a one-generation lead: the 100-terabit Tomahawk 6 Ethernet switch has shipped for over a year, a next-gen 200-terabit switch tapes out this quarter, and the company is the de facto CPO standard (1.6T DSPs, CW and EML lasers).

Jericho3 and Jericho4 fabric solutions enable the largest multi-hyperscaler cross-data-center deployments. Networking carries richer margins than XPUs/ASICs, partly offsetting mix dilution.

Gross-margin mix dynamics

Consolidated gross margin fell 230 bps YoY to 77.1% and is guided to ~74% in Q3 as fast-growing, lower-margin AI/TPU revenue and semiconductors become a larger share of mix. Management stressed this is not a structural semiconductor-margin change: semiconductor gross margin held around 70% and operating leverage kept operating margin flat at ~67%.

Investors were explicitly urged to model semiconductor and software margins separately given shifting revenue mix.

Non-AI semiconductor cyclical recovery

Non-AI semiconductor revenue was $4.2B, up 6% YoY, with bookings exceeding $6B — which management cited as a clear indication of a path toward full cyclical recovery. Broadband, server storage and enterprise networking rose, partially offset by a seasonal decline in wireless. Q3 non-AI revenue is guided to ~$4.5B, up 12% YoY.

Infrastructure software and VMware momentum

Software revenue was $7.2B, up 9% YoY, with ARR sustaining 17% YoY growth and a 93% gross margin. The newly released VMware Cloud Foundation 9.1 adds heterogeneous compute support across AMD, Intel and NVIDIA platforms plus enterprise AI-inferencing support, and strong global server demand is driving robust on-prem private-cloud adoption.

Q3 software revenue is guided to ~$8.9B, up 31% YoY; management sees no negative agentic-AI impact given the hypervisor-level positioning.

Bookings, supply and visibility into 2028

AI bookings exceeded $30B in the quarter against $10.8B shipped, as customers order far ahead to secure not just wafers, HBM and DRAM but also power. Management says visibility now runs to 2028 (versus roughly 2027 three months ago). On supply, Hock said Broadcom has secured needs for 2026-2027 and is working on 2028-2029, and can generally meet incremental customer demand for wafers and HBM.

Inventory rose to 86 days (from 68 in Q1) as a deliberate build ahead of H2 AI acceleration.

Summary of the company's earnings call; see how call summaries are prepared. Not investment advice.