--- title: "AVGO: AI firing on all cylinders; a fiercer rival to NVDA?---" type: "Topics" locale: "en" url: "https://longbridge.com/en/dolphin/post/39070789.md" description: "AVGO released Q1 FY26 results (ended Jan 2026) after-hours on Mar 5 (Beijing time). 1) Overall performance: revenue was $19.3bn, +29% YoY, in line with market expectations ($19.2bn).Revenue rose $1.3bn QoQ, mainly driven by AI. GPM was 68.1%. Excluding acquisition-related amortization and restructuring charges, Adj. GPM was 75.8%..." datetime: "2026-03-05T01:53:23.000Z" locales: - [en](https://longbridge.com/en/dolphin/post/39070789.md) - [zh-CN](https://longbridge.com/zh-CN/dolphin/post/39070789.md) - [zh-HK](https://longbridge.com/zh-HK/dolphin/post/39070789.md) author: "[Dolphin Research](https://longbridge.com/en/dolphin.md)" generator: "portal-rs" --- # AVGO: AI firing on all cylinders; a fiercer rival to NVDA?--- AVGO released Q1 FY26 results (ended Jan 2026) after the US market close on Mar 5 (Beijing time). Details follow. **1\. Overall results**: **Q1 revenue was $19.3bn (+29% YoY), in line with the Street ($19.2bn)**. QoQ rose by $1.3bn, driven by AI. GPM was 68.1%. **On an operating basis (ex acquisition amortization and restructuring), GPM was 75.8%, down 80bps QoQ**, reflecting mix shift as custom ASICs carry lower margins. **2\. Semiconductor**: **Revenue was $12.5bn, up $1.5bn QoQ**, with AI contributing most of the increment. Details: **①**$Broadcom(AVGO.US) **AI: $8.4bn, up $1.9bn QoQ, ahead of the Street ($8.2bn)**. Growth was led by higher Google TPU shipments across the quarter. **Google and Meta have recently raised 2026 capex outlooks, and AVGO expects AI revenue of $10.7bn next quarter, up $2.3bn QoQ**. **② Non-AI: $4.1bn, roughly flat YoY**. Non-AI trends were stable. **3\. Infrastructure software**: Revenue was $6.8bn, +1.4% YoY. Earlier growth stemmed from VMware integration and pricing changes (moving from perpetual licenses to subscriptions). **The M&A-driven surge has ended, and future growth will hinge on VMware’s subscription-led organic expansion**. **4\. Opex**: Core opex (R&D + S&M) was $3.98bn, down slightly QoQ. Scale benefits helped lower core opex ratio to ~21%. Over the past two years, stock-based comp has increased materially (now nearly half of opex). Ex SBC, core opex was $2.04bn, down $90mn QoQ. **5\. Inventory: $2.96bn this quarter, up 30% QoQ**. Unlike the usual single-digit QoQ moves, **the jump reflects strong demand and deliberate early stocking**. **6\. VMware integration**: Dolphin Research tracks Total Debt/LTM Adj. EBITDA, which fell further to 2 this quarter. The ratio has returned to pre-acquisition levels, indicating VMware’s debt impact has been absorbed within two years. **7\. Guidance**: **Q2 FY26 revenue is guided to around $22.0bn, above the Street ($20.8bn), with Adj. EBITDA margin at 68% vs. Street 67%. AI revenue is expected to rise to $10.7bn next quarter**. **Dolphin Research view: AI growth accelerates; explicit guidance lifts confidence** AVGO’s revenue and margins met expectations. Top-line growth was **primarily driven by AI**. **On an operating basis, GPM was 75.8%**, down slightly QoQ due to the larger mix of lower-margin ASICs. **For next quarter, revenue is guided to $22.0bn, up $2.7bn QoQ and above consensus ($20.8bn)**. Growth should be led by AI. **1) M&A impacts:** **Total Debt/LTM Adj. EBITDA fell to 2, back to pre-deal levels, indicating VMware’s debt impact was absorbed within two years**. **2) AI performance**: AI revenue was $8.4bn this quarter, up $1.9bn QoQ. **Management guides Q2 AI revenue of $10.7bn, up $2.3bn QoQ, above the Street ($9.7bn)**. **With VMware integration done, investor focus is squarely on AVGO’s AI trajectory**: **a) Hyperscaler capex is the engine**: Hyperscalers are the key buyers of AVGO custom ASICs, directly shaping AI expectations. **Based on guidance, Dolphin Research estimates combined 2026 capex of Google, Meta, Microsoft and Amazon could reach $660bn, with YoY growth over 60%, underpinning the AI chip market**. **b) AI chip share**: Nvidia remains dominant, with AVGO playing catch-up. **AVGO’s AI chip share is ~10% today and could rise toward ~20%, supported by demand from Google TPU and Anthropic**. **c) Product iteration**: AVGO has secured orders from Google, Meta and OpenAI. **Google’s TPU is the core product currently, and its performance will directly drive AI revenue**. TPUv7 is in mass production. Versus Nvidia GPUs, **TPUv7 in FP8 is broadly comparable to Nvidia’s B200 (mass production in Q4 2024), leaving Google roughly a year behind**. The biggest differences today: ① **Google supports up to FP8**, emphasizing stability and large-scale cluster efficiency. ② **Nvidia prioritizes speed, introducing NVFP4 in Blackwell to double inference throughput vs. FP8**. **Putting (a+b+c) together, hyperscalers are still lifting capex and AVGO’s share is rising**. Yet the stock has been muted, likely due to concerns about capex sustainability. **Meta guides 2026 capex of $115–135bn, over 50% of revenue, well above the prior 20–25% range**. **Google and Microsoft are also pushing capex to 40%+ of revenue**. Meta’s revenue growth is ~15–20%, while 2026 capex/revenue hits 50–60%, suggesting limited room to increase further. That implies hyperscaler capex growth may not be sustainable at current levels. **Hence, the market widely expects capex growth to slow sharply around 2027, ending the high-growth, multiple-expansion phase**. For AVGO in FY26, this is a period of earnings growth and valuation digestion, with the multiple now below 30x. **Industry headwinds weigh especially on Nvidia**. AVGO faces clearer risks: while it is a cost-down option for cloud operators, Google may diversify to Taiwan vendors or in-house. **For example, Google’s next-gen TPUv8 is being co-developed by AVGO and MediaTek (MTK). If legacy clients do not add orders, the market worries about share shifting to Taiwan alternatives**. **Order growth from existing clients is a key question**. Also, unlike Nvidia’s structurally rising margins, AVGO’s margin pressure from ASIC mix is another investor concern. **With a current mkt cap of ~$1.5tn, AVGO trades at ~27x PE on FY26 post-tax core operating profit** (assumes +74% YoY revenue, 72% operating GPM, 9.7% tax). AVGO’s valuation remains above Nvidia (~20x PE), reflecting ASIC share gains, even as sector multiples have corrected. **On forward growth, management provided a 2027 AI outlook**. Across six core AI customers (Google, Anthropic, Meta, OpenAI, etc.), combined compute demand approaches 10GW, implying over $100bn in AI revenue. **Explicit guidance should ease near-term worries and support sentiment**. More important will be management’s detail on Taiwan outsourcing vs. in-house and margin trajectory, the key overhangs on valuation. Dolphin Research’s detailed take on AVGO’s results follows below: **I. AVGO’s key businesses** Recent growth has been driven by AI and the VMware consolidation, making custom AI ASICs and VMware pricing strategy the focal points. By segment: **1) Semiconductor solutions**: Benefiting from AI revenue growth, mainly demand for custom ASICs from Google, Meta and ByteDance. Non-AI was largely flat. **AI**: Incremental growth mainly from Google TPU shipments. **AVGO now has six core AI customers (including Google, Anthropic, Meta, OpenAI, etc.), with OpenAI becoming a substantive customer**. **For 2027, the six customers’ combined compute demand is near 10GW**, implying $100bn+ in AI revenue. Anthropic exceeds 3GW, and OpenAI will ramp its first XPU (>1GW). **2) Infrastructure software**: With VMware consolidated, software is ~40% of revenue. AVGO raised VMware prices as it shifted from perpetual to subscription, which boosted revenue, but that impact is fading. **II. Consolidated results: AI is the core driver** **2.1 Revenue** **AVGO delivered Q1 FY26 revenue of $19.3bn (+29.5% YoY), in line with the Street ($19.2bn)**. YoY growth was mainly AI-driven. **QoQ, total revenue rose $1.3bn**, with AI adding $1.9bn QoQ and software down sequentially. **2.2 Gross profit** **Gross profit was $13.2bn, +30% YoY**. Reported GPM was 68.1%, up slightly QoQ. **On an operating basis (ex amortization and restructuring), GPM was 75.8%, down 80bps QoQ**, due to the higher mix of lower-margin custom ASICs. **2.3 Operating expenses** **Opex was $4.6bn, down slightly QoQ**. Ex SBC, **core opex (R&D + S&M) was $2.04bn, down $90mn QoQ**. Post VMware consolidation, AVGO’s opex optimization is largely complete. **2.4 Profit** **Net income was $7.35bn in Q1 FY26**. Dolphin Research prefers core operating profit (= GP - R&D - S&M) as a cleaner view. **Core operating profit was $10.65bn, up $900mn QoQ, driven by AI**. **2.5 EBITDA** AVGO emphasizes Adj. EBITDA% given its M&A-heavy model. **Dolphin Research estimates Q1 FY26 Adj. EBITDA% at 68%, vs. prior guidance of 67%**. Leverage improved: **Total Debt/LTM Adj. EBITDA fell to 2**, back to pre-deal levels on AI-driven EBITDA growth. This suggests VMware’s impact was digested within two years, and AVGO may resume pursuing new deals. **III. Segment detail: six customers could contribute $100bn in AI revenue** AVGO’s two pillars are semiconductor solutions and infrastructure software. Within these: **1) Semiconductor solutions**: networking (AI), wireless, storage connectivity, broadband, industrial & other. **2) Infrastructure software**: VMware, CA, Symantec, Brocade, etc. **3.1 Semiconductor solutions** **Q1 FY26 semiconductor revenue was $12.5bn, +52% YoY**. Growth was mainly AI-driven, with non-AI stable. **1) AI** **AI is the core earnings driver**. AI revenue was $8.4bn, up $1.9bn QoQ, re-accelerating on Google TPU shipments. **AI revenue is currently concentrated in three customers (Google, Meta, ByteDance)**. With Google and Meta lifting capex, AVGO guides Q2 AI revenue of $10.7bn, up $2.3bn QoQ. **AVGO’s ASIC roster now includes six disclosed customers: Google, Meta, ByteDance, Anthropic, a fifth customer (~$1bn), and newly OpenAI**. This indicates prior framework agreements are converting and OpenAI is now a substantive customer. Near term, focus remains on output from the three current customers, especially the ramp of TPUv6 and TPUv7. **Orders from Anthropic and the fifth customer (~$1bn) will ship in H2 2026, making full-year AI revenue back-half weighted**. Addressing growth uncertainty, **management outlined a 2027 AI outlook: with six core customers (Google, Anthropic, Meta, OpenAI, etc.), combined compute demand is near 10GW, implying $100bn+ in AI revenue**. Anthropic tops 3GW, and OpenAI will mass-produce its first XPU (>1GW). **The outlook should bolster near-term confidence**. That said, hyperscaler capex ratios are already elevated. Meta’s 2026 capex/revenue >50% limits further upside, which remains a key overhang for company and sector multiples. **2) Non-AI** **Non-AI semiconductor revenue was $4.1bn this quarter, roughly flat YoY**. The non-AI portfolio includes enterprise storage, broadband, wireless, and industrial & other. Enterprise networking, broadband and server storage grew YoY, offsetting seasonal weakness in wireless. **3.2 Infrastructure software** **Q1 FY26 infrastructure software revenue was $6.8bn, +1.4% YoY**. VMware integration impacts have been digested; focus shifts to organic growth. Software comprises VMware plus legacy CA, Symantec and Brocade. The legacy stack runs at roughly $2.0bn per quarter, **so the key watch is VMware**. **Dolphin Research estimates VMware contributed about $4.6bn this quarter**. License-to-subscription migration is now above 85%. As subscription penetration rises, VMware and software revenue should still grow, but not at M&A-integration pace. **With the leverage ratio down to 2, VMware has been absorbed**. AVGO is no longer breaking out VMware detail; AI is the primary focus. \ **Risk disclosure & statement:**[**Dolphin Research Disclaimer & General Disclosure**](https://support.longbridge.global/topics/misc/dolphin-disclaimer) ### Related Stocks - [GOOGL.US](https://longbridge.com/en/quote/GOOGL.US.md) - [META.US](https://longbridge.com/en/quote/META.US.md) - [AVGO.US](https://longbridge.com/en/quote/AVGO.US.md) - [NVDA.US](https://longbridge.com/en/quote/NVDA.US.md) - [GOOG.US](https://longbridge.com/en/quote/GOOG.US.md) --- > **Disclaimer: This article is for reference only and does not constitute any investment advice.**