Softbank Group Corp. · TSE:9984
Softbank — Arm Holdings — Q2 FYE25 Earnings Presentation
10 slides from Softbank's earnings presentation, November 2024
Slide 1: Arm Holdings plc SoftBank Investor Briefing
Arm Holdings plc SoftBank Investor Briefing Ian Thornton Vice President, Investor Relations AI-generated image November 2024 © 2024 Arm
Slide 2: Nomenclature difference between Arm and SoftBank Group
Nomenclature difference between Arm and SoftBank Group Fiscal year ending March 31, 2025 Fiscal year ending March 31, 2025 FY2024 FYE25 © 2024 Arm 2
Slide 3: Q2 FYE25: From Revenues to Profits
Q2 FYE25: From Revenues to Profits 97% 39% Non-GAAP Non-GAAP Gross Margin Operating Margin $844m $24m $820m License Revenue $330m $494m $326m $317m Royalty Revenue $514m Revenue Non-GAAP Non-GAAP Non-GAAP Non-GAAP Non-GAAP COS Gross Profit Operating Operating Net Income Expense Income Notes: 1. Depr
Slide 4: Semiconductor industry recovered after inventory correction
Semiconductor industry recovered after inventory correction After slow start to the year, semiconductor industry has recovered, growing 23% YoY in September to record levels Industry demand driven by AI and consumer. Automotive and industrial electronics are weaker. Smartphone market grew 4.0% YoY i
Slide 5: Armv9 adoption driving royalty growth
Armv9 adoption driving royalty growth Royalty Revenue by Architecture 100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% FYE16 FYE17 FYE18 FYE19 FYE20 FYE21 FYE22 FYE23 FYE24 Q1 FYE25* Q2 FYE25* Armv7 & older Armv8 Armv9 Armv9 commands a higher royalty per chip than prior architectures Armv9 adoption has s
Slide 6: Compute Subsystems are further increasing royalty per chip
Compute Subsystems are further increasing royalty per chip Arm’s Compute Subsystems (CSS) are being Arm is building an ecosystem adopted across all target markets to support CSS customers Partners have taped out multiple chips and chiplets combining Arm CSS with products and services across the ecos
Slide 7: Accelerating AI on Arm-based devices (Smartphone)
Accelerating AI on Arm-based devices (Smartphone*) Prompt Processing Token Generation Llama 3.2 3B WikiText-2 PPL (prompt = 128 tokens) (output = 128 tokens) 100 18 10 90 16 9 Small 80 14 8 degradation ~5x improvement 70 from Arm CPU ~3x improvement 7 12 60 optimized kernel 6 Perplexity token/sec to
Slide 8: Accelerating AI on Arm-based devices (Server)
Accelerating AI on Arm-based devices (Server*) Prompt Processing Token Generation Llama 3 70B WikiText-2 PPL (prompt = 128 tokens) (output = 128 tokens) 120 120 4 3.5 Small 100 100 degradation 3 80 80 2.5 tokens/sec tokens/sec >5x improvement ~3x improvement Perplexity 60 from Arm CPU 60 optimized k
Slide 9: Near-Term Guidance
Near-Term Guidance Q3 FYE 25 FYE 25 Revenue ($m) $920m - $970m $3.80b - $4.10b Non-GAAP Operating Expense ($m)1 ~$525m ~$2.05b Non-GAAP fully diluted earnings per share ($)1 $0.32 - $0.36 $1.45 - $1.65 (1) For more information and definitions of the non-GAAP measures see the “Key Financial and Opera
Slide 10: To be replaced
To be replaced © 2024 Arm 10