“{"visible_text": "Exhibit 8: Sensitivity of an annual \"required\" hyperscaler AI revenues to ROIC assumption. Annual \"required\" hyperscaler AI revenues based on average annual capex in 2026 and 2027 under different ROIC targets ($ billions). Y-axis: Annual \"required\" hypersc…”
关联内容单元 #190 · 原始来源原始全文
{"visible_text": "Exhibit 8: Sensitivity of an annual \"required\" hyperscaler AI revenues to ROIC assumption. Annual \"required\" hyperscaler AI revenues based on average annual capex in 2026 and 2027 under different ROIC targets ($ billions). Y-axis: Annual \"required\" hyperscaler AI revenues ($0 to $1,000). X-axis: Annual hyperscaler ROIC (0% breakeven, 10%, 20%, 30%). Hyperscalers include AMZN, GOOGL, META, MSFT, ORCL, and SPCX. \"Required\" AI revenues equal the sum of the estimated annual depreciation, annual operating expenses, and return on invested capital using gross capex, based on GS estimate of average annual AI capex in 2026 and 2027. See Sizing the AI Economy Needed to Justify ROIC on Hyperscaler AI Capex for more details. Source: Goldman Sachs Global Investment Research.", "image_type": "bar_chart", "companies": ["AMZN", "GOOGL", "META", "MSFT", "ORCL", "SPCX"], "industries": ["AI", "Cloud Computing", "Hyperscale Data Centers"], "key_numbers": {"required_AI_revenue_at_0pct_ROIC_breakeven_bn": 308, "required_AI_revenue_at_10pct_ROIC_bn": 417, "required_AI_revenue_at_20pct_ROIC_bn": 526, "required_AI_revenue_at_30pct_ROIC_bn": 636, "reference_year": "2026 and 2027", "revenue_uom": "$ billions"}, "table_data": [["Annual Hyperscaler ROIC", "Annual Required Hyperscaler AI Revenues ($ billions)"], ["0% (breakeven)", "$308"], ["10%", "$417"], ["20%", "$526"], ["30%", "$636"]], "chart_insights": "Required AI revenues rise linearly with the assumed ROIC target. Each 10-percentage-point increase in ROIC adds approximately $109-$110 billion to required revenues. Breakeven scenario requires $308B vs $636B at 30% ROIC, a ~2.1x difference. Implied annual AI capex (2026-2027 avg) underpins these revenue thresholds.", "key_information": "Goldman Sachs sensitivity analysis showing the AI revenue hyperscalers (AMZN, GOOGL, META, MSFT, ORCL, SPCX) must generate in 2026-2027 to justify their AI capex under different ROIC assumptions. Required revenue = depreciation + operating expenses + return on invested capital on gross capex. Even at breakeven (0% ROIC), $308B in AI revenues is needed, scaling to $636B at 30% ROIC.", "source_date_clues": {"source": "Goldman Sachs Global Investment Research", "reference_period": "average annual AI capex in 2026 and 2027", "report_topic": "Sizing the AI Economy Needed to Justify ROIC on Hyperscaler AI Capex"}, "research_relevance": "Critical for assessing whether hyperscaler AI investments can generate sufficient returns. Frames the magnitude of AI revenue opportunity needed to validate current capex commitments. Useful for evaluating AI monetisation prospects across major cloud/tech companies and the sustainability of current AI infrastructure spending pace."}AI 解读
Goldman Sachs has quantified the annual AI revenue thresholds hyperscalers must generate to justify their AI capex under varying ROIC targets, with breakeven requiring $308B and a 30% ROIC target requiring $636B in annual AI revenues, based on 2026-2027 average capex estimates.