Google Cloud 在 2026 年 Q2 同比增长 82%，达到 248 亿美元，超过 223 亿美元的市场一致预期。仅这个数字就足以成为任何财报的头条。更重要的信号在于它所趋同的方向：Google Cloud 的增速如今已与 NVIDIA 如出一辙。
Google 本季度确实开始销售硬件了。CFO Anat Ashkenazi 告诉投资者，Google“开始确认 TPU 系统销售的收入，我们在 Q2 首次将这些系统交付至客户数据中心。”1 但她明确表示，这笔金额很小：云业务增长“即使在剔除 TPU 系统销售的影响后仍显著加速”，而 TPU 收入的大部分将在 2027 年落地。这种趋同另有来源。Google Cloud 按小时出租 AI 算力，NVIDIA 则直接销售芯片，而这两条收入线都源自同一需求的下游——训练和运行 AI 模型。
这张图展示了这种趋同。Google Cloud 与 NVIDIA 的数据中心业务板块如今勾勒出同一条曲线，一条因租赁收入而上扬，另一条因整机销售而上扬，NVIDIA 在规模上领先。
订单储备也指向同一方向。Google Cloud 的积压订单达到 5140 亿美元，较一年前的 1060 亿美元增长 385%，单季度增长超过 500 亿美元，其中略超一半将在 24 个月内转化为收入。与其他超大规模云厂商相比，四大云服务商目前合计持有超过 2 万亿美元的合同需求。
此外，Google 还享有规模经济。云业务营业利润为 88 亿美元，同比增长超过两倍，营业利润率从一年前的 20.7% 扩大至 35.6%。
与 AWS 相比，利润率差距正在迅速缩小。
| 分部 | 营业利润 | 营收 | 营业利润率 | 利润率同比变化 |
|---|---|---|---|---|
| AWS | $14.2b | $37.6b | 37.7% | -1.8 pts |
| Google Cloud | $8.8b | $24.8b | 35.6% | +14.9 pts |
时间口径有所不同：AWS 尚未公布 2026 日历年第二季度数据，因此其数字为 2026 年第一季度，而 Google Cloud 的为 2026 年第二季度。AWS 的绝对美元规模仍然更大，但其利润率同比下滑近两个百分点，而 Google Cloud 的利润率则扩大近十五个百分点，将一年前 19 个百分点的差距缩小到仅剩两个百分点。
资本投资的回报率清晰可见，因此预计资本支出将进一步增加。Google 将 2026 年全年资本支出指引从 $180b 至 $190b 上调至 $195b 至 $205b，并表示预计资本支出“将在 2027 年大幅增长”。综合 Amazon、Microsoft、Meta 和 Oracle 来看，行业 AI 基础设施支出明年将远超 $1t。
我们从未见过规模化企业能以如此之快的速度增长，且盈利能力如此之强。
-
Alphabet 2026 年第二季度财报电话会议，2026 年 7 月 22 日。所有引用的数字与言论均来自该会议。此框架最早由 SemiAnalysis (@SemiAnalysis_) 提出。↩︎
Google Cloud grew 82% year-over-year to $24.8b in Q2 2026, beating the $22.3b consensus. The number alone would headline any earnings report. The more important signal is what it converges toward: Google Cloud’s growth rate now mirrors NVIDIA’s.
Google did start selling hardware this quarter. CFO Anat Ashkenazi told investors Google “began to recognize revenues from TPU system sales, which we delivered to customer data centers for the first time in Q2.”1 But she was clear the dollars were small: cloud growth “accelerated meaningfully even after excluding the impact of TPU system sales,” with the bulk of TPU revenue landing in 2027. The convergence has a different source. Google Cloud rents AI compute by the hour, NVIDIA sells the chips outright, & both revenue lines are downstream of the same demand to train & run AI models.
The chart shows the convergence. Google Cloud & NVIDIA’s Data Center segment now trace the same curve, one bending up on rental revenue, the other on unit sales, NVIDIA ahead on scale.
The pipeline points the same direction. Google Cloud’s backlog reached $514b, up 385% from $106b a year ago & up more than $50b in a single quarter, with just over half set to convert to revenue within 24 months. Set against the other hyperscalers, the four largest cloud providers now carry more than $2t in contracted demand.
Plus, Google enjoys economies of scale. Cloud operating income was $8.8b, more than tripling year-over-year, & operating margin expanded from 20.7% a year ago to 35.6%.
Set against AWS, the margin gap is closing fast.
| Segment | Operating income | Revenue | Operating margin | Margin change YoY |
|---|---|---|---|---|
| AWS | $14.2b | $37.6b | 37.7% | -1.8 pts |
| Google Cloud | $8.8b | $24.8b | 35.6% | +14.9 pts |
The timing differs: AWS has not yet reported calendar Q2 2026, so its figures are Q1 2026, while Google Cloud’s are Q2 2026. AWS remains larger in absolute dollars, but its margin slipped nearly two points year-over-year while Google Cloud’s expanded almost fifteen, narrowing a gap that was 19 points a year ago to barely two.
The ROI on capital investment is clear, so more capex is projected. Google raised full-year 2026 capex guidance to $195b to $205b, up from $180b to $190b, & said it expects capex “to increase significantly in 2027.” Across Amazon, Microsoft, Meta, & Oracle, industry AI infrastructure spend is on a path well beyond $1t next year.
We’ve never seen businesses at scale grow this fast & this profitably.
-
Alphabet Q2 2026 earnings call, July 22, 2026. All quoted figures & remarks from the . Framing first flagged by SemiAnalysis (@SemiAnalysis_). ↩︎