II 全球机构情报
盛宝银行 · 2026/09/02

推动人工智能发展的公司表现优于构建人工智能的公司

前往官网原文 ↗
自动质量提示本研报仍可阅读,但 AI 分析或译文低于优选质量阈值,已进入改进队列。重要判断请同时核对官网原文。
完整研报正文
完整中文译文

推动人工智能发展的公司表现优于构建人工智能的公司

要点:

能源股的表现明确优于大型科技股:地缘政治供应中断以及炼油和天然气利润率飙升正在带来即时的现金流收益,而超大规模企业则面临不断攀升的人工智能投资成本以及未来回报的不确定性。

盈利差距反映了不同的投资周期:科技公司正在向人工智能基础设施投入前所未有的资金,而相关收入和资本回报的大部分仍需要数年时间才能实现。

能源公司正在将稀缺性变现:战争、制裁和贸易流中断推高了原油、成品油和天然气的价格,立即提升了现有资产的现金流。

炼油业务变得异常有利可图:美国3-2-1裂解价差已达到每桶63美元,而欧洲利润率更强劲,反映出柴油价格超过每桶200美元以及天然气成本高昂。

由人工智能投资热潮推动的科技行业继续实现强劲的收入增长,但今年一些最强劲的股票回报却来自那些为构建和运营其背后基础设施提供能源的公司。

这种对比十分鲜明。表中列出的七大能源巨头——埃克森美孚、挪威国家石油公司、康菲石油、壳牌、道达尔能源、雪佛龙和英国石油——平均同比回报率为38.5%,而“科技七巨头”股票的平均同比回报率为18%。今年迄今的表现尤为强劲,七只能源股均至少上涨25%,其中挪威国家石油公司领涨71%,而“科技七巨头”的表现则从苹果上涨19.6%到特斯拉下跌20.8%不等。请注意,这些是简单的等权计算,并非正式指数回报,且过往表现不代表未来结果。

这种分化并不意味着人工智能故事正在失败。科技盈利仍然强劲,但维持这种增长的成本正在迅速上升。Meta第二季度收入同比增长28%,达到60.8亿美元,但资本支出达到31.1亿美元,自由现金流降至仅784百万美元。该公司目前预计全年资本支出为130–145亿美元。

亚马逊也提供了类似的例证。AWS营业收入从一年前的10.2亿美元跃升至16.6亿美元,但过去12个月的自由现金流转为7.6亿美元的流出,因为物业和设备的购买增加了66.1亿美元,主要反映对人工智能基础设施的投资。

超大规模企业能够为这些项目提供资金,但投资者越来越关注最终回报的时间和规模。今天的许多支出必须在未来人工智能需求、利用率和定价能力能够被确知之前投入。

对于“能源七巨头”来说,其经济逻辑几乎相反。大宗商品价格上涨立即使得在运营中的生产、加工和基础设施资产的收入和现金流提升。多年来的克制投资限制了闲置产能,而战争、制裁和航运路线中断减少了原油、成品油和天然气的供应。

后果已体现在盈利中。埃克森美孚公布第二季度盈利为14.5亿美元,是去年同期7.1亿美元的两倍多。经营活动现金流达到23.6亿美元,自由现金流为17.2亿美元,得益于价格上涨、利润率提高以及创纪录的第二季度柴油产量。雪佛龙盈利12.1亿美元,资本回报率达到21%,并报告了创纪录的美国产量和炼油厂吞吐量。

Equinor是欧洲争夺能源安全的明显受益者。其第二季度调整后营业收入达到11.5亿美元,而净营业收入比去年同期增长了一倍多。该公司在欧洲天然气实现价格为每百万英热单位15.8美元,液体燃料实现价格为每桶97.9美元,原油交易和炼油业务的强劲表现提供了额外支撑。

炼油资产已成为另一个巨额利润来源。美国3-2-1裂解价差——将三桶原油转化为两桶汽油和一桶馏分油的理论利润——目前接近每桶63美元,而去年同期约为每桶25美元。对于能够保持高开工率的炼油商来说,这意味着在扣除运营和其他成本前有可观的利润率。

欧洲的情况更为紧张,柴油价格已升至每桶200美元以上,比布伦特原油高出105美元,远高于去年同期28美元的价差水平。这一溢价反映了俄罗斯和中东炼油厂供应中断,以及该地区炼油能力有限和欧洲对进口原油及成品油的持续依赖。因此,结合上游生产、交易业务和炼油资产的公司正从价值链的多个环节获益。

欧洲天然气提供了同样引人注目的例子。欧盟天然气交易价格接近每桶油当量142美元,比布伦特原油高出45美元,约为美国天然气价格的八倍。欧洲本土供应短缺和对海运液化天然气(越来越多来自美国,而中东供应接近停摆)的依赖,造成了巨大的区域溢价,为Equinor等供应商带来了意外之财。

因此,表现差距反映的不仅仅是资金从科技转向传统能源的暂时轮动。七巨头今天正花费巨额资金支撑预期中的未来收益。而“能源七巨头”正从稀缺性、高价格和现有基础设施中收获非凡的现金流。

人工智能仍是一个强大的结构性增长主题。然而,其对电力、天然气、备用发电和电网基础设施日益增长的需求意味着,其最直接的利润不仅流向了技术开发者,也流向了那些提供运行技术所需分子和电子的公司。

如需市场评论和见解——而非交易建议——请关注我,并在Twitter和Substack上参与讨论

完整英文原文

Key Points:

Energy shares have decisively outperformed Big Tech: Geopolitical supply disruptions and soaring refining and gas margins are delivering immediate cash-flow gains, while hyperscalers face mounting AI investment costs and uncertainty over future returns.

The earnings gap reflects different investment cycles: Technology companies are committing unprecedented sums to AI infrastructure, while much of the associated revenue and return on capital remains years away.

Energy companies are monetising scarcity today: Wars, sanctions and disrupted trade flows have lifted crude oil, refined products and natural gas prices, immediately boosting cash flow from existing assets.

Refining has become exceptionally profitable: The US 3-2-1 crack spread has reached USD 63 per barrel, while even stronger European margins reflect diesel above USD 200 per barrel and expensive natural gas.

The technology sector fuelled by the artificial-intelligence investment boom continues to generate strong revenue growth, but some of the strongest equity returns this year have come from the companies providing the energy needed to build and operate the infrastructure behind it.

The contrast is striking. The seven energy majors shown in the table - ExxonMobil, Equinor, ConocoPhillips, Shell, TotalEnergies, Chevron and BP - have delivered an average year-over-year return of 38.5%, compared with 18% for the Magnificent Seven technology stocks. The performance has been particular strong so far this year with all seven energy stocks having gained at least 25%, led by Equinor at 71%, while performance among the Magnificent Seven ranges from Apple’s 19.6% rise to Tesla’s 20.8% decline. Please note, that these are simple equal-weight calculations, not formal index returns, and that past performance is not indicative of future results.

The divergence does not mean the AI story is failing. Technology earnings remain formidable, but the cost of maintaining that growth is escalating rapidly. Meta’s second-quarter revenue rose 28% to USD 60.8 billion, yet capital expenditure reached USD 31.1 billion and free cash flow fell to just USD 784 million. The company now expects full-year capital spending of USD 130–145 billion.

Amazon offers a similar illustration. AWS operating income jumped to USD 16.6 billion from USD 10.2 billion a year earlier, but trailing 12-month free cash flow swung to an outflow of USD 7.6 billion as purchases of property and equipment increased by USD 66.1 billion, primarily reflecting investment in AI infrastructure.

The hyperscalers can finance these programmes, but investors are increasingly focused on the timing and scale of the eventual returns. Much of today’s spending must be committed before future AI demand, utilisation rates and pricing power can be known with confidence.

For the Energy Seven, the economics are almost the reverse. Higher commodity prices immediately lift revenue and cash flow from producing, processing and infrastructure assets already in operation. Years of restrained investment have limited spare capacity, while wars, sanctions and disrupted shipping routes have reduced the availability of crude oil, refined products and natural gas.

The consequences are already visible in earnings. ExxonMobil reported second-quarter earnings of USD 14.5 billion, more than double the USD 7.1 billion earned a year earlier. Cash flow from operations reached USD 23.6 billion and free cash flow USD 17.2 billion, supported by higher prices, stronger margins and record second-quarter diesel production. Chevron earned USD 12.1 billion, achieved a 21% return on capital employed and reported record US production and refinery throughput.

Equinor has been a particularly clear beneficiary of Europe’s scramble for secure energy. Its second-quarter adjusted operating income reached USD 11.5 billion, while net operating income more than doubled from a year earlier. The company realised USD 15.8 per MMBtu for European gas and USD 97.9 per barrel for liquids, with strong crude trading and refining performance providing additional support.

Refining assets have become another major source of exceptional profits. The US 3-2-1 crack spread - the theoretical margin from converting three barrels of crude into two barrels of gasoline and one barrel of distillate - is trading near USD 63 per barrel, compared with around USD 25 per barrel this time last year. That implies a remarkable margin before operating and other costs for refiners able to maintain high utilisation.

Conditions are even tighter in Europe, where diesel has risen above USD 200 per barrel, more than USD 105 above Brent and well above the USD 28 per barrel at which the spread traded this time last year. The premium reflects disrupted supply from Russia and Middle East refiners as well as limited regional refining capacity and Europe’s continued dependence on imported crude and products. Companies combining upstream production, trading operations and refining assets are therefore benefiting at several stages of the value chain.

European natural gas provides an equally dramatic example. EU gas is trading near USD 142 per barrel of oil equivalent, more than USD 45 above Brent and around eight times the prevailing US natural gas price. Europe’s shortage of indigenous supply and dependence on seaborne LNG, increasingly sourced from the US as Middle Eastern supplies remain close to offline, have created a substantial regional premium, handing a windfall to suppliers such as Equinor.

The performance gap therefore reflects more than a temporary rotation from technology into traditional energy. The Magnificent Seven are spending vast sums today to support earnings expected tomorrow. The Energy Seven are harvesting exceptional cash flow today from scarcity, elevated prices and infrastructure that already exists.

AI remains a powerful structural growth theme. However, its expanding requirement for electricity, natural gas, backup generation and grid infrastructure means some of its most immediate profits are accruing not only to those developing the technology, but also to those supplying the molecules and electrons needed to run it.

For market commentary and insights - not trading advice - follow me and join the conversation on Twitter and Substack

预览 PDF
1 / 110%

正在载入文档……

AI 分析
由 AI 依据上文研报生成 · 非原文直译、非机构原话 · 重要判断请核对官网原文
关键论点
  • 能源股跑赢大型科技股:七大能源巨头平均年回报率38.5%,而科技七巨头为18%。
  • 能源公司正在变现当前的稀缺性:战争、制裁和贸易中断推高原油、成品油和天然气价格。
  • 炼油利润率飙升:美国3-2-1裂解价差接近63美元/桶,欧洲柴油超过200美元/桶。
  • 欧洲天然气溢价:欧盟天然气接近142美元/桶油当量,较布伦特原油高约45美元,利好Equinor等供应商。
  • 超大规模企业的AI资本支出沉重:Meta第二季度资本支出311亿美元,自由现金流降至7.84亿美元;亚马逊TTM自由现金流转为-76亿美元。
  • AI的电力与基础设施需求可能使能源供应商获益。
风险
  • AI投资可能无法产生预期回报,影响科技估值。
  • 如果地缘政治紧张局势缓解或供应正常化,能源价格可能下跌。
  • 炼油利润率具有周期性,可能回归历史平均水平。
  • 监管变化可能影响能源生产和贸易。
  • 高能源价格可能抑制全球经济增长。