别谈有机增长,AI迫使印度中型IT企业重写并购剧本

在 Coforge 第二季度财报电话会上,一位分析师请首席财务官 Saurabh Goel 解释合并增长与内生增长之间的差距。按固定汇率计算,营收环比增长 10.9%,而内生增长为 3.8%,这表明收购在多大程度上补足了公司的增长。
Goel 将这一差异部分归因于第二季度合并了 Encora 整整一个季度的营收,而第一季度仅合并了两个月。Coforge 以 22090 亿卢比收购的数字工程公司 Encora,贡献了 13.82 亿至 13.91 亿卢比营收。在第一季度计入的两个月里,该公司贡献了约 9.4 亿卢比。
第二季度不仅 Coforge 忙碌,Cyient 和 Mphasis 也同样如此,后者收购了数据工程与决策智能公司。与此同时,Persistent 预计将在 2027 年第一季度完成对 Nagarro 的收购。
它们的业绩显示出这些交易正如何影响营收、利润率和订单储备。
各公司正通过不同指标来披露 AI 相关业务。Coforge 表示,其 AI 主导的工程、数据集成和云业务贡献了合并营收的 86%。Mphasis 称,AI 主导的交易占其 43.8 亿卢比净新增合同总价值(TCV)的 66%。Zensar 报告称,受 AI 影响的交易占其订单量的 36%,而 Sonata Software 则将 AI 主导的订单预订量列为 25%。
中型企业将不得不建立专业领域的能力。收购专业公司是实现这一目标的一种方式,Gartner 杰出副总裁分析师 Milind Govekar 对 AIM 表示。“它们中的大多数是在通过收购获取新技能、新技术,”他说。很少有公司收购主要是为了消灭竞争对手或获取客户。
Coforge 的进展最快。它已将 Encora 整合进自身业务,公司不再区分这两个实体。Coforge 的合并 EBIT 利润率为 16%,而其独立业务为 17%。整合成本达到约 ₹160 crore,超过公司 ₹141 crore 至 ₹150 crore 的指引。
此外,Cyient 以 ₹2,049 crore 收购了 AI 原生数据工程公司 TAO Digital。TAO 在并入后的第一个月贡献了 ₹90 crore。剔除该收购,按固定汇率计算,有机收入环比增长 1.6%,同比增长 0.4%。
CFO Shrinivas Kulkarni 在财报电话会上表示,TAO 的年化收入运行率已从收购时的约 8000 万美元(₹752 crore)升至近 1 亿美元(₹940 crore)。他预计未来 12 个月将达到 ₹1,034 crore,同时提醒不应根据单月收入进行外推。Cyient 目前计划将 TAO 的 AI 数据工程能力推广到现有客户。该公司本季度的技术订单量增长 39.7%。
Persistent 拟收购 Nagarro 是本文讨论的最大一笔交易。该公司获得了 Nagarro 约 94.04% 的股本和投票权,其中不包括库存股,但包括其现有持股。该交易尚待 RBI 批准以及奥地利和西班牙的外国投资许可。
此次收购将借助 Nagarro 在罗马尼亚、葡萄牙、匈牙利和波兰的交付中心,扩大 Persistent 在欧洲的业务版图。以美元计算,Persistent 第二季度欧洲收入同比下降 8.5%,公司将这归因于两起与特定客户相关的事件。北美收入增长 22.5%。
Persistent 在第二季度计入了与 Nagarro 相关的 1.85 亿卢比交易成本,包括咨询、融资和法律费用。其 PAT 利润率为 7.5%,运营 PAT 利润率为 10.6%。该公司预计未来两个季度还将产生 2.35 亿至 2.82 亿卢比成本,使累计一次性费用达到约 4.2 亿至 4.67 亿卢比。
Mphasis 也通过收购来增加特定能力和客户关系。该公司于 4 月收购了加拿大决策智能公司 Theory and Practice Business Intelligence(TAP),并预计在 11 月完成对 IT 人力公司 Red Oak Tech 客户合同的收购,价值最高达 2.65 亿卢比。CFO Aravind Viswanathan 在财报电话会上将 TAP 相关的摊销列为利润率的压力因素。
Govekar 表示,客户可能更青睐行业特定的 AI 应用。“人们不会买通用 AI。他们会买特定 AI。”
印度 IT 服务商通常在前沿 AI 公司开发的大语言模型之上进行构建。Gartner 副总裁分析师 DD Mishra 表示,它们的优势在于理解客户的运营,并将这些模型应用于特定的业务需求。
Coforge 和 Persistent 的高管在各自的财报电话会上提出了类似论点。Singh 表示,Coforge 的大型交易涉及为单个企业量身定制的解决方案。Persistent CTO Nitish Shrivastava 表示,差异化正从模型本身转向模型周边的上下文。
这种重应用而非重所有权的倾向,在此前与 AIM 的对话中也有所体现。Mphasis 全球交付主管 Ravi Vasantraj 表示,拥有知识产权本身并非主要考量,而 Coforge 首席 AI 商务官 Anup Nair 表示,企业客户未必想要专有平台。
然而,印度 IT 公司并非唯一在拓展企业级 AI 业务的服务商。本季度,Persistent 成为 Anthropic 的 Claude 合作伙伴网络中的首选合作伙伴。Birlasoft 则被指定为 OpenAI Select Partner。
这些合作让两家公司得以接入上述实验室的模型与合作伙伴生态,同时也让印度 IT 服务商与一些公司走得更近——这些公司最终可能在同一批企业实施与转型业务上展开竞争。
根据 Gartner 的印度 IT 支出预测,其预计 2027 年印度 IT 服务支出将增长 6.2%,而软件支出将增长 19.3%。
Coforge 将 2027 财年的 EBIT 利润率指引从 15.5% 上调至 16%,并目标在第四季度达到 17%。以卢比计算,营收同比增长 53.6%,而一般及行政费用上升 15.6%。
Singh 表示,Coforge 不会把所有生产率提升带来的节省都让渡给客户。他说:“在某些交易中,你必须把一部分让给客户。”在另一些交易中,公司则保留这些节省。
Persistent 正在试验与成果挂钩的定价模式,不过一笔大额交易的成本拖累了利润率。其 EBIT 利润率降至 15%,环比下降 100 个基点,同比下降 130 个基点。
这份为期 6.5 年、价值 ₹6,110 crore 的 Nagarro 交易涵盖 20 款软件的工程服务,包括第三方许可与云成本。它每季度增加约 ₹38 crore 的摊销,并使利润率下降约 90 个基点。Kalra 描述了一种收入分成安排,并设有与客户保证成本节省挂钩的兜底条款。
Mphasis 将 EBIT 利润率维持在 14.8%,接近其 14.75%–15.75% 区间的下限。早期 AI 项目需要更多现场工程师,推高了交付成本。
Rakesh 表示,较短期的合同也能加快收入转化。一份为期六个月的合约所能带来的近期收入,可能与一份 TCV 高出 10 倍的五年期合同相当,因为较短合约的收入是在小得多的期间内确认的。Zensar 还预计,随着 AI 采用加速,交易规模将进一步缩小。
Mishra 表示,客户仍在检验 AI 部署能否带来预期回报。一些客户在部署未获成功后回退了 AI 智能体,并对供应商进行了处罚。Coforge 报告称,其在 225 家以上客户中开展了 520 多个 AI 项目,其中 260 个智能体处于实际运行状态。
收购让中型 IT 企业获得了专业能力和客户关系,而 AI 正在改变它们交付和定价服务的方式。它们的第二季度业绩显示出整合这些业务的成本,以及随着企业 AI 采用扩大,寻找新的差异化来源所面临的压力。
译文由英文原文机器翻译,术语按行业惯例统一,可在文末展开原文核对。
查看英文原文(Forget Organic Growth, AI is Forcing Mid-Cap Indian IT to Reinvent M&A Playbook)
During Coforge’s Q2 earnings call, an analyst asked Chief Financial Officer Saurabh Goel to explain the gap between consolidated and organic growth. Revenue grew 10.9% sequentially in constant currency, against organic growth of 3.8%, indicating how much acquisitions are supplementing the company's growth.
Goel attributed the difference partly to the consolidation of a full quarter of Encora's revenue in Q2, compared to only two months in Q1. Encora, the digital engineering firm Coforge acquired for ₹22,090 crore, contributed ₹1,382–1,391 crore in revenue. It contributed about ₹940 crore during the two months included in Q1.
Q2 was busy not just for Coforge, but also for Cyient and Mphasis, which acquired data engineering and decision intelligence firms. Meanwhile, Persistent is expected to complete its acquisition of Nagarro in the first quarter of 2027.
Their results show how these deals are affecting revenue, margins, and deal pipelines.
Companies are reporting AI activity through different metrics. Coforge said its AI-led engineering, data integration, and cloud businesses contributed 86% of consolidated revenue. Mphasis said AI-led deals accounted for 66% of its ₹4,380 crore in net new total contract value (TCV). Zensar reported that AI-influenced deals made up 36% of its order book, while Sonata Software put AI-led order bookings at 25%.
Acquisitions Buy Skills More Than AI
Mid-tier companies will have to build specialist expertise. Acquiring specialist firms is one way to do so, Gartner’s Distinguished Vice President Analyst Milind Govekar tells AIM. “Most of them are acquiring to get new skills, new technology,” he says. Few are buying primarily to eliminate rivals or gain customers.
Coforge is furthest along. It has already integrated Encora into its business, and the company no longer distinguishes between the two entities. Coforge's consolidated EBIT margin stood at 16%, against 17% for its standalone business. Integration costs reached about ₹160 crore, exceeding the company's guidance of ₹141 crore to ₹150 crore.
Also, Cyient acquired AI-native data engineering firm TAO Digital for ₹2,049 crore. TAO contributed ₹90 crore in its first month. Excluding the acquisition, organic revenue grew 1.6% sequentially and 0.4% year-on-year in constant currency.
CFO Shrinivas Kulkarni said during the earnings call that TAO's annual revenue run rate had risen from about $80 million (₹752 crore) at acquisition to nearly $100 million (₹940 crore). He expects it to reach ₹1,034 crore over the next 12 months, while cautioning against extrapolating from one month's revenue. Cyient now plans to take TAO's AI data engineering capabilities to existing customers. The company's technology order intake rose 39.7% during the quarter.
Persistent's proposed acquisition of Nagarro is the largest deal discussed here. It secured approximately 94.04% of Nagarro's share capital and voting rights, excluding treasury shares and including its existing holding. The deal awaits RBI approval and foreign investment clearances in Austria and Spain.
The acquisition would expand Persistent's European presence through Nagarro's delivery centres in Romania, Portugal, Hungary, and Poland. Persistent's European revenue fell 8.5% year-on-year in Q2 in US dollar terms, which it attributed to two client-specific events. North American revenue grew 22.5%.
Persistent booked ₹185 crore in Nagarro-related transaction costs in Q2, including advisory, financing, and legal expenses. Its PAT margin was 7.5% and operating PAT margin 10.6%. It expects another ₹235–282 crore in costs over the next two quarters, taking cumulative one-time expenses to approximately ₹420–467 crore.
Mphasis has also pursued acquisitions to add specific capabilities and customer relationships. It acquired Canadian decision-intelligence firm Theory and Practice Business Intelligence (TAP) in April and expects to complete its acquisition of IT staffing firm Red Oak Tech's customer contracts, worth up to ₹265 crore, in November. CFO Aravind Viswanathan cited TAP-related amortisation as a pressure on margins during the earnings call.
Where IT Can Differentiate
Customers are likely to favour industry-specific AI applications, Govekar says. “People won’t buy generic AI. They will buy specific AI.”
Indian IT providers generally build on large language models developed by frontier AI companies. Gartner Vice President Analyst DD Mishra says their advantage lies in understanding customers' operations and applying those models to specific business needs.
Coforge and Persistent executives made similar arguments during their respective earnings calls. Singh said Coforge’s large deals involve solutions tailored to individual enterprises. Persistent CTO Nitish Shrivastava said differentiation is shifting from the models themselves to the context around them.
This emphasis on application over ownership also came through in earlier conversations with AIM. Mphasis Global Delivery Head Ravi Vasantraj said owning the intellectual property itself is not the main consideration, while Coforge Chief AI Commercial Officer Anup Nair said enterprise customers do not necessarily want proprietary platforms.
However, Indian IT companies are not the only providers pursuing enterprise AI engagements. Persistent became a preferred partner in Anthropic's Claude Partner Network during the quarter. Birlasoft was named an OpenAI Select Partner.
The partnerships give both companies access to the labs' models and partner ecosystems, while bringing Indian IT providers closer to companies that could eventually compete for the same enterprise implementation and transformation work.
Pricing Pressure Tests the Model
Gartner expects IT services spending in India to grow 6.2% in 2027, against 19.3% for software, according to its India IT spending forecast.
Coforge raised its FY27 EBIT margin guidance from 15.5% to 16% and is targeting 17% by Q4. Revenue grew 53.6% year-on-year in rupee terms, while general and administrative costs rose 15.6%.
Singh said Coforge does not pass all productivity savings to customers. “In some deals, you have to pass some to the client,” he said. In others, it retains the savings.
Persistent is experimenting with outcome-linked pricing, although costs from a large deal have weighed on margins. Its EBIT margin fell to 15%, down 100 basis points sequentially and 130 basis points year-on-year.
The 6.5-year, ₹6,110 crore Nagarro deal covers engineering for 20 software products, including third-party licence and cloud costs. It added about ₹38 crore in quarterly amortisation and reduced margins by about 90 basis points. Kalra described a revenue-sharing arrangement with a backstop tied to assured customer cost savings.
Mphasis held its EBIT margin at 14.8%, near the bottom of its 14.75–15.75% range. Early AI programmes have required more on-site engineers, raising delivery costs.
Shorter contracts can also accelerate revenue conversion, Rakesh said. A six-month engagement can generate as much near-term revenue as a five-year contract with 10x the TCV, since the revenue from the shorter engagement is recognised over a much smaller period. Zensar also expects deal sizes to shrink further as AI adoption accelerates.
Clients are still testing whether AI deployments deliver expected returns, Mishra says. Some have rolled back AI agents after unsuccessful deployments and penalised providers. Coforge reported more than 520 AI engagements across over 225 clients, with 260 agents in live operations.
Acquisitions have given mid-tier IT firms specialist capabilities and customer relationships, while AI is changing how they deliver and price services. Their Q2 results show the costs of integrating these businesses and the pressure to find new sources of differentiation as enterprise AI adoption expands.
正文保留英文原文,标题与摘要由机器翻译。
原文发布于 Analytics India Magazine(Forget Organic Growth, AI is Forcing Mid-Cap Indian IT to Reinvent M&A Playbook),标题、摘要与正文由机器翻译,术语按行业惯例统一,未经人工逐句校订。