Why Didn’t India Build Its Own Silicon Valley?印度为何没能建立起自己的硅谷？

Dwaipayan Banerjee’s new book investigates India’s role in the history of 20th-century tech development, and what caused it always to seem a step behind the West.德瓦伊帕扬·班纳吉（Dwaipayan Banerjee）的新书探讨了印度在20世纪科技发展史中的角色，以及为何它总是显得比西方落后一步。

By Angela SainiAugust 5, 2026作者：安吉拉·赛尼（Angela Saini），2026年8月5日

• Science & Technology科学与技术

• History历史

• Politics政治

IN THE 1970S, my father was among thousands of science and engineering graduates who left India for the West. The United States and Britain had flung open their doors to Asian immigrants. Leaving India made sense for the superfluously educated, whom India was ambitious enough to train even while lacking the infrastructure to employ them. The West, however, wasn’t always what they had hoped for. Like my dad, many ended up running small businesses or finding quiet jobs in research or manufacturing, hitting glass ceilings if they tried to move any higher. But as their numbers grew and discrimination softened, Indian-born engineers would eventually take over the biggest technology firms in the world, from Google (Sundar Pichai) to Microsoft (Satya Nadella) and Adobe (Shantanu Narayen). In Fremont, the closest major East Bay city to California’s Silicon Valley, nearly a third of residents are Indian Americans.20世纪70年代，我的父亲是成千上万离开印度前往西方寻求发展的科学和工程专业毕业生之一。美国和英国当时向亚洲移民敞开了大门。对于那些受过高等教育、但印度国内缺乏相应基础设施来提供就业的人来说，离开印度是合理的选择。然而，西方并不总是他们所期望的样子。像我父亲一样，许多人最终经营小生意，或在科研或制造业中从事默默无闻的工作，一旦试图晋升就会遇到“玻璃天花板”。但随着人数的增加和歧视的缓解，印度出生的工程师最终接管了世界上一些最大的科技公司，从谷歌（桑达尔·皮查伊）到微软（萨提亚·纳德拉）再到 Adobe（山塔努·纳拉延）。在加州硅谷附近最大的东湾城市弗里蒙特，近三分之一的居民是印度裔美国人。

Their success leaves us to ask why India never built a Silicon Valley of its own.他们的成功让我们不禁要问，为什么印度从未建立起自己的硅谷。

It is more of a puzzle than it seems. While it’s true that India is best known for backend IT services, it lags so far behind China and the United States in scientific research that it’s scarcely mentioned in the same breath anymore. For decades, however, the outcome of the global technological arms race was far from certain. For a while, India was a contender. After wresting independence from the British in 1947 (when my father was two years old), Prime Minister Jawaharlal Nehru bet big on science and technology, nurturing what he termed India’s “scientific temper,” a quality that in late antiquity had led its thinkers to discover “the zero,” a foundational concept in modern mathematics. India started its own space program as early as 1962. It launched its first satellite in 1975. It established the Indian Institutes of Technology to rival the technical colleges of the West.这比表面看起来更令人费解。虽然印度以后端 IT 服务闻名，但它在科学研究方面落后于中国和美国太远，以至于在提到这些国家时几乎不再提及印度。然而，几十年来，全球科技军备竞赛的结果远未确定。曾几何时，印度也是竞争者。1947年从英国手中夺回独立（当时我父亲两岁）后，总理贾瓦哈拉尔·尼赫鲁在科学技术上投入巨资，培育了他所称的印度的“科学气质”——正是这种特质在古代促使印度思想家发现了现代数学的基础概念“零”。印度早在1962年就启动了自己的太空计划。1975年，它发射了第一颗卫星。它还建立了印度理工学院，以抗衡西方的技术学院。

“In the 1950s, 1960s, and 1970s, policymakers across the Global South adopted these radical economic ideas,” writes MIT anthropologist Dwaipayan Banerjee in his new book Computing in the Age of Decolonization: India’s Lost Technological Revolution, which traces the nation’s early computing successes and failures. For some Asian nations, these bets paid off. South Korea and Taiwan bolstered their manufacturing bases and offered incentives to companies that carried out innovative research. Today, the two countries together are the world’s largest net exporters of computer chips. In India, the ingredients were present, but somehow the bread didn’t rise. India retreated into low-cost outsourcing and offshoring services, becoming more of a servant than a competitor to Silicon Valley. Unlike South Korea and Taiwan, India imports far more computer chips than it makes.麻省理工学院人类学家德瓦伊帕扬·班纳吉在其新书《去殖民化时代的计算：印度失落的科技革命》中写道：“在20世纪50年代、60年代和70年代，全球南方的决策者们采用了这些激进的经济理念。”该书追溯了印度在计算领域的早期成功与失败。对于一些亚洲国家来说，这些赌注得到了回报。韩国和台湾加强了制造业基础，并为进行创新研究的公司提供了激励措施。如今，这两个国家加起来是世界上最大的计算机芯片净出口国。在印度，虽然具备了各种要素，但面包却没能发起来。印度退回到了低成本的外包和离岸服务，与其说是硅谷的竞争对手，不如说是它的仆人。与韩国和台湾不同，印度进口的计算机芯片远多于其生产的。

“This was not the future that Indian policymakers hoped for,” Banerjee observes. The postindependence dream had been to harness high technology to repair the damage of colonial underdevelopment, perhaps even to leapfrog the nations that had once ruled the country. As the historian S. Irfan Habib has noted, “for Nehru, the solution to India’s problems lay in a combination of socialism and science, of technology and heavy industrialisation.” Instead, notes Banerjee, “India provides the raw labor that drives Global North research and manufacturing.”“这并不是印度决策者所期望的未来，”班纳吉指出。独立后的梦想是利用高科技来修复殖民时期欠发达造成的损害，甚至可能赶超曾经统治这个国家的国家。正如历史学家 S·伊尔凡·哈比卜所指出的，“对于尼赫鲁来说，印度问题的解决方案在于社会主义与科学、技术与重工业化的结合。”然而班纳吉指出，“印度提供的是驱动全球北方科研和制造业的原始劳动力。”

Indeed, he adds, American computing dominance is premised on underdevelopment elsewhere. If it had not depressed the technological ascent of countries like India and absorbed their engineering talent, the United States would never have become the computing power it is now.事实上，他补充道，美国的计算霸权是以其他地区的不发达为前提的。如果不是因为它抑制了像印度这样的国家的技术崛起并吸收了它们的工程人才，美国绝不会成为今天这样的计算强国。

This is more than a story of bad decisions and lost opportunities. For Banerjee, it’s a cautionary tale about the enduring effects of empire: the long psychological reach of providing cheap labor and raw materials for two centuries while the colonizers reaped the real profits. In the 18th century, India had been one of the world’s biggest exporters of fine cotton textiles. By the early 19th century, its own cotton was being processed at an industrial scale by English textile mills, making Britain unimaginably rich while India languished.这不仅仅是一个关于错误决策和错失良机的故事。在班纳吉看来，这是一个关于帝国持久影响的警示故事：在两个世纪里提供廉价劳动力和原材料，而殖民者却攫取了真正的利润，这种影响具有深远的心理层面。18世纪，印度曾是世界上最大的精纺棉织品出口国之一。到了19世纪初，它自己的棉花被英国纺织厂进行工业化加工，使英国变得富得流油，而印度却陷入了停滞。

“Computing has become the new cotton,” he argues.“计算已经成为新的棉花，”他认为。

Banerjee’s case study comes by way of a single machine: the Tata Institute of Fundamental Research (TIFR) Automatic Calculator. Developed in Bombay in the 1950s, it was India’s first homegrown digital computer. Having witnessed how foreign powers kept India industrially stagnant by rationing access to technology, scientists and engineers at TIFR resolved to build a device sans foreign parts or input. In this way, they hoped to show that they could be self-reliant, in true Gandhian style.班纳吉的案例研究源于一台机器：塔塔基础研究所（TIFR）自动计算器。它于20世纪50年代在孟买开发，是印度第一台国产数字计算机。在目睹了外国势力如何通过限制技术准入使印度工业停滞不前后，TIFR 的科学家和工程师们决心建造一台无需外国零件或投入的设备。通过这种方式，他们希望以真正的甘地风格展示他们能够自力更生。

The internationally celebrated physicist Homi J. Bhabha, who was the founding director of TIFR and the driving force behind India’s nuclear program, recognized that Indian science would be lost in a backwater without fast computers. Paper and pencil were no longer enough to get Nobel Prize–winning research done. The future lay in processing power. The ENIAC at the University of Pennsylvania (the first general-purpose electronic computer, announced in 1946), mathematician John von Neumann’s IAS machine at Princeton (operational by 1952), and the IBM 701 (the company’s first commercial digital electronic computer, introduced in 1952) were turning the United States into the center of the scientific world.国际知名的物理学家霍米·J·巴巴（Homi J. Bhabha）是 TIFR 的创始主任，也是印度核计划背后的推动力。他认识到，如果没有快速的计算机，印度科学将迷失在死水之中。纸和笔已不足以完成诺贝尔奖级别的研究。未来在于处理能力。宾夕法尼亚大学的 ENIAC（1946年宣布的第一台通用电子计算机）、数学家约翰·冯·诺依曼在普林斯顿的 IAS 机器（1952年投入运行）以及 IBM 701（该公司1952年推出的第一台商用数字电子计算机）正将美国变成科学世界的中心。

Bhabha became a power broker, furiously working to ensure that scientists at TIFR had the equipment they needed. Leveraging his personal contacts and considerable charisma, he earned both the government’s backing and the support of left-leaning Western researchers who felt that formerly colonized countries had been deliberately overshadowed by “a kind of apartheid in global scientific expertise,” Banerjee writes. Although India was officially nonaligned, the Cold War played out behind the scenes.巴巴成为了一名权力掮客，积极奔走以确保 TIFR 的科学家拥有他们所需的设备。班纳吉写道，利用个人关系和相当大的人格魅力，他赢得了政府的支持，也获得了西方左翼研究人员的支持，他们认为曾经被殖民的国家被“全球科学专业知识中的一种种族隔离”故意遮蔽了。尽管印度官方不结盟，但冷战仍在幕后上演。

Still, the challenges for TIFR were immense. By the time work started on the Automatic Calculator, it had been only a decade since the Bengal famine of 1943. India was beset by poverty, malnutrition, and illiteracy. Agricultural scientists were desperately trying to boost crop yields to stop the country from having to import food. This was the era in which a single computer was the size of a large room and cost millions of dollars. Indian researchers would have to build theirs on the tightest of budgets, and without a space of their own; initially, they were operating out of temporary quarters in the Bombay Yacht Club. Their plan was all the more remarkable, adds Banerjee, because no other Asian country at the time, with the possible exception of Japan, had yet established a viable national computing program.尽管如此，TIFR 面临的挑战依然巨大。在自动计算器开始研发时，距离1943年孟加拉饥荒仅过去十年。印度深受贫困、营养不良和文盲的困扰。农业科学家正拼命提高作物产量，以阻止国家进口粮食。那是一个单台计算机就有大房间那么大、耗资数百万美元的时代。印度研究人员必须在最紧张的预算下、且没有自己空间的情况下建造它；起初，他们在孟买游艇俱乐部的临时住所里办公。班纳吉补充说，他们的计划之所以引人注目，是因为当时除了日本外，没有其他亚洲国家建立了可行的国家计算计划。

Their ideological commitment to self-sufficiency was admirable, but it also slowed them down. It was rare for a university or private company anywhere in the world to build a computer using only its own parts and sourcing everything locally. The in-house workshop at TIFR nevertheless “assembled 2700 vacuum tubes, 1700 germanium diodes, and 12,500 resistors,” Banerjee writes.他们对自给自足的意识形态承诺令人钦佩，但也拖慢了他们的进度。在世界任何地方，大学或私营公司仅使用自己的零件并在本地采购所有东西来制造计算机都是罕见的。尽管如此，TIFR 的内部车间还是“组装了2700个真空管、1700个锗二极管和12500个电阻器，”班纳吉写道。

This spirit of making do with what they had, known in Hindi as “jugaad,” has since become a hallmark of Indian technology. The country is infamous for its low-cost work-arounds, crafty reverse engineering, and hacks that somehow get things done with much less than others seem to need. One legendary example is India’s 2023 lunar lander mission, Chandrayaan-3, successfully executed with only 75 million dollars of funding (reporters were quick to point out at the time that the recent Hollywood movies Gravity and Interstellar each had bigger budgets).这种利用现有条件凑合的精神，在印地语中被称为“jugaad”，此后成为了印度技术的标志。这个国家以其低成本的变通方法、巧妙的逆向工程和在资源匮乏的情况下完成任务的黑客手段而闻名。一个传奇的例子是印度2023年的月球着陆任务“月船3号”，仅以7500万美元的资金成功执行（当时记者们很快指出，好莱坞电影《地心引力》和《星际穿越》的预算都比这多）。

We root for the underdog, and the tale of the TIFR Automatic Calculator is the quintessential underdog story. In the 1960s, India did what even many developed nations were not brave enough to do. It saw the future with more clarity than any nation, save the United States, and it chased its dream in defiance of critics at home and abroad who accused it of wasting money. More importantly, it proved that it had the expertise to build a world-class computer on a relative shoestring.我们支持弱者，而 TIFR 自动计算器的故事是典型的弱者故事。20世纪60年代，印度做了许多发达国家都不敢做的事情。它比除美国以外的任何国家都更清晰地看到了未来，并无视国内外批评它浪费金钱的声音，追求着自己的梦想。更重要的是，它证明了自己有能力在相对拮据的情况下建造世界一流的计算机。

The machine, one of the fastest in the world, was officially unveiled in 1962. Indian researchers complained, however, that its speed had come at the price of memory, thereby sowing critical programming problems. It had the same main memory capacity as the IBM 701, which was by then a decade old, and an order of magnitude smaller than IBM’s newest computer, the 709, which didn’t need to make any trade-offs between speed and memory at all.这台机器是世界上最快的机器之一，于1962年正式亮相。然而，印度研究人员抱怨说，它的速度是以牺牲内存为代价的，从而引发了关键的编程问题。它的主内存容量与当时已经有十年历史的 IBM 701 相同，比 IBM 最新的计算机 709 小了一个数量级，而后者在速度和内存之间根本不需要做任何权衡。

By 1965, the year after Nehru died of a heart attack, the TIFR Automatic Calculator had been disassembled for parts. In the archives, Banerjee finds “a sense of disillusionment with the entire experience.”到了1965年，即尼赫鲁因心脏病去世的第二年，TIFR 自动计算器已被拆解。在档案中，班纳吉发现了“一种对整个经历的幻灭感”。

Bhabha, too, died the following year in a plane crash. In hindsight, the entire project was an object lesson in the difficulties of postcolonial development. Skilled as the scientists at TIFR were, their MacGyvering could get them only so far. As “a direct result of British colonialism,” Banerjee explains, India’s postindependence economy was fragile, and high technology required capital. There was just no escaping it, and it is even truer now than it was then. Today’s race to artificial general intelligence, for instance, is turning out to be a matter of who can throw the most money at it. The likely success of companies like OpenAI and Anthropic is measured less by their actual products than by how close each is to a trillion-dollar valuation. By falling back on what could be done quickly and cheaply, by copying what others were doing just to prove they could do it themselves, India relegated itself to precisely where visionary leaders like Nehru did not want it to be—it became a follower.巴巴也在次年死于空难。回想起来，整个项目是后殖民发展困难的一个典型教训。尽管 TIFR 的科学家技术娴熟，但他们的“百战天龙”式做法只能让他们走这么远。正如班纳吉所解释的那样，作为“英国殖民主义的直接后果”，印度独立后的经济非常脆弱，而高科技需要资本。这是无法逃避的，而且现在比当时更是如此。例如，当今对通用人工智能的竞赛正演变成谁能投入更多资金的问题。像 OpenAI 和 Anthropic 这样的公司的成功可能性，与其说是通过其实际产品来衡量，不如说是通过它们离万亿市值有多近来衡量。通过退回到可以快速且廉价完成的事情上，通过复制别人正在做的事情来证明自己也能做到，印度将自己降级到了尼赫鲁等富有远见的领导人所不希望它处于的位置——它成了一个追随者。

IBM emerged in the 1960s as the undisputed winner. And it was capital that got it there. Its computers were so powerful that Bhabha himself was convinced that TIFR would need to import one even as his institute labored feverishly to produce its own. In the end, TIFR bought a computer from another American firm, but it was too late by then to prevent IBM from establishing a near-monopoly in India.IBM 在20世纪60年代成为了无可争议的赢家。正是资本让它做到了这一点。它的计算机功能如此强大，以至于巴巴本人也确信，即使他的研究所正狂热地努力生产自己的计算机，TIFR 也需要进口一台。最终，TIFR 从另一家美国公司购买了一台计算机，但为时已晚，无法阻止 IBM 在印度建立近乎垄断的地位。

“By the mid-1960s,” writes Banerjee, “IBM’s dominance of Indian computing was nearly absolute.” Universities across the country installed its systems. If this weren’t embarrassing enough, many of these IBM machines were reconditioned older models, discarded by customers in the United States or Europe in favor of newer ones. American firms saw the Indian market as a lucrative dumping ground, a chance to make extra profit on obsolete machines. Even by 1972, “106 of India’s 145 active computers” were still produced by IBM.“到了20世纪60年代中期，”班纳吉写道，“IBM 对印度计算领域的统治几乎是绝对的。”全国各地的大学都安装了它的系统。如果这还不够尴尬的话，许多这些 IBM 机器都是翻新的旧型号，是美国或欧洲客户为了更新型号而丢弃的。美国公司将印度市场视为一个有利可图的倾销场，是一个从过时机器上赚取额外利润的机会。即使到了1972年，“印度145台在用计算机中有106台”仍由 IBM 生产。

This is not to say that nothing useful came out of TIFR manufacturing its own device. As Banerjee notes, the act of making a computer was itself instructive, effectively preparing a generation of Indian researchers for the digital age. Even the scientists who knew that their computer was unlikely to rival IBM’s understood that it provided proof of concept, establishing India as a technological power to be reckoned with. The underdog did win, kind of.这并不是说 TIFR 制造自己的设备没有带来任何有用的东西。正如班纳吉所指出的，制造计算机的行为本身就具有指导意义，有效地为一代印度研究人员做好了进入数字时代的准备。即使是那些知道自己的计算机不太可能与 IBM 竞争的科学家也明白，它提供了概念验证，确立了印度作为不可忽视的技术大国的地位。弱者确实赢了，某种程度上。

The stranger part of this story, though, is why India is still the underdog 60 years later. At a recent film screening for a documentary about one of Silicon Valley’s earliest Indian immigrant success stories—the computer networking pioneer Kanwal Rekhi, once described by Fortune as the “Godfather of Silicon Valley’s Indian Mafia”—India was described in much the same terms as before. It remains a work in progress, stuck in a perpetual race with its own ambitions, never quite crossing the finish line. Expectations remain low even as other Asian countries surge ahead.然而，这个故事更奇怪的部分是，为什么60年后的今天，印度仍然是弱者。在最近一场关于硅谷最早的印度移民成功故事之一的纪录片放映会上——主角是计算机网络先驱坎瓦尔·雷基（Kanwal Rekhi），他曾被《财富》杂志描述为“硅谷印度黑手党的教父”——印度被描述得和以前一样。它仍然是一个半成品，陷入了与自身野心的永恒赛跑中，从未真正冲过终点线。尽管其他亚洲国家突飞猛进，但人们对印度的期望依然很低。

For Banerjee, at least some of this failure can be explained by IBM’s unfair early business practices. The machines it sold in India were not just flagrantly overpriced and shamefully outdated (sometimes four years older than the latest models marketed in the United States); they also didn’t allow anyone to tinker with the hardware. As a result, students and researchers were forced to work on software development, since “this was the only aspect of the machine that was accessible for experimentation and learning.” Concomitantly, President Lyndon Johnson’s immigration reforms in 1965 invited skilled workers to move to the United States and bring their families. When they saw the limitations they faced at home in India, young engineers knew they had to leave. These graduates, explains Banerjee, “became integral to US computing.”对于班纳吉来说，这种失败至少部分可以用 IBM 早期的不公平商业行为来解释。它在印度销售的机器不仅价格高得离谱，而且极其过时（有时比在美国销售的最新型号还要旧四年）；它们还不允许任何人对硬件进行修补。结果，学生和研究人员被迫从事软件开发，因为“这是机器唯一可供实验和学习的方面”。与此同时，林登·约翰逊总统1965年的移民改革邀请技术工人移居美国并带上家人。当年轻的工程师们看到他们在印度国内面临的局限性时，他们知道自己必须离开。班纳吉解释说，这些毕业生“成为了美国计算领域不可或缺的一部分”。

In 1978, the exasperated Indian government finally forced IBM out of India. The country went on to find its strength in software. Designers created new operating systems and applications. The government introduced special economic zones for foreign companies that wanted to develop software in India, and it also offered generous incentives to Indian firms exporting software to the US. These policies brought in much-needed dollars, but they also caused problems. Western firms began soaking up Indian programming talent. Everything became geared toward serving lucrative international markets rather than domestic ones. Regretfully, as Banerjee notes, “service provision had become an end in itself.”1978年，忍无可忍的印度政府终于将 IBM 赶出了印度。该国随后在软件领域找到了自己的优势。设计师们创建了新的操作系统和应用程序。政府为希望在印度开发软件的外国公司设立了经济特区，并向出口软件到美国的印度公司提供了慷慨的激励措施。这些政策带来了急需的美元，但也引发了问题。西方公司开始吸纳印度的编程人才。一切都变得旨在服务于利润丰厚的国际市场，而不是国内市场。遗憾的是，正如班纳吉所指出的，“提供服务本身成了目的。”

The rest is history. As they became more focused on services, Indian technology firms turned away from basic research and development and continued importing hardware. They left valuable manufacturing capability, including the production of semiconductors, to others. I visited the sprawling Bengaluru campus of the Indian IT services company Infosys back in 2010. On the surface, it was a landscaped city within a city, with giant food courts and sports grounds, resembling Google’s headquarters in Silicon Valley. But as I learned, much of this was a smoke screen. As Banerjee also observes, “the hardware running its servers is designed abroad; the innovations powering its software are patented elsewhere.”剩下的就是历史了。随着它们越来越专注于服务，印度科技公司背离了基础研发，并继续进口硬件。它们将宝贵的制造能力（包括半导体生产）留给了别人。2010年，我参观了印度 IT 服务公司印孚瑟斯（Infosys）位于班加罗尔的庞大园区。从表面上看，这是一座城市中的景观城市，有巨大的美食广场和运动场，类似于谷歌在硅谷的总部。但正如我所了解到的，这在很大程度上是一种掩饰。正如班纳吉也观察到的那样，“运行其服务器的硬件是在国外设计的；驱动其软件的创新是在其他地方申请专利的。”

But then, the race for technological supremacy never really ends. The fact that India has not yet built its own Silicon Valley doesn’t mean it never will. My father died this year, taking with him memories of a time when there was no grander ambition for an Indian engineer than to work for a big American or European corporation. In his final days this past spring, that era was already fading. The Trump administration’s shortsighted immigration policy has made the United States a far less attractive prospect for highly skilled workers. Paired with the vicious anti-immigration turn across Britain and the rest of Europe, the drip feed of brains from east to west is slowing.但话说回来，追求技术霸权的竞赛永远不会真正结束。印度尚未建立自己的硅谷并不意味着它永远不会。我父亲今年去世了，带走了他那个时代的记忆，那时对于一名印度工程师来说，没有比为美国或欧洲大公司工作更宏大的野心了。在今年春天他生命的最后日子里，那个时代已经消逝。特朗普政府短视的移民政策使美国对高技能工人的吸引力大打折扣。再加上英国和欧洲其他地区恶毒的反移民转向，从东到西的人才输送正在放缓。

Western nations may not realize it, but this will hasten their decline. In the 15th century, astronomers would travel from Kandahar to Samarkand to sit at the beating heart of scientific progress, hoping to see the stars at the observatory of the great mathematician Ulugh Beg. In the 19th and 20th centuries, the best minds flocked to Europe and the United States. Seeing the writing on the wall, my 13-year-old son tells me that he wants to go to college in China or Singapore. The race continues.西方国家可能没有意识到，但这将加速它们的衰落。在15世纪，天文学家会从坎大哈前往撒马尔罕，坐在科学进步的中心，希望在伟大的数学家乌鲁伯格的天文台观测星空。在19世纪和20世纪，最优秀的人才涌向欧洲和美国。看到了大势所趋，我13岁的儿子告诉我，他想去中国或新加坡上大学。竞赛仍在继续。