标签： Asia

Fueled by ultra-affordable data plans, India has rapidly emerged as one of the world’s largest and fastest-growing internet markets, boasting more than one billion broadband subscribers as of late 2025. Behind this digital success story, however, lies a critical, underreported vulnerability in the physical infrastructure that powers the nation’s global connectivity: an overconcentration of critical intercontinental submarine cables in one small stretch of Mumbai coastline, paired with aging infrastructure, a lack of domestic repair capacity, and bureaucratic bottlenecks that threaten to derail India’s ambitious digital and AI-driven growth plans.

The core risk is concentrated in a mere six-kilometer stretch along Mumbai’s Arabian Sea coast near Versova, a densely populated northwestern neighborhood of the city. Of India’s total 18 intercontinental submarine cables— the invisible optical fiber highways that carry 99% of global cross-border data traffic as light pulses across 1.4 million kilometers of ocean floor worldwide— at least 13 come ashore within this tiny geographic zone. Combined, these cables carry up to 95% of India’s total international bandwidth destined for Europe, Africa, and West Asia.

Anwesha Sen, a researcher at the Takshashila Institution, an Indian public policy think tank, warns that this extreme geographic concentration creates an existential critical vulnerability. If multiple cables were cut at the same time, whether by accidental damage or intentional disruption, the vast majority of India’s westbound connectivity would be severely impacted. While modern cable architecture does allow for traffic to be rerouted through other landing points across India’s southern coast, such as Chennai and Kochi, experts note that even minor performance degradation—increased data latency—can carry devastating financial costs for key sectors including banking, national stock exchanges, cloud service providers, and government communications networks. AI infrastructure, which depends on constant high-volume, high-speed data transfers, is uniquely sensitive to even small disruptions to connectivity.

This vulnerability points to a larger paradox at the heart of India’s digital transformation: while the country now accounts for roughly 20% of global data consumption and ranks among the world’s largest internet markets, it controls only a tiny fraction of the global undersea cable infrastructure that powers this digital activity. Data from the Broadband India Forum (BIF) shows that India is connected to just 3% of the world’s total submarine cable network, and hosts only 21 of the roughly 1,900 active and planned cable landing stations (the coastal facilities that connect undersea cables to onshore terrestrial networks) globally—equal to just 1% of the world’s total capacity. “We need to grow the number of cable landing stations several fold,” says Aruna Sundararajan, BIF chairperson.

Compounding the overconcentration risk is the age of India’s existing cable network. A recent analysis from the Takshashila Institution found that 11 of India’s 18 international submarine cables are more than 20 years old, rapidly approaching the standard 25-year nominal operational lifespan for these systems. While routine faults are far more common than deliberate sabotage—with an estimated 150 to 200 cable disruptions annually worldwide, most caused by commercial fishing activity, dragging ship anchors, natural seabed hazards, or equipment failure—India has a second major gap in its preparedness: it has no domestic submarine cable repair vessels. All repair work in Indian territorial waters requires bringing in foreign vessels from ports in Dubai or Singapore, which must then obtain multiple operating licenses before work can begin, adding weeks of delay to an already time-sensitive process.

Amajit Gupta, group CEO and managing director of Lightstorm, a leading Indian digital infrastructure firm, explains that this dependence on foreign repair capacity is the most impactful structural weakness facing India’s connectivity network. “The bigger structural issue for India isn’t rerouting capability—it’s repair capacity. That combination – dependence on foreign repair capacity and regulatory lead times – represents the real bottleneck,” he says. Gupta points to the 2024 Red Sea shipping and cable disruptions as a case study of how unplanned outages can create cascading delays, when global traffic had to be rerouted thousands of miles east through Singapore to reach North American networks.

Beyond repair gaps, India also faces a growing mismatch between exploding demand for international bandwidth driven by AI and data center expansion, and the slow pace of new subsea cable development. Government data shows that India’s installed data center capacity has grown more than fourfold since 2020, jumping from 375MW to 1,575MW today, as companies rush to build computing capacity to power domestic AI model development and global cloud services. All of these facilities rely on fast, reliable subsea connectivity to global networks, but planning and deploying a new submarine cable system takes between five and six years total, meaning new capacity cannot keep up with the rapid growth of data center demand.

Sundararajan identifies fragmented governance and bureaucratic red tape as the single biggest barrier to expanding India’s subsea infrastructure. No fewer than 16 separate government agencies across telecom, defense, home affairs, shipping, port authorities, environmental bodies, and state governments are involved in approving new cable projects and repair operations. Just constructing a single new cable landing station requires roughly 50 separate clearances across more than a dozen ministries. “A core issue is fragmented governance, not a lack of technical capability,” she says. Sundararajan recommends that India’s Department of Telecommunications be designated as the lead central agency, with a single-window approval system to cut wait times for repair vessel permits and new project approvals. She also calls for the government to pre-approve multiple cable landing sites across India’s east and west coasts as well as island territories, so developers do not need to start the approval process from scratch for every new project—an approach that has helped Singapore become a leading regional cable hub through predictable regulation and streamlined processes.

Pooja Bhatt, an associate professor at OP Jindal Global University, notes that the mismatch between India’s digital ambitions and its inadequate core infrastructure can no longer be ignored. “Given its young demographics and ‘Digital India’ aspirations, it makes sense for the nation to build and maintain its own internet infrastructure for security and governance purposes,” she says. The Indian government has already taken one small step to address the issue, officially designating submarine cable systems as critical national telecommunications infrastructure, a classification that industry leaders hope will unlock faster approvals, support the development of a domestic repair fleet, and enable the creation of legally protected zones around cable routes to reduce the risk of accidental damage from fishing and anchoring.

Emerging technologies may also help reduce risk over time: Distributed Acoustic Sensing, a new monitoring tool, can turn existing optical fibers into distributed sensors that detect vibrations from ships, anchors, or underwater vehicles, allowing operators to identify potential threats before a cable is severed. But industry leaders agree that the most critical fixes are low-tech: expanding the number of geographically dispersed cable landing stations, streamlining regulatory approval processes, protecting existing cable routes, and building a domestic fleet of repair vessels.

Gupta notes that connectivity naturally clustered around Mumbai for decades, thanks to existing demand and infrastructure, but that this historic concentration is exactly the risk that the industry must move away from. “But that concentration is exactly the risk the industry needs to design away from,” he says. The shift to greater diversity is already underway: four new submarine cable systems are currently being commissioned across India, with three more in advanced planning stages. One of the new projects, I-2SEA, led by a Lightstorm-led consortium connecting India to Malaysia and Singapore, will feature two separate landing points in India: one at Machilipatnam, which creates a shorter route to the major technology hub of Hyderabad, and a new landing site in South Chennai, adding much-needed geographic diversity to existing networks.

For a country that consumes data at one of the fastest rates in the world and is racing to build capacity for an AI-driven economy, the physical infrastructure that powers its global connectivity remains surprisingly fragile. While India is rapidly expanding its data center capacity and positioning itself as a global digital leader, the open question remains whether it can upgrade its subsea cable network, landing station capacity, and repair ecosystem fast enough to keep pace with its own ambition.

For decades, theme parks have relied on roller coasters, themed character meet-and-greets, and elaborate live shows to draw millions of visitors annually. Now, a new project in South Korea is pushing the boundaries of leisure entertainment by asking a forward-thinking question: could robots become the core of the next generation of theme park experiences?

At the center of this experimental shift is Galaxy Robot Park, a sprawling 16,500-square-meter purpose-built venue that showcases how robotic technology can be integrated into leisure attractions. Unlike traditional theme parks that blend human staff and mechanical rides, this innovative park centers cutting-edge robotics as its main draw, with one of the most popular showcases featuring humanoid robot performers choreographed to dance to chart-topping K-pop hits. The synchronized, precise movements of these robots offer a one-of-a-kind spectator experience that cannot be replicated by human performance teams, drawing curiosity from both tech enthusiasts and casual theme park visitors.

This venture represents a growing global trend of integrating advanced automation and artificial intelligence into the entertainment and leisure sectors. Theme park operators around the world have been testing robotic elements for years, from automated ride systems to interactive robot guides, but Galaxy Robot Park is one of the first full-scale venues dedicated entirely to robot-centric attractions. Industry analysts point out that the experiment could reshape how theme parks design experiences, offering lower long-term operational costs for venue owners while opening up new possibilities for performances and interactions that push the limits of what human performers can do. At the same time, the project raises questions about the future of human employment in the leisure sector, and how audiences will respond to an entirely robot-driven leisure experience.

Recent reports have revealed that multiple major Chinese technology firms—including Alibaba, ByteDance, and Tencent—have gained remote access to computing power powered by Nvidia’s cutting-edge GB300 chips via third-party data centers located in Thailand, Malaysia, and Japan. The incident comes just days after a White House accusation against Chinese AI startup Moonshot AI for using advanced Nvidia processors, laying bare a critical flaw in Washington’s long-running strategy of restricting China’s access to top-tier semiconductor technology.

This case underscores a simple but consequential truth in an era of deeply interconnected global technology: blocking direct shipments of advanced chips to China does not equate to cutting off Chinese access to their computing power. As digital infrastructure and tech networks grow increasingly cross-border, controlling the physical movement of hardware has failed to control how that hardware’s capabilities are used.

For years, the United States and China have implemented parallel strategies to reduce mutual economic and technological dependency. Washington has imposed steep tariffs on Chinese goods, coordinated with allied nations to enforce sweeping export controls targeting China’s strategically critical sectors, and restricted Beijing’s access to cutting-edge American technologies. In response, Beijing has prioritized building greater financial and technological self-reliance, reduced its dependence on U.S.-dominated global payment systems, and leveraged its dominant position in key global supply chains—most notably rare earth materials—to protect its core economic and national security interests.

Data from the 2026 Stanford AI Index illustrates just how far China has advanced in artificial intelligence development despite U.S. restrictions. The report finds that the performance gap between the world’s leading U.S. and Chinese AI models has narrowed significantly in recent years. While the U.S. still outpaces China in producing the highest-tier AI models and high-impact research patents, China now leads the world in the volume of AI academic publications, total citation counts, overall AI patent output, and annual installations of industrial robots.

These trends confirm two key developments: China is rapidly building its capacity to develop indigenous AI technology, and U.S. efforts to slow China’s technological progress have hit inherent limits. Beyond that, the global semiconductor and AI ecosystem itself undermines Washington’s restrictive approach. The U.S. hosts more than 5,400 data centers—over 10 times the number hosted by any other single nation—and nearly all of the world’s most advanced chips are manufactured by Taiwan Semiconductor Manufacturing Company (TSMC), leaving the U.S. AI supply chain dependent on a single major foundry. In short, America’s own ambition to maintain global leadership in AI relies heavily on interconnected international production networks.

The current state of U.S. semiconductor policy toward China traces back to a decision by the Trump administration to resume sales of Nvidia’s previous-generation H200 AI chips to “approved Chinese customers,” while maintaining a full embargo on the latest Blackwell architecture chips. The policy was designed as a calculated gamble: by granting access to an older chip generation while blocking cutting-edge options, Washington sought to shape China’s AI development trajectory and strengthen its bargaining hand in broader trade talks with Beijing.

But this strategy has not delivered the intended results. By deploying H200 chips within mainland China to train domestic AI models, China has continued to advance its homegrown AI sector while drawing benefits from existing American technology, all without building long-term dependence on Washington. This has effectively neutralized the leverage the Trump administration aimed to gain.

Worse for U.S. containment efforts, the perception in Beijing that Washington is actively trying to stifle China’s technological rise has only accelerated its push to build a fully self-reliant domestic semiconductor industry. Even Nvidia CEO Jensen Huang has acknowledged that U.S. withdrawal from the Chinese semiconductor market has accelerated growth of China’s domestic chip sector.

The limits of U.S. economic and trade pressure extend far beyond semiconductor technology. In August, the White House released a report titled “The Great Transplant Scam,” which accused Chinese exporters of routing goods through more than 40 third-party countries to evade U.S. tariffs. The report, while highlighting what Washington frames as evasion, inadvertently reveals the inherent limits of Washington’s goal to reduce U.S. economic dependence on China. While the U.S. can erect trade barriers to cut direct imports of Chinese goods, it cannot stop China-linked global supply chains from adapting to the new rules. The report itself even acknowledges that much of the production shift across third countries stems from “legitimate changes” in manufacturing, investment, and sourcing, meaning that not all China-connected goods entering the U.S. via intermediates represent illegal activity—many companies are simply relocating production to mitigate geopolitical risk.

Independent analysis backs up this reality. Research from the Peterson Institute for International Economics, a Washington-based think tank, shows that even after years of U.S. tariffs, Chinese goods and services remain deeply integrated into U.S. imports from third-party countries. While direct bilateral trade between the two nations has fallen, Chinese inputs have not been removed from global supply chains. A separate survey from Nikkei Asia also found that Chinese companies have expanded their global market share even amid U.S. tariff barriers.

At the same time, China remains closely tied to U.S. consumer and industrial demand. As direct bilateral trade has fallen under U.S. trade barriers, China’s exports of electronics, computing hardware, and circuit board assemblies to neighboring Asian economies have actually increased. Much of this production feeds into the ongoing U.S. AI boom, meaning Chinese manufacturers are indirectly benefiting from Washington’s own investments in AI development. This does not amount to illegal transshipment of goods into the U.S.; it simply demonstrates that the core economic ties between the two nations remain intact, even as trade routes have shifted.

With multiple studies confirming that Chinese inputs still reach the U.S. via intermediaries like Vietnam and Mexico, the Trump administration appears to have accepted that a full economic “divorce” between the world’s two largest economies is not feasible in the near term. Its policy has thus shifted away from the earlier goal of full decoupling, toward a new strategy focused on extracting economic concessions from Beijing—including securing Chinese commitments to address U.S. concerns over critical mineral supply chains, purchase large quantities of Boeing commercial aircraft, and reopen markets to American agricultural goods.

This shift does not mean the U.S. has abandoned pressure tactics entirely. Washington is now pursuing what it frames as selective engagement: maintaining open trade with China in non-sensitive sectors while continuing to restrict access to technologies deemed critical to national security. The emerging strategy is designed to increase non-sensitive trade with Beijing to preserve broader economic leverage, while continuing to block technology access in sensitive areas. For its part, Beijing has doubled down on building domestic technological capacity, reducing its exposure to foreign suppliers, and building its own counter-leverage against Washington.

The Trump administration’s clear policy shift from full decoupling to selective engagement has reshaped the U.S.-China economic and technological standoff. Washington continues to deploy tariffs and technology restrictions to secure benefits for American manufacturers and farmers, while China uses its supply chain dominance and reduced reliance on American semiconductors to strengthen its negotiating position. In this carefully calibrated standoff, both sides are leaning into their respective strengths: Washington applies targeted economic pressure, while Beijing finds workarounds to blunt that pressure. All eyes now turn to the upcoming summit between Chinese President Xi Jinping and U.S. President Donald Trump in Washington later this month, where the two sides will determine whether this current dynamic paves the way for a major new bilateral trade agreement.