Dragoman Digest

The difficulty of leaving deters foreign investors from setting up shop in India

Labour protections designed to preserve manufacturing jobs also prevent their creation

India requires an average of 4.3 years to close a factory. In Singapore the average is one year, compared to 15 months in Germany, and one to two years in the UK. General Motors’ (GM) four-year exit from its Pune plant (announced in 2020) is a case in point. Despite offering severance pay equivalent to 110 days’ salary per year of service, GM faced lawsuits from 1,100 employees. Meanwhile, Maharashtra’s labour minister refused to grant GM permission to close, declaring the company should absorb losses and stage a comeback. Courts then mandated that GM continue to pay workers half their salaries until the disputes were resolved. Several interested buyers explicitly refused to purchase the facility because they feared inheriting these labour obligations, forcing GM to secure signed waivers from every worker before ultimately completing the sale to Hyundai in 2023.

These barriers facing companies seeking to exit or close operations – replicated across India – create market distortions. Only three percent of Indian factories close annually. This is among the lowest rates globally. By contrast, Vietnam closes 17 percent of factories annually, China 10 percent, and the US nine percent. This artificially suppressed closure rate means approximately 20 percent of manufacturing companies operate as zombie firms – consuming capital while producing nothing. Officials hold significant discretionary power over factory closures. Consequently, approval outcomes vary based on political calculations at the state level rather than consistent legal standards. When Ford closed two plants in 2021, Tamil Nadu approved the Chennai closure because officials prioritised preserving future investment relationships. Conversely, in Gujarat, the exit only succeeded because Tata agreed to retain all workers.

US tariffs on India, set at 50 percent for the time being, should provide an impetus for India to pursue competitiveness enhancing reforms. Yet these reforms remain politically difficult. The same labour protections that deter foreign manufacturers have broad-based popularity in India. This leaves India’s manufacturing sector trapped between protecting politically powerful incumbent workers and attracting the investment flows needed to build competitive manufacturing at scale.

Anthropic leads OpenAI in enterprise adoption yet trails drastically in valuation

Investors favour mass appeal over sustainable revenue growth

US-based AI startup Anthropic continues to extend its lead in corporate market share over rival OpenAI. As of mid-2025, Anthropic held 32 percent of enterprise model usage versus OpenAI’s 25 percent. Yet, OpenAI’s October US$500 billion valuation is significantly higher than Anthropic’s US$183 billion valuation from September. While Anthropic has concentrated on corporate customers, OpenAI has pursued mass-market consumer adoption. Anthropic’s over 300,000 business and enterprise customers generate 80 percent of its revenue and produce substantially higher revenue per user than OpenAI’s 800 million weekly ChatGPT users. Only around 30 percent of OpenAI’s revenue comes from business users. Anthropic’s considerably higher revenue growth rate – nearly tripling annually versus OpenAI’s 50 percent expansion – puts it on track to overtake OpenAI’s revenue by 2027.

OpenAI’s projected US$13 billion in annual revenue appears insignificant compared to computing commitments that could reach US$1 trillion subject to milestone-based deployments. These include a US$300 billion five-year deal with Oracle for data centre capacity, plus conditional chip agreements with Nvidia, AMD, and Broadcom totalling another US$200-300 billion. Monthly subscription fees of US$20 to US$200 cannot offset the massive costs of AI development and infrastructure. This leaves the company without a clear path to profitability beyond advertising. On this front, OpenAI faces direct competition from Google, whose upcoming Gemini 3.0 model (Google’s ChatGPT equivalent) will combine superior AI capabilities with decades of advertising infrastructure. Anthropic faces much less uncertainty around its ability to monetise investments. Its focus on measurable enterprise productivity gains has driven significant adoption in the corporate coding market (42 percent to OpenAI’s 21 percent), legal documentation, and billing applications. Microsoft’s addition of Anthropic’s Claude model to its Copilot suite despite exclusive ties to OpenAI was a clear vote of confidence in Anthropic’s offering.

Both companies rely on circular financial structures. Google and Amazon are investing billions in Anthropic while serving as its cloud providers. However, OpenAI’s arrangements dwarf Anthropic’s in scale and complexity. Nvidia is investing up to US$100 billion in OpenAI while selling it chips worth the same amount. AMD offered 10 percent of its stock, worth US$36 billion, as a sweetener for orders. OpenAI’s trillion-dollar commitments, backed by vendors whose stock prices benefit from announcing the deals, obscure whether demand reflects genuine business viability.

US tech giants move to reduce supply chain dependence on China

Untangling deeply embedded supply chains at speed will prove challenging

US technology companies are pushing to build alternative production options and reduce reliance on China amid ongoing uncertainty caused by tariffs and export controls. Firms such as Microsoft, Google, and Amazon Web Services (AWS) are among those actively working to establish non-China production bases. These US tech behemoths are following in the footsteps of hardware leader Dell, which in 2023 pledged to phase out China-made chips by the following year. HP has also set an internal goal to shift up to 70 percent of its computer production out of China within three years. Microsoft has specifically instructed suppliers to prepare “out of China” production options for its Surface notebook computers and data centre server products. Its most ambitious target is for over 80 percent of components for the Surface notebook to be sourced from outside China by 2026. AWS is moving critical AI data centre server production away from Chinese suppliers, while Google has expanded server production in Thailand as part of its broader “China+1” strategy.

While relocating final assembly operations is relatively straightforward, shifting manufacturing at the component level is a far more fundamental change. Microsoft’s 2026 goal will likely prove especially challenging. Most US tech companies remain deeply reliant on increasingly advanced and cost-competitive Chinese partners for key parts such as printed circuit boards (PCB), cables, and connectors. AWS, for instance, continues to source from Chinese PCB supplier SYE. In the case of Apple, Chinese suppliers have, somewhat ironically, been crucial to helping it establish supply chains outside China. With no individual country possessing China’s efficiency or scale, more diversified supply chains will likely present added costs.

ASEAN Power Grid gets a second lease on life

Ambitious plans for a regional power grid will still face steep costs and political hurdles

Momentum is building for the Association of Southeast Asian Nations’ (ASEAN) Power Grid (APG). First envisioned more than three decades ago, the APG aims to connect ASEAN’s disparate national and sub-national power grids. It will be crucial to enabling Southeast Asia’s energy transition (given the intermittency of renewables) and meeting surging demand from energy-intensive industries. Some studies indicate the APG will lower the cost of decarbonisation by US$800 billion (11 percent) and reduce electricity costs by nearly four percent. For years, the APG was effectively stillborn. However, in recent years, projects like the Monsoon Wind Power Project in southern Laos, a 600MW onshore wind farm that exports power to Vietnam, have acted as a proof of concept. The Asian Development Bank (ADB) has recently pledged US$10 billion to help mobilise financing for the APG. Singapore – which imports 95 percent of its energy and wants to replace gas consumption with imported renewables – has also emerged as a key champion. ASEAN plans to expand cross-border capacity to 17.5 GW by 2040, up from 7.7 GW in 2024.

A shift in momentum notwithstanding, the build-out of the APG is still expected to face significant obstacles. Only half of the 18 planned cross-border interconnection projects have been completed, most of them land-based. Many projects, including more technologically complicated subsea cable links, are still in the early planning stages. The ADB’s financing will hopefully help mobilise investment but is relatively minor compared to the estimated overall cost of US$764 billion required for building cross-border transmission lines and associated infrastructure. Fragmented regulations and divergent technical standards present further obstacles. ASEAN lacks consistent technical standards for operating multi-country grids and trading power across borders. Small scale power exports from Laos to Singapore via Thailand and Malaysia, the most multilateral of any of the cross-border ASEAN initiatives, faced protracted wrangling over grid transit fees. Whilst none of these obstacles are inherently insurmountable, a much greater degree of political cooperation will be required to realise the APG.

Chinese EV and battery investment in Europe faces headwinds

European countries are increasingly cautious of Chinese investment

Chinese-owned EV and battery production facilities across Eastern and Central Europe experienced rapid expansion in the early 2020s, driven by the efforts of Chinese companies to localise production and reduce tariff risks. The crown jewels of China’s EV and battery industries, BYD and CATL, have both made large investments. CATL has invested well over €10 billion in battery facilities across Spain, Hungary, and Germany. For its part, BYD is building auto plants in both Hungary and Türkiye. Many less globally prominent Chinese companies and their suppliers have followed suit. However, Chinese investment momentum in Europe has since slowed sharply, amid the EU’s imposition of countervailing tariffs on Chinese EVs late last year and weaker-than-expected demand. Newly announced Chinese EV projects fell by 79 percent in 2024 from an average of €15 billion in 2022 and 2023 to just €3.1 billion. Major projects have been cancelled, such as Svolt Energy’s €4.2 billion battery plants in Germany and Nuode’s €500 million battery plant in Belgium.

Policymakers in Europe are increasingly concerned that Chinese investments have limited benefit for Europe’s economy and its vast auto supplier sector. This has prompted Brussels to consider making Chinese auto and battery investments contingent on forced technology transfers and a set share of local content. Fearing that BYD’s €4 billion Hungary assembly plant would merely serve as a conduit for tariff evasion, EU officials have launched an investigation into the facility under the EU’s Foreign Subsidies Regulation mechanism. European EV demand has also proven weaker than anticipated, accounting for only about 16 percent of new car registrations. Meanwhile, automakers like Volkswagen and BMW have stepped up efforts to reclaim the European market by introducing competitively priced, low-cost EVs. With these constraints, Chinese investment is increasingly flowing to Southeast Asia.