#329 Look AheadTen Predictions for 2026, A Claim on China's Geopolitical Leverage Due to Rare Earths, and an Anticipated Pax Silica Plus India
India Policy Watch: Predictions 2026Insights on current policy issues in India—RSJThe usual routine for the first post of the year is my list of predictions. This is a bit of an indulgence. But like all indulgences, I love doing them. There isn’t any science to them, nor are these so specific with binary outcomes that you can truly gauge how accurate you were at the end of the year. Over the years, however, I have tried to be more specific with predictions with a tighter range of figures wherever possible and avoided extrapolating a trend to arrive at a prediction or two. There’s a pleasure in looking at trends and then taking an imaginative, bucking-the-trend leap about where a thing or two will end up. But that was when things were normal and boring. And, when peering into the future brought a minor frisson of excitement into my boring life. We live in the age of Trump now, where daily events and announcements are such sharp departures from the past that you can only shake your head in disbelief and update your priors. This may seem like making predictions daunting. On the contrary, I think it makes them more fun. So, I had more than my usual share of fun in coming up with the ten predictions for 2026. There’s only one key difference about this set of predictions from those made in previous years. In the past, I tried to keep my predictions individually distinct but mutually compatible. Simply put, the consequences of one prediction, were it to come true, wouldn’t negate the other predictions. I have done away with this kind of compatibility. This is because Trump is shaking up geopolitics in a manner that all bets are off on what this year could bring in terms of one-off events. Any of these one-off events, if they come true, could obliterate every other prediction about the economy or the market. With that out of the way, here are my somewhat specific predictions for 2026.
Those are my predictions. I will wait and watch. Global Policy Watch: The Long History of Non-fuel Mineral GeopoliticsGlobal issues relevant to India—Pranay KotasthaneThe weaponisation of minerals seems unprecedented, but it really isn’t. While people are well aware of instances when fuels (oil, gas, etc.) were weaponised, past cases of non-fuel minerals being weaponised are hard to recall. That creates a perception that China’s hold over rare earths is unprecedented and undefeatable. Yet, there are multiple prominent cases in which the geopolitical leverage countries had because of their dominance over critical minerals dissipated quickly. It’s not as if these countries stopped producing the minerals; they just couldn’t weaponise them for too long. Take the case of Cobalt. It was the in-fashion dual-use material of the 1970s. One tonne of it went into every F-16’s jet engine. The US imported almost all of it, most of which came from then-Zaire (Democratic Republic of Congo). The USSR was signing mineral agreements with African countries, providing technical assistance in exchange for a share in total output. Modelled on the OPEC, there were cartels for Bauxite, Copper, and the like. A US Congressman termed the situation the ‘materials myopia’ of the West. And then, things came to a head when some rebels occupied mining regions in Zaire, leading to a supply cutoff and a sudden price increase. And yet, there was no repeat of the 1973 Oil Crisis. What really happened? A combination of several moves eventually reduced the strategic edge that Cobalt-producing countries enjoyed. First, the high prices made Cobalt mining in other, more stable countries lucrative—mining in Canada, Brazil, and Australia picked up. Even as these new mining projects were taking shape, prior research on substitutes for Cobalt was commercialised much more quickly. The more abundant Molybdenum began displacing Cobalt in certain alloys as jet engine makers warmed to new composites and carbon fibres. Next, companies began to economise. Just like we read reports today that every electric car, submarine, or jet fighter uses X kgs of neodymium, without which the world would come to a grinding halt, there were similar concerns about Cobalt in the 1970s. But the demand proved more elastic than people had assumed. Non-priority uses of cobalt were reduced by using other materials in alloys, and when substitution wasn’t possible, efficiency improvements helped contain overall demand. The fourth leg was government intervention. Military stockpiles were built to meet emergency defence applications. For the broader commercial market, governments also took up price-risk mechanisms and offtake guarantees. The net result wasn’t that Congo was displaced as the world’s major Cobalt producer; it still accounts for over 70 per cent of global supply today. Rather, Cobalt itself was displaced as the hitherto indispensable material necessary for dual-use applications. Once substitutes were developed, the hackles were lowered, Cobalt stopped being a geopolitical talking point, and the world became comfortable with buying Cobalt from Congo once again. Several such instances exist in the near past. China cut off tungsten exports to the US during the Korean War in the early 1950s. China had been supplying half of American demand at a dirt-cheap $16 per pound. Then came the operationalisation of the Defence Production Act of 1950, which allowed the government to offer offtake guarantees, write off investments, and provide low-interest loans. The US government guaranteed a price of $63 per pound to domestic miners, and within three years, domestic firms were producing thrice the government’s reuired demand. Similarly, Mercury was another dual-use material of the 1950s, crucial for walkie-talkies and ammunition. A looming shortage and rising prices caused widespread furore. But through the same process of market-led substitution and alternative sources backed by government guarantees, the strategic significance of Mercury went down. I see a version of these Cobalt, Tungsten, and Mercury stories playing out in the case of China’s weaponisation of rare earths, too. As I have written in multiple editions, substitutes will be operationalised, alternative sources will emerge, and efficiency improvements will reduce per-unit requirements. Additionally, the Law of One Price (LOOP) will play its role. China’s extensive export controls have meant that the international prices of Chinese rare earth oxides (say in Europe) are at least three times the domestic prices in China. Under such stark conditions, many Chinese companies will be willing to take the risk of evading controls in exchange for the promise of extravagant profits. Shell companies might be set up, and third countries might be used to convert ‘black’ rare earths into ‘white’. LOOP will kick in, and the price differential will wither away. Given these trends and in the spirit of making falsifiable predictions, I claim that by December 2028, China will lose its geopolitical leverage due to its dominance in rare earths. While it will remain the world’s largest rare earths producer, the strategic importance of these materials will decline. China will stop putting additional export controls on these technologies and begin diluting existing ones by that time. Hope this newsletter will be alive three years from now to test this prediction. But if not, I am sure some sharp readers are keeping track. Matsyanyaaya: Pax Silica! Ab Tera Kya Hoga India? Part 2Big fish eating small fish = Foreign Policy in action—Pranay Kotasthane and Tannmay Kumarr Baid(An edited version of this article first appeared in India’s World on 16th January 2026) In January 2026, the United States envoy, Sergio Gor, invited India to join Pax Silica. Pax Silica is a US-led effort with an aim to align partners across semiconductors, AI infrastructure, and critical minerals. The stated objective of the initiative is to reduce dependencies in the technology supply chain and shape how the AI economy is built. This invite came barely a month after New Delhi was conspicuously absent from the initiative’s inaugural summit in Washington. Some suggested that the absence reflected India’s declining relevance in Washington’s technology calculus, while others inferred that it proves India was a lightweight in the semiconductor domain. Despite the alarmism, one of us (Pranay) had blogged that India’s initial absence from Pax Silica matters far less than the noise suggested, and that India would eventually become a part of the grouping. That position seems to have held its ground. The core reason is simple: India already sits at the centre of the global semiconductor and AI ecosystem, with or without formal membership. The absence reflected ongoing trade frictions and tariff disputes, rather than a downgrade of India’s role in the US technology strategy. India is deeply embeddedFramed as a new organising principle for the global silicon and AI supply chain, Pax Silica aims to coordinate trusted partners through policy alignment, joint projects, and investment coordination to diversify supply chains, with an outward focus on reducing dependence on China. That ambition was understandable. But even then, the way the initiative was launched suggested it was more a political signal than an operational framework. Read in that light, India’s absence weakened Pax Silica on paper, but it did not materially alter or affect India’s position in the global technology ecosystem. India is, and remains structurally embedded in the semiconductor value chain. Pax Silica does not, and cannot, change that in the foreseeable future. This is important because semiconductor design is explicitly identified as a core pillar of Pax Silica. In this segment, India is already indispensable. Nearly every top-25 fabless semiconductor company by revenue operates large design centres in India. Crucially, these are not peripheral support offices. Most of them work on core intellectual property and system-level design. These are the processes that are at the heart of chip development cycles. This massive base that global companies have in India is driven by incentives, talent availability, and ecosystem depth. The Indian government claims that nearly 20 per cent of the global chip design talent is in India. Any attempt to bypass India’s design base would impose higher costs and slower innovation on the very firms Pax Silica seeks to support. In that sense, India is already inside the tent where it matters most upstream. This design strength also explains why India’s absence from a diplomatic grouping did not imply exclusion from future technology gains. The semiconductor supply chain is not neatly modular. Design, software, and systems integration are tightly coupled with downstream manufacturing and deployment. Pax Silica may seek to coordinate trusted production nodes, but it cannot exclude India, which is where a significant chunk of the thinking, testing, and iteration already takes place, and hope to succeed. Manufacturing cannot be wished awayIndia’s manufacturing role, while still emerging, is no longer hypothetical. Eight OSAT (Outsourced Semiconductor Assembly and Test) plants and one commercial CMOS fab are under active construction. While these projects do not compete with Taiwan or South Korea on advanced logic nodes, they do not need to. Assembly, testing, and packaging are becoming increasingly cutting-edge to squeeze out better performance without necessarily making transistors smaller. This is also where diversification is most feasible and easiest to achieve. In the coming decade, India is likely to emerge as a meaningful OSAT node. If the objective is to reduce concentration rather than just signal alignment, then having capacity in a large, independent market is a huge asset, not an inconvenience. The same logic applies to artificial intelligence. Pax Silica places heavy emphasis on AI infrastructure and deployment, yet India sits close to the centre of both. Indian software firms are well-positioned in what Kai-Fu Lee has described as “enterprise AI”, where models are integrated into business processes rather than built from scratch. This adoption, rather than just creation, is where most economic value is likely to accrue over the next decade. India is also one of the largest potential markets for AI deployment across finance, logistics, healthcare, manufacturing, and public services. An effort to shape the global AI economy while leaving out both major integrators and a market of this scale will struggle to deliver returns for participating firms. AI supply chains do not end at data centres. They end where systems are deployed at scale. Looking AheadThe first cut of participants in Pax Silica reflected diplomatic convenience rather than supply chain completeness. We have seen this before. Coalitions built around technology tend to expand in phases. Early announcements prioritise ease of alignment over comprehensiveness. There is also a risk of over-securitisation. Treating AI and semiconductors purely through a national security lens can narrow choice and raise costs. Recent easing of certain chip export controls and deals with China in rare earths suggest that even Washington recognises the limits of blunt restriction. China will continue to remain a player in the AI, semiconductor and critical minerals market in the future, and an attempt to cut it off entirely is neither feasible nor economically wise. The aim must remain to reduce single-source dependence. Timing and next stepsA more plausible explanation for India’s initial absence lies in trade frictions rather than strategic distance. Tariff disputes continue to weigh on India–US economic ties. It is reasonable to hypothesise that India may have chosen not to join a high-visibility economic grouping while being singled out on trade, or that the US may not have chosen to include India at that moment. This made the absence a bargaining position, not an outright rejection of cooperation, as subsequent events now suggest. There is precedent for this approach. India was not a founding member of the Mineral Security Partnership either. That triggered similar concern at the time. India joined later, once the contours of the grouping were clearer and its interests better reflected. Pax Silica has followed a similar trajectory. Semiconductor and AI firms with deep exposure to India have little incentive to see India permanently excluded from any framework that shapes their operating environment. Even in December, there was little reason for alarm. Semiconductor firms have continued to expand design and engineering work in India. AI providers have continued to invest in Indian data centres and deployment to monetise their models. OSAT capacity under construction did not hinge on the outcome of one summit. India’s continued priority should be domestic execution: improving manufacturing competitiveness, reducing regulatory friction, scaling packaging and testing capacity, and focusing on AI deployment. Delivering on these projects will be the make-or-break factor for India. HomeWorkReading and listening recommendations on public policy matters
|
