Impact Of The West Asian War On AI Development and Diffusion
Authors
Executive Summary
This report finds that the war in West Asia exposes fragilities in the global AI supply chain that impact the national AI strategies of most countries, including India. The impact can be analysed along three dimensions.
First, the AI stack extends beyond data, models, compute and energy prioritised in the national strategies of most states. Disruptions in helium and bromine supplies, and energy flows because of the effective closure of the Strait of Hormuz, have short-term impacts on the world’s production of advanced semiconductor chips and high bandwidth memory that is concentrated in Korea and Taiwan, and used for AI training and inference.
Second, AI infrastructure has entered the theatre of conflict in two ways simultaneously. The Islamic Revolutionary Guard Corps designated AWS and Oracle facilities in the UAE and Bahrain as “legitimate targets”, marking the first confirmed kinetic strikes on commercial cloud infrastructure in history. At the same time, the battlefield use of AI-enabled decision-support systems such as Lavender, Gospel, and Maven by Israeli and American forces generated targets at combat speed. This raises questions about whether meaningful human control is possible. AI is now both an object and an instrument of war.
Third, a protracted war risks elevated energy prices, rising input and insurance costs, and stagflation pressures that could stress the AI investment cycle at its historical peak. Geopolitical uncertainty is likely to accelerate market concentration in a few firms with deep pockets, furthering the divide and raising costs of access and adoption for emerging economies.
For India, a few implications follow. Frontier AI’s emergence as an instrument of US foreign policy argues for higher defence AI R&D expenditure and diplomatic advocacy for “meaningful human control” in lethal autonomous weapons systems. Risks of capital flight from emerging markets and rising costs necessitate frugal innovation and supporting open-tech alternatives. On the positive side, India may benefit from positioning itself as a data centre hub as an alternative to the Gulf. But reforms that unlock private-sector participation in the generation, transmission, and distribution of energy and provide an effective path for clean energy transition would be key to activating the required levers.
Introduction
India’s primary concerns vis-a-vis the war in West Asia pertain to its 9 million strong diaspora, their safety and associated remittance flows as well as energy dependencies from the region. This paper moves away from these important, pressing issues and studies a quieter set of second and third order impacts that the war has revealed and accelerated in the global artificial intelligence development and diffusion.
The paper is divided into four sections - supply chain fragilities; the dual role of AI infrastructure in the war - as a subject of kinetic strikes and as an instrument of warfighting through decision-support systems; the war’s effects on the AI investment landscape; and finally, implications for India.
Supply Chain Fragility
The primary building blocks of the AI stack include data, models, compute, and energy. In addition to these, infrastructure, investments, talent and the regulatory ecosystem have been the focus for building a resilient AI supply chain. Most national AI policies, including India’s AI Mission, target one of these different building blocks. However, the Iran war has revealed that the global supply chains of the technology are more complex and often go down to the input materials involved.
Input Materials
Consider helium, which is extremely useful in magnetic resonance imaging machines, semiconductor manufacturing and aerospace applications. Thanks to its excellent thermal conductivity and inertness, it serves as a coolant for maintaining a stable environment during the semiconductor fabrication process. Its supply chains are fairly concentrated, dominated by the US, Russia and Qatar. Being one of the lightest gases, it leaks over time, making stockpiling expensive and technically demanding.
QatarEnergy declared force majeure in March 2026 following strikes on key facilities in Ras Laffan Industrial City. This has effectively taken 30% of global capacity of helium, produced as a byproduct of LNG refining, offline 1. However, the element is one of several inputs, and high-tech industries command enough purchasing power to secure supply at elevated prices and from other sources, even as other buyers get squeezed. Bromine is another input in the semiconductor etching process whose supply is being crunched by the war and leading to a price surge – two-thirds of the world’s bromine comes from Israel and Jordan 2. While the Iran war is unlikely to trigger a helium crisis, short-term disruptions are already underway while prices respond and alternative sources are being explored. This exposes the fragility of technology supply chains to geopolitical risks.
Energy
A more significant disruption has been caused by the energy crisis triggered by the war. South Korean firms like Samsung Electronics and SK Hynix account for nearly 80 per cent of the global high-bandwidth memory (HBM) and 70 per cent of the dynamic random access memory (DRAM) market 3. The country is heavily dependent on fossil fuels for its energy needs, with 70 per cent imported from West Asia through the Strait of Hormuz 4. A similar story plays out in Taiwan, where 90 per cent of the world’s most advanced semiconductors are manufactured 5. While supply has not been affected yet, the sustained high input costs are putting pressure on profit margins and could accelerate efforts to raise chip prices.
Timing these risks is the real tax for end users. They cannot hedge against an unknown reopening date, so they pay for dual-sourcing, strategic stockpiles, forward contracts at wartime surge prices. Insurance premia for passage through the Strait of Hormuz, which were a small line item prior to the war, have skyrocketed with states stepping up to insure losses and provide naval escorts 6. As resilience takes precedence over efficiency, firms with weaker balance sheets will fold sooner, concentrating the market in hyperscalers that can absorb the increased costs.
AI Infrastructure Becomes Part of the Theatre of Conflict
AI Use in the War
In the past few years, AI has transitioned from a supportive role in military operations to a primary driver of mass target identification and kinetic execution. The speed of this transition has outpaced the development of international humanitarian law – a classic pacing problem, where rapid, exponential advancements in tech outpace the slow, incremental development of laws, regulations, and social structures designed to manage them.
Israel has invested heavily in integrating AI into their decision support systems (DSS), with systems such as Lavender, which shortlists individuals for targeting based on suspected affiliation with armed groups; The Gospel (Habsora), which generates target lists of buildings and structures; and Where’s Daddy?, which tracks individuals’ locations ahead of a potential strike by tracking their phones 7. The US is also extensively using DSS such as Palantir’s Maven Smart System to analyse surveillance data, create targeting lists and enable target-prioritisation 8.
These systems expedite the decision-making process by processing vast amounts of data – Israel’s Lavender DSS was reportedly used to generate over 37,000 targets in the opening weeks of the Gaza war, and the US was able to hit over 1000 targets within the first 24 hours of the Iran war 9. However, it risks the collapse of human control as officers defer to machine suggestions because they cannot process the information at machine speeds. The human in the loop becomes ceremonial.
AI Infrastructure Seen as a Legitimate Target
While firms such as Palantir and Corsight AI are the vendors for the AI-based DSS, there are several other tech companies that provide the enabling infrastructure. These include big-tech companies such as Amazon, Anthropic, Google, Microsoft, and OpenAI 10 .
In a confrontation with militarily outmatched adversaries, Iran has resorted to tactics that exert political and economic pressure. In an unprecedented move, the Islamic Revolutionary Guard Corps (IRGC) designated Amazon Web Services and Oracle’s facilities in UAE and Bahrain as “legitimate targets” because they hosted military applications and proceeded with the first confirmed military strikes on commercial cloud data centres in history 11.
The strikes themselves caused minor disruptions to operations, as cloud data has redundancies built in. However, they set a precedent for a new category of infrastructure in need of military protection from kinetic strikes, in addition to cyber threats. For the US, there are several implications for policy, such as establishing a credible deterrence, intel sharing with industry or serving as a backup insurer for such risks 12.
West Asia as an AI Hub?
The Gulf Cooperation Council states, particularly Qatar, the United Arab Emirates, and Saudi Arabia, are making a strategic push to diversify their economies beyond oil towards emerging technologies such as AI. To position themselves as a regional hub, some of the states have proposed the idea of digital embassies, where foreign governments and companies can govern their data under their own terms regardless of the location of the datacentres 13. The Gulf states are an attractive destination for data centres due to their inexpensive energy, favourable geography relative to other emerging markets and access to capital from sovereign wealth funds.
The dynamics of the AI race between the US and China are a key determinant of who gets access to the building blocks of frontier AI. Saudi Arabia, Qatar and the Emirates, having signed access-for-investment deals with the US 14. They get access to state-of-the-art AI chips in exchange for investments with US partners. Further, the Pax Silica, which is a US initiative that aims to secure the AI supply chain among a strategic bloc of allies and partners, has Qatar and the United Arab Emirates as signatories 15.
For the Gulf states, the intent to partner with the likes of Google, Microsoft, OpenAI and others could also be seen as a way to lock in a security guarantee from the US 16. However, the emerging precedent of attacks targeting AI infrastructure means that data centres, energy facilities, and sub-sea cables connecting the region to the global internet are now vulnerable targets. Given the high costs, including long-term power contracts, land agreements, and fibre connectivity, invested in already operational facilities, it’s unlikely that hyperscalers will relocate built capacity. But the marginal investment decision for new capacity rollouts would factor in this added geopolitical risk, alongside other location-evaluation metrics such as tax incentives and power-grid stability.
Investment Under Stress
Fragile Pre-War Baseline
The war arrived at the peak of the largest and most concentrated capital expenditure cycle in the history of the technology sector. The five US frontier AI labs and hyperscalers — Amazon, Microsoft, Alphabet, Meta, and Oracle — had committed between USD 660 billion and USD 770 billion of capex in 2026, roughly three-quarters of it directed at AI infrastructure 17. The scale of the investments by US companies exceeds that of monumental past capital expenditures, such as the Manhattan Project, the moon landing, or the broadband buildout. Aggregate hyperscaler capex had risen from USD 256 billion in 2024 to USD 443 billion in 2025, growing at an annualised rate of roughly 72 per cent since the second quarter of 2023 18.
Notably, this buildout is being funded through debt, and not internal cash generation and is yet to be tested through a credit cycle. Hyperscalers raised USD 121 billion in new debt in 2025 19. In February 2026, Alphabet Inc. issued a rare 100-year bond (century bond) primarily meant to finance AI infrastructure that was oversubscribed 10 times. Critics have raised concerns about the aggressive accounting practices employed by these firms, such as extending the “useful life” of AI assets, despite the rapid obsolescence of high-end GPUs 20. The important question is whether the eventual adoption, diffusion and revenue from AI services will scale fast enough to cover these recurring depreciation charges.
AI Investments
In the scenario of a prolonged war, including the Gulf countries participating, the world may be headed for stagflation pressures like those of the 1970s: high energy prices contributing to global inflation while economic activity stagnates. This is likely to have an impact on the profitability of frontier AI labs and hyperscalers. Returns on AI investments now need to price in greater geopolitical and market concentration risks. This is likely to stretch the timelines between investment and return, even leading to valuation compressions for AI companies and concentrating market power in favour of companies with deeper pockets. A potential capital redistribution away from emerging markets can threaten access, diffusion and developmental goals in these societies.
Globally, whether AI adoption timelines will be compressed for productivity gains or stretched due to a declining global GDP growth rate remains to be seen.
Implications for India
AI Sovereignty and Defence AI Policy
The India AI mission has built baseline domestic capabilities through support for indigenous large language models, with applications in agriculture, healthcare and education being the focus. However, to a large extent, Indians access frontier AI through two channels: proprietary models mostly from US labs (Anthropic, Google, OpenAI) and open weight models from Chinese and US labs (Z.ai, Alibaba, DeepSeek, Google, NVIDIA, OpenAI) 21.
The Iran war and related events in the past few months have highlighted the sovereignty risks. The US and Israel have used frontier AI capabilities on the battlefield to enable decision-making at combat pace. Meanwhile, the Pentagon-Anthropic standoff shows that the US-state can get AI companies to comply with their demands – when Anthropic drew red lines on surveillance and autonomous weapons, it was designated a “supply chain risk” and OpenAI, Google and xAI accepted the “all lawful use” standard. This highlights the fact that frontier AI can be a tool of US foreign policy. Most of the frontier AI companies are American and currently restrict access of military applications to the US and the country’s closest allies.
The importance of increasing defence R&D expenditure is evident – it is currently a small share of the 2% GDP defence budget. India should also push diplomatically for “meaningful human control” in lethal autonomous weapons systems and AI-enabled targeting. On the civilian side, India should pursue plurilateral efforts with partners such as the EU, Japan and Korea to achieve shared sovereignty 22. This can be done by supporting open-tech alternatives where concentration is highest. Some examples include support for open standard instruction set architectures such as RISC-V, open source alternatives to NVIDIA’s proprietary programming platform CUDA, open-source model development, and harmonised evaluation.
Infrastructure and Supply Chain Resilience
The war highlights risks to AI infrastructure. The resilience measures that are likely to follow add capex and opex costs across the AI stack that will eventually be borne by end-users.
Iran reportedly threatening to cut Red Sea cables during the war 23 was the second stress test on India’s digital backbone in recent months. The Red Sea cable cuts in September 2025 caused internet shutdowns and latency across large parts of the country. Over 60% of India’s internet traffic transits through West Asia and lands at Mumbai 24. The other landing points at Chennai and Kochi have a smaller presence, making it a critical vulnerability. Addressing it requires terrestrial backbone redundancy, landing diversification and indigenous repair capabilities – India currently depends on foreign cable-repair fleets – to be built up.
The Iranian strikes on AWS facilities in the UAE and Bahrain, and the IRGC’s “legitimate target” framing, establish a precedent that data centres could be military targets. For hyperscalers, this makes geopolitical risk a site-selection criterion alongside power, water, land and state policies.
The short-term disruptions caused by the war expanded the perception of the AI supply chain to include materials few considered strategic – Helium, Bromine and energy from West Asia are crucial inputs for semiconductor manufacturing in East Asia. Building on Beijing’s weaponisation of its processing dominance in several minerals, including gallium, germanium, graphite, and other rare earths in its trade war with the US, the war in West Asia is likely to further the diversification imperative.
Economic and Strategic Factors
The economic fallout of the war will impact AI development and diffusion in several ways, depending on how prolonged the conflict lasts. A prolonged war with sustained high energy prices could contribute to global inflation and stagnation of economic activity. Some of the different dimensions of this impact are described below.
Rising costs will favour better-resourced companies that can sustain longer timelines between investment and return. Building resilience against supply chain risks passes on the increased costs of energy, semiconductors and insurance to the companies building AI infrastructure. With demand already outpacing supply, investments into fabs by companies such as TSMC won’t yield chips until a few years later – those with deeper pockets will be able to lock-in multi-year allocations while new entrants face restricted access at spot prices.
Global equity capital that was chasing returns in emerging markets or second-tier AI labs might move to safer assets in uncertain scenarios. The increase in oil prices also leads to weakening of emerging market currencies such as the Indian rupee due to higher costs of oil imports and capital outflows seeking the dollar as a safe haven. Since most AI capex is US dollar-denominated, it inflates the bill of materials significantly.
The GCC states, with their access to capital, energy and favourable location had positioned themselves as a key pillar for the diffusion of the American AI stack. With the attacks on data centres in the Gulf, capital allocation now needs to factor in geopolitical risks due to security guarantees, insurance, and location diversification. This opens up an opportunity for India to be an alternative data centre hub to the Gulf. India’s national data centre policy already offers a 20-year tax holiday, along with states offering other incentives related to access to land, energy and water. The Indian electricity system has major problems that limit private-sector participation and the transition to greener energy sources. Centralised planning controls resource allocation, which might not respond to market demands; transmission and distribution losses due to theft also add pressure to the finances; and there is no effective path for clean energy transition 25. Reforms to India’s electricity sector aimed at increasing private-sector participation are under consideration 26 but a much-needed comprehensive overhaul faces political economy pressures. In addition, safety would be a key factor in relocation decisions for AI researchers and their families. Depending on how long the war stretches, talent flows to the Gulf could be impacted.
Constraints on access to capital might lead to increased pressures for frugal innovation. DeepSeek and other Chinese open weight models are gaining popularity across Africa, aided by strategic partnerships with firms such as Huawei 27. The widespread adoption of competing ecosystems might be a better outcome for most countries than one dominated by a few big tech companies from a single country.
Lastly, one of the spillover effects of this war may be that the impact of generative AI on the workforce and the broader economy, unfolding simultaneously with the effects of the war, is difficult to mitigate. On the heels of the aftereffects of a global pandemic and supply chain readjustments forced by the tariff wars, the war in Iran might serve as an additional confounding variable, making it harder to gauge AI’s true impact.
Footnotes
Béchard, Deni Ellis. “The AI Boom Is Dangerously Dependent on Helium.” Scientific American, March 18, 2026. Link↩︎
Kharpal, Arjun, and Dylan Butts. “How the Iran War and Rising Energy Prices Are Threatening Semiconductor Demand.” CNBC, March 10, 2026. Link↩︎
Trueman, Charlotte. “Samsung and SK Hynix to Scale Up Memory Production Capacity in 2026 to Meet AI Demand.” Data Center Dynamics. January 2026. Link↩︎
Sharma, Abhishek. “The Hormuz Crisis and South Korea’s Energy–Alliance Nexus.” Observer Research Foundation, April 13, 2026. Link↩︎
Li, Kevin. “The Iran War Tests Taiwan’s Energy Resilience.” Atlantic Council, March 19, 2026. Link↩︎
Humairah, Laila. “Hormuz Becomes World’s Most Expensive Waterway after 300% Surge in Risk Premiums.” Euronews, March 16, 2026. Link↩︎
Serhan, Yasmeen. “How Israel Uses AI in Gaza–And What It Might Mean for the Future of Warfare.” Time, December 18, 2024, Link↩︎
Mohammed, Noor. “The Proliferation of AI-Enabled Military Technology in the Middle East.” International Institute for Strategic Studies, April 2026. Link↩︎
ibid.↩︎
ibid.↩︎
Al Jazeera Staff. “Iran Declares US-Israeli Economic, Banking Interests in Region Are Targets.” Al Jazeera, March 11, 2026. Link↩︎
Harding, Emily. “Data Is Now the Front Line of Warfare.” Center for Strategic and International Studies, March 19, 2026. Link↩︎
G42. “G42 Introduces Digital Embassies and Greenshield to Make AI Sovereignty Portable.” Press release, January 20, 2026. Link↩︎
Yacoubian, Mona, and Samuel Zabin. “If Compute Is the New Oil, War in the Gulf Significantly Raises the Stakes.” Center for Strategic and International Studies, February 27, 2026. Link↩︎
U.S. Department of State. “Pax Silica.” Under Secretary for Economic Affairs. Accessed April 20, 2026. Link↩︎
Cook, Steven A. “For the Gulf States, Investment in AI Is Partly About U.S. Protection.” Foreign Policy, February 23, 2026. Link↩︎
Srivastava, Prakhar, and Anirban Sen. “Big Tech to Invest about $650 Billion in AI in 2026, Bridgewater Says.” Reuters, February 23, 2026. Link↩︎
Juniewicz, Isabel. “Hyperscaler Capex Has Quadrupled since GPT-4’s Release.” Epoch AI, February 26, 2026. Link↩︎
Bair, Theodore Jr., and Bryan Steele. “Record-Breaking AI-Related Debt Issuance in 2025.” Global Macro Views (blog). Mellon Investments Corporation, December 2025. Link↩︎
Tiger Newspress. “Michael Burry Sounds Alarm on $176B Depreciation Gap among Tech Giants.” Tiger Brokers, November 12, 2025. Link↩︎
Artificial Analysis. “LLM Leaderboard — Comparison of over 100 AI Models from OpenAI, Google, DeepSeek & Others.” Accessed April 20, 2026. Link↩︎
Sharma, Lokendra, Abhishek Kadiyala, Arindam Goswami, Ashwin Prasad, and Nitin Pai. “India Must Architect Sectoral Plurilateral Blocs to Overcome Geopolitical Coercion.” Takshashila Institution, January 2026. Link↩︎
TOI Tech Desk. “Iran Reportedly Threatening to Cut Red Sea Cables: What It Means for Internet Users across the World, Including India.” Times of India, March 2026. Link↩︎
Communications Today. “Undersea Cable Risks in West Asia — What It Means for India’s Connectivity.” March 2026. Link↩︎
Shah, Ajay, and Susan Thomas. “Electricity Cost Recovery and the Political Imagination: A Comparison between Private and Public Distribution in India’s Biggest Cities.” The Leap Blog, November 2, 2025. Link↩︎
Ministry of Power. “Draft National Electricity Policy (NEP), 2026 Released for Public Consultation with Stakeholders.” Press release. Press Information Bureau, January 21, 2026. Link↩︎
Microsoft Corporate Blogs. “Global AI Adoption in 2025 — A Widening Digital Divide.” Microsoft, January 8, 2026. Link↩︎