# AI and digital platforms: Value chain, from upstream to the customer Model builders and apps selling AI as assistants, APIs and copilots, not hardware Source page: https://www.materialinsights.org/industries/ai_platforms/value-chain/ #### Value chain, from upstream to the customer **Criticality to the chain:** 5 · Chokepoint 4 · Critical 3 · Important 2 · Supporting 1 · Commodity service Stage 1 **AI chipsAI accelerator and chip designers**5 · Chokepoint Design the GPUs and custom accelerators that train and run models, and sell them with networking and software stacks to clouds and labs.^[5](https://siliconanalysts.com/analysis/nvidia-ai-accelerator-market-share-2024-2026) **Why chokepoint:** Nvidia took an estimated 80-90% of AI accelerator revenue in 2025, and it, AMD and most custom chips are fabricated and CoWoS-packaged by one foundry, TSMC in Taiwan, so supply is rationed with no near-term substitute. Technologies GPU and TPU architectures High-bandwidth memory (HBM) Advanced packaging (CoWoS) High-speed interconnects (NVLink, InfiniBand, Ethernet) CUDA and other software stacks Depends on Commodities [Si Silicon](https://www.materialinsights.org/commodities/silicon/)[Cu Copper](https://www.materialinsights.org/commodities/copper/)[Au Gold](https://www.materialinsights.org/commodities/gold/)[He Helium](https://www.materialinsights.org/commodities/helium/)[Ta Tantalum](https://www.materialinsights.org/commodities/tantalum/) Chemicals [ESG Electronic specialty gases](https://www.materialinsights.org/chemicals/electronic_gases/)[PR Photoresists](https://www.materialinsights.org/chemicals/photoresists/)[UPW Ultra-pure wet chemicals](https://www.materialinsights.org/chemicals/wet_chemicals/)[EPX Epoxy resins](https://www.materialinsights.org/chemicals/epoxy/) Energy [Dist Distribution grids, mini-grids and interconnectors](https://www.materialinsights.org/energy/e_distribution/) Industries [Semiconductors](https://www.materialinsights.org/industries/semiconductors/)**Examples:** Nvidia, AMD, Google (TPU), Broadcom Stage 2 **Compute infrastructureHyperscale clouds and GPU cloud operators**4 · Critical Build and run data centres full of accelerators and rent training and inference capacity to labs, enterprises and governments.^[1](https://www.gartner.com/en/newsroom/press-releases/2026-1-15-gartner-says-worldwide-ai-spending-will-total-2-point-5-trillion-dollars-in-2026)^,^[3](https://www.iea.org/reports/energy-and-ai/energy-demand-from-ai)^,^[6](https://www.iea.org/news/data-centre-electricity-use-surged-in-2025-even-with-tightening-bottlenecks-driving-a-scramble-for-solutions)^,^[7](https://www.crnasia.com/news-network/2026/cloud-market-share-q1-2026-aws-microsoft-google-battling-in-ai-era) **Why critical:** Three hyperscalers held about 63% of world cloud infrastructure spending in early 2026, and only a few firms can finance and power gigawatt-scale campuses, which wait years for grid connections and gas turbines. Technologies GPU clusters with liquid cooling High-voltage grid connections Software-defined networking Model serving platforms On-site storage and backup power Depends on Commodities [Cu Copper](https://www.materialinsights.org/commodities/copper/)[Al Aluminium](https://www.materialinsights.org/commodities/aluminum/)[St Steel](https://www.materialinsights.org/commodities/steel/)[H₂O Water](https://www.materialinsights.org/commodities/water/)[Gas Natural gas](https://www.materialinsights.org/energy/natural_gas/) Chemicals [HFC Fluorochemicals and refrigerants](https://www.materialinsights.org/chemicals/fluoro_intermediates/)[WT Water treatment chemicals](https://www.materialinsights.org/chemicals/water_treatment/) Energy [DC Data centres](https://www.materialinsights.org/energy/e_data_centres/)[HVAC AC transmission](https://www.materialinsights.org/energy/e_hvac/)[CCGT Gas turbines and combined cycle](https://www.materialinsights.org/energy/e_gas_turbines/)[LFP Lithium iron phosphate batteries](https://www.materialinsights.org/energy/e_li_lfp/) Industries [Data centres and cloud](https://www.materialinsights.org/industries/data_centres/)[Computers and electronics](https://www.materialinsights.org/industries/computers_electronics/)[Electrical equipment](https://www.materialinsights.org/industries/electrical_equipment/)**Examples:** Microsoft Azure, Amazon Web Services, Google Cloud, CoreWeave, Yotta Data Services · ≈ 55% of AI spending in 2025 (calculated from)^[1](https://www.gartner.com/en/newsroom/press-releases/2026-1-15-gartner-says-worldwide-ai-spending-will-total-2-point-5-trillion-dollars-in-2026) Stage 3 **Data, talent and capitalData licensors and labelling firms**3 · Important License text, images, video and code for training, and supply human labelling and expert feedback to tune models. **Why important:** Many sources and labelling firms exist, but high-quality licensed and expert data is scarce and copyright disputes can cut off supply. Technologies Data licensing and rights management Annotation platforms Reinforcement learning from human feedback Synthetic data generation Depends on Energy [DC Data centres](https://www.materialinsights.org/energy/e_data_centres/) Industries [Media and entertainment](https://www.materialinsights.org/industries/media_entertainment/)[Software and IT services](https://www.materialinsights.org/industries/software_it/)**Examples:** Scale AI, Reddit, Shutterstock, iMerit **Universities and public research labs**3 · Important Train AI researchers and engineers, publish methods and benchmarks, and build open models and datasets, including for Indian languages. **Why important:** Talent pipelines take years, but many universities worldwide train AI researchers. Technologies Open-source models and datasets Benchmarks and evaluations Academic compute clusters Indian-language corpora Depends on Energy [DC Data centres](https://www.materialinsights.org/energy/e_data_centres/) Industries [Education and research](https://www.materialinsights.org/industries/education_research/)**Examples:** Stanford University, Mila, IIT Madras (AI4Bharat), IISc Bengaluru **Venture capital and strategic investors**3 · Important Fund model labs, chip start-ups and AI applications with equity, compute credits and debt for data centres.^[2](https://hai.stanford.edu/ai-index/2026-ai-index-report/economy) **Why important:** Corporate AI investment hit US$ 581.7 billion in 2025, but it depends on a few very large backers and on capital market sentiment. Technologies Venture and growth equity Compute-for-equity deals Project finance for data centres Sovereign AI funds Depends on Industries [Finance: banking, insurance, markets](https://www.materialinsights.org/industries/finance/)**Examples:** Microsoft, SoftBank Vision Fund, Sequoia Capital, Peak XV Partners Stage 4 **Foundation modelsFoundation model developers**4 · Critical Train large language, image and multimodal models and sell access through apps and APIs to consumers, developers and enterprises.^[2](https://hai.stanford.edu/ai-index/2026-ai-index-report/economy) **Why critical:** A handful of labs build frontier models, with annual compute bills of US$ 16.3 billion for OpenAI and US$ 6.8 billion for Anthropic in 2025, though open models give partial substitutes. Technologies Transformer and mixture-of-experts architectures Large-scale distributed training Reinforcement learning and alignment Inference optimisation Safety evaluations Depends on Energy [DC Data centres](https://www.materialinsights.org/energy/e_data_centres/) Industries [Data centres and cloud](https://www.materialinsights.org/industries/data_centres/)[Education and research](https://www.materialinsights.org/industries/education_research/)[Media and entertainment](https://www.materialinsights.org/industries/media_entertainment/)[Finance: banking, insurance, markets](https://www.materialinsights.org/industries/finance/)**Examples:** OpenAI, Anthropic, Google DeepMind, Meta, Sarvam AI Stage 5 **Tooling and integrationData, MLOps and model hosting platforms**2 · Supporting Provide the data warehouses, model hubs, fine-tuning and monitoring tools companies use to build their own AI systems. **Why supporting:** Many vendors and open-source tools compete, and switching takes weeks for most workloads. Technologies Data lakehouses Vector databases Model hubs and fine-tuning Evaluation and monitoring Agent frameworks Depends on Energy [DC Data centres](https://www.materialinsights.org/energy/e_data_centres/) Industries [Software and IT services](https://www.materialinsights.org/industries/software_it/)[Data centres and cloud](https://www.materialinsights.org/industries/data_centres/)**Examples:** Databricks, Snowflake, Hugging Face, Fractal Analytics **IT services firms and AI integrators**2 · Supporting Integrate models into enterprise systems, build agents and copilots for clients, and run AI operations under services contracts.^[1](https://www.gartner.com/en/newsroom/press-releases/2026-1-15-gartner-says-worldwide-ai-spending-will-total-2-point-5-trillion-dollars-in-2026) **Why supporting:** Thousands of integrators compete for AI services work, which Gartner puts at about US$ 439 billion in 2025, so clients can switch within weeks. Technologies Retrieval-augmented generation Workflow and agent integration Legacy system modernisation AI governance and testing Depends on Energy [Bld Energy in buildings](https://www.materialinsights.org/energy/e_buildings/) Industries [Software and IT services](https://www.materialinsights.org/industries/software_it/)**Examples:** Accenture, Tata Consultancy Services, Infosys, IBM Consulting Stage 6 **Applications and platformsAI applications and consumer digital platforms**2 · Supporting Offer assistants, search, copilots and AI features in social, productivity and vertical apps, earning subscriptions, usage fees or ads. **Why supporting:** Thousands of apps compete and users can switch easily, though platforms with large user bases have strong distribution advantages. Technologies Chat and voice assistants Coding and office copilots AI search Recommendation feeds Agents that act in software Depends on Energy [DC Data centres](https://www.materialinsights.org/energy/e_data_centres/) Industries [Software and IT services](https://www.materialinsights.org/industries/software_it/)[Data centres and cloud](https://www.materialinsights.org/industries/data_centres/)[Finance: banking, insurance, markets](https://www.materialinsights.org/industries/finance/)**Examples:** OpenAI (ChatGPT), Microsoft (Copilot), Perplexity, Meta (Meta AI), Zoho (Zia) ## Other sections - [Overview](https://www.materialinsights.org/industries/ai_platforms/index.md) - [Where ₹100 goes](https://www.materialinsights.org/industries/ai_platforms/cost-mix/index.md) - [Price links with suppliers](https://www.materialinsights.org/industries/ai_platforms/price-links/index.md) - [Example story](https://www.materialinsights.org/industries/ai_platforms/story/index.md) - [Linked industries](https://www.materialinsights.org/industries/ai_platforms/linked-industries/index.md) - [Materials drawn on](https://www.materialinsights.org/industries/ai_platforms/materials/index.md) - [Chokepoints and risks](https://www.materialinsights.org/industries/ai_platforms/risks/index.md) - [Sources](https://www.materialinsights.org/industries/ai_platforms/sources/index.md)