India context
India’s economy is a developing mixed system where the state remains pivotal in strategic sectors. It is the world’s fourth-largest economy by nominal GDP and third by PPP, yet per capita ranks are lower (around 136th by nominal GDP and 119th by PPP). After decades of Licence Raj controls modelled on Soviet-style protectionism, the 1991 liberalisation opened markets and catalysed private enterprise. Today, with roughly 1,900 public sector undertakings and state control over railways, highways, core banking and insurance, telecom, space, ports and essential utilities, India combines an assertive public sector with rising private champions—an important context for any AI and semiconductor scale-up.
The announcement
OpenAI has partnered with Broadcom to build a planned 10 gigawatts (GW) of custom AI chips, with deployment beginning in the second half of 2026 and completing by the end of 2029. For Indian readers tracking Digital India, Make in India and the semiconductor PLI scheme, this multi-year roadmap signals a fresh wave of global AI compute build-outs that will shape supply chains, energy demand, data centre design, and talent requirements across markets, including India.
| Item | Detail |
|---|---|
| Total AI chip capacity | 10 GW (10,000 MW) planned |
| Partners | OpenAI and Broadcom |
| Deployment start | H2 2026 |
| Target completion | Q4 2029 |
| Relevance to India | Semiconductor design, data centres, green energy, AI services |
Scale: 10 GW
Ten GW is industrial-scale computing. In power terms, 10,000 MW running continuously would consume up to 87.6 TWh per year (10,000 MW × 8,760 hours). Even at a 90% utilisation factor, that’s roughly 78.8 TWh annually. For comparison, this is larger than the entire annual electricity use of many countries, underscoring how AI training and inference now demand power, cooling, and networking at utility scale. Expect knock-on effects in high-capacity networking, advanced packaging, and liquid cooling—areas where Indian engineering services already operate globally.
Why Broadcom
Broadcom is a long-standing designer of custom silicon and high-speed networking. For a company like OpenAI, aligning with a partner that can integrate compute, interconnect, and system design at scale reduces supply risk and potentially optimises total cost of ownership. While specifics of node, packaging or foundry are undisclosed, the takeaway is clear: hyperscale AI is moving toward vertically optimised silicon stacks, a direction that opens opportunities for India’s VLSI design centres, EDA talent, and verification services.
India’s semiconductor push
India’s semiconductor strategy marries public capacity with private execution. The India Semiconductor Mission and associated incentives—commonly cited around Rs 76,000 crore—aim to catalyse fabs, OSAT/ATMP, and compound semiconductors. With 1,900+ PSUs and strong state involvement in power, transport and telecom, India can coordinate land, utilities, and infrastructure while private players bring agility. As global leaders commit multi-gigawatt compute roadmaps, Indian states vying for fabs and advanced packaging can position themselves across design, assembly, test, and specialised components.
Data centres in India
India’s data centre market is scaling with multi‑gigawatt pipelines across Mumbai, Chennai, Hyderabad, and NCR, aided by state policies, improved fibre backbones, and subsea cable landings. AI-ready facilities require high-density racks, liquid cooling, and low-latency networks—capabilities Indian operators are now building. For Indian enterprises, the OpenAI–Broadcom plan signals that AI workloads will keep growing faster than general compute, making AI-optimised data centres and sovereign cloud strategies a priority.
Power and sustainability
AI capacity at 10 GW puts a spotlight on clean power. India has crossed 180 GW of installed renewable capacity when including large hydro, and is expanding solar and wind rapidly. Aligning AI clusters with firm, green power—through hybrid PPAs, storage, and grid upgrades—can reduce both cost and carbon intensity. Expect greater interest in round-the-clock renewables, transmission buildouts, and demand response as AI data centres seek stable, sustainable power within India’s evolving grid.
Public sector’s role
Given India’s substantial state role in railways, highways, power, ports, and telecommunications, policy coordination can accelerate AI infrastructure: expedited right‑of‑way for fibre; dedicated substations for high‑density campuses; efficient water allocation for cooling; and export-friendly rules for advanced equipment. The same public-sector heft that built national highways and digital rails can underpin AI-era infrastructure—while ensuring security, privacy, and resilience for critical workloads.
Jobs and services
For India’s tech workforce, AI silicon and systems create demand in RTL design, physical verification, firmware, HPC networking, and data centre engineering. India’s IT services exports exceed $200 billion, and AI transformation across BFSI, healthcare, manufacturing, and public services will add higher-value work in model optimisation, inference cost reduction, data governance, and safety evaluations. For startups, niches abound: cooling tech, AIOps, silicon validation, and energy optimisation software tuned for AI clusters.
What to watch
Key markers from now to 2029 include: procurement scale (how fast 10 GW ramps from H2 2026), advances in packaging and interconnect, energy sourcing (PPAs and storage), and supply-chain localisation for components. In India, watch for new semiconductor design centres, ATMP investments, green energy tie‑ups for AI campuses, and state-level policies aligning land, power, and high‑capacity networking with Digital India and Make in India goals.
Conclusion
The OpenAI–Broadcom plan to deploy 10 GW of custom AI chips by 2029 confirms that AI is becoming a utility‑scale industry. For India—the world’s fourth‑largest economy by nominal GDP and third by PPP—this is a strategic opening: leverage public-sector strengths, catalyse private innovation, and build competitive moats in design, packaging, data centres, and green power. The economies that align compute, energy, and talent will define the next decade of AI.
As India scales its semiconductor and AI ambitions, what specific capability—design talent, green power, advanced packaging, or data centre density—should the country prioritise first to capture the most value from this global AI compute wave?
