The Embodied Conglomerate: Decoding Mukesh Ambani’s AI Manifesto

On December 30, 2025, Mr. Mukesh Ambani, Chairman and Managing Director of Reliance Industries Limited (RIL), addressed the conglomerate's workforce of 4,00,000+ employees. In his address, he unveiled the draft Reliance AI Manifesto. Mr. Ambani described artificial intelligence (AI) as "the most consequential technological development in human history" and positioned RIL to lead India's AI revolution, mirroring its earlier leadership in the country's digital transformation. He emphasized that the world has only seen "the tip of the iceberg" of AI's possibilities, which can solve complex global problems if used wisely.

Mr. Ambani called for a transformative adoption of AI to turn RIL into an AI-native deep-tech company with advanced manufacturing capabilities. The core ambition is to achieve a ten-fold (10x) improvement in productivity, velocity, efficiency, quality, and outcomes across the entire workforce by fundamentally re-thinking workflows.

This corporate message has been widely read as a motivational or operational directive. That reading, I argue, is incomplete. For a conglomerate of RIL’s scale, sprawl, and complexity, such a statement functions less as internal morale-building and more as an institutional signal: about where intelligence will sit in the organisation, how capital will be deployed, and what kind of enterprise RIL intends to become in the AI era.

The real significance of the message, as I see it, lies not in the number '10x', but in where that productivity is expected to materialise—inside physical systems, industrial processes, and asset-heavy operations that historically resist rapid efficiency gains. Read this way, the message aligns with a deeper strategic pivot: RIL positioning itself as an embodied AI conglomerate, rather than a cloud- or software-first adopter.

From Digitisation to Closure: Eliminating Operational Latency

At the centre of Mr. Ambani’s articulation is a multi-layered Digital Functional Core, designed to eliminate what he describes as “digital breaks”—manual or semi-manual transitions between data capture, analysis, and execution. This framing is important. It acknowledges that most large enterprises are already “digitised” in a superficial sense, yet remain structurally inefficient because intelligence does not flow continuously into action.

RIL's move toward smaller, cross-functional execution units—described internally as 'pods'—suggests an attempt to close control loops across procurement, production, logistics, energy management, and retail operations. The ambition is not simply faster reporting or better dashboards, but autonomous or near-autonomous workflows where sensing, decision-making, and actuation occur with minimal human mediation.

For a conglomerate spanning India's biggest crude-oil refinery, telecom network, retail chain, and (soon to be) India's biggest renewable energy equipment ecosystem, this is not an IT upgrade. It is a re-architecture of how decisions are embedded into matter.

Why Intelligence Must Move to the Edge

This is where the discussion shifts from generic enterprise AI to something more consequential: embodied and edge-deployed intelligence.

In asset-intensive environments—refineries, petro-chemical plants, giga-factories—latency is not a technical inconvenience but an economic constraint. Control decisions delayed by even tens of milliseconds can translate into energy loss, throughput inefficiencies, safety buffers, or excess wear. In such contexts, the logic of cloud-first AI weakens rapidly.

Recent discussions in the global AI community, especially around edge computing and AI-embedded devices and chips, point to an emerging consensus:  sustainable and meaningful AI must reside inside the device/machine, not merely observe it from afar. Thus, in the case for industrial AI, sensors, valves, turbines, conveyors, robotic arms, and power electronics increasingly need to host their own inference and optimisation layers.

For RIL, this is less a technological preference than a structural necessity. Its scale and physical dispersion make centralised intelligence economically suboptimal. Embodiment—AI embedded into hardware and processes—is the only viable path to sustained productivity gains.

The Partner Play: Google’s Role in the Pivot

Through its subsidiary Reliance Intelligence, RIL has struck a massive deal to use Google’s AI stack, including their specialized Tensor Processing Units (TPUs), to power RIL’s 1 GW data centers.

While Google provides the "high-speed engine" (Gemini 2.5 Pro and TPU-powered cloud), RIL is layering its own JioBrain on top. JioBrain acts as the "driver" that understands the specific Indian industrial context, ensuring that while the tech is global, the application is purely Reliance.

The Strategic Convergence

Seen through this lens, the Reliance–Google AI partnership takes on sharper meaning.

Google is a global leader in foundational models, distributed computing, and AI tooling—but it remains comparatively underexposed to continuous, real-world industrial deployment. RIL, conversely, operates one of the densest concentrations of physical infrastructure under unified ownership anywhere in the world, but does not seek to reinvent foundational AI from first principles.

The partnership, therefore, appears less transactional and more complementary. Google contributes scalable cognition; RIL provides matter, deployment surfaces, and demand certainty. If successful, this convergence could produce AI systems trained not merely on text and images, but on decades of operational data from energy systems, supply chains, and industrial processes — precisely the domains where AI’s economic impact has historically been slow to materialise.

Institutional Memory as a Force Multiplier

One particular advantage RIL holds over younger companies is the depth of its institutional memory. Decades of operational data, process refinements, and tacit engineering knowledge already exist within the organisation. AI, in this context, functions less as a replacement for human expertise and more as a mechanism for stabilising and replicating it at scale.

Encoding experienced decision-making into models—whether for maintenance, energy optimisation, or supply planning—allows that knowledge to persist beyond individual tenures. This is not digital immortality in a literal sense, but it is a meaningful shift in how large organisations retain and transmit competence.

A New Industrial Revolution

The implications of this embodied AI approach extend across RIL's entire industrial footprint, with the potential to fundamentally reshape how each sector operates.

In green energy, the integration of AI into manufacturing processes could dramatically accelerate RIL's path to producing 100 GW of renewable energy equipment. By deploying models that can simulate and optimize assembly in real-time, the company aims to manufacture solar photovoltaic modules and electrolyzers with unprecedented speed and cost efficiency. This isn't merely automation—it's the compression of the entire learning curve of manufacturing into systems that continuously refine themselves.

Within conventional energy and petrochemicals, the transformation is equally profound. Legacy infrastructure—refineries, chemical plants, and their complex arrays of distillation columns, reactors, heat exchangers, and rotating equipment—can be retrofitted with embedded intelligence that transforms them from static assets into adaptive systems. Critical process equipment becomes capable of self-diagnosis and predictive maintenance, anticipating component degradation before it manifests, optimizing energy flows and reaction conditions in real-time, and sustaining operations at theoretical maximum efficiency with minimal downtime. The result is infrastructure that ages more gracefully while extracting more value from every barrel processed and every molecule transformed.

The retail and supply chain operations present perhaps the most visible transformation. AI-driven logistics enable what RIL describes internally as the "Batch of One"—hyper-personalized inventory management that drastically reduces waste while improving responsiveness. For sectors like textiles and fast-moving consumer goods, where excess inventory and supply-demand mismatches have always been costly, this represents a fundamental shift. Products reach Reliance Trends or JioMart outlets with what Mr. Ambani calls "surgical precision," guided by predictive models that understand demand patterns at granular geographic and temporal scales.

What unifies these applications is a common architectural principle: intelligence moves from being a service consumed by the organization to being a capability distributed throughout it. RIL's ambition is to function as a self-optimizing organism, where autonomous or semi-autonomous systems handle routine optimization, allowing human expertise to focus on strategic decisions and exceptional circumstances. 

In this vision, AI becomes not a tool wielded by the enterprise, but part of its operational substrate—embedded in the matter and processes that constitute industrial activity itself.

Conclusion: The Emergence of the AI-Native Conglomerate (and Disruptor)

Mr. Ambani’s AI emphasis should be read not as hype, but as a declaration of intent: RIL aims to embed intelligence directly into the physical substrate of the economy. 

If this RILxAI strategy succeeds, RIL will not just become more productive—it will help define what an AI-native industrial ecosystem looks like in practice. 

And in that process, RIL could become a market-disruptor (again) — by crashing the cost of industrial intelligence for all of India.