Mahindra Group chairman Anand Mahindra has publicly endorsed Palantir CEO Alex Karp's theory, arguing that the true competitive advantage and lasting value in artificial intelligence will increasingly stem from the application layer rather than just the underlying frontier AI models or compute infrastructure.
The Smartphone Analogy for AI
Drawing parallels to the smartphone revolution, Mahindra highlighted in a recent post on X (formerly Twitter) and in a shareholder letter to Tech Mahindra that while the core technology of a smartphone is remarkable, its indispensable nature came from the applications and experiences built on top of it. He argues that AI's trajectory will be similar: the ecosystem that develops specialized, secure applications will create lasting value, not just the creators of chips or foundational models.
Three Layers of AI: Compute, Models, Applications
Mahindra agreed with Karp's framework, which identifies AI as comprising three interconnected layers:
- Compute: The essential computing infrastructure that powers AI operations.
- Models: The large language models (LLMs) and other generative models that produce intelligence.
- Applications: The software solutions built on top of these models, transforming raw AI intelligence into practical, enterprise-grade tools.
He emphasized that critical infrastructure cannot solely rely on a model; a robust application layer is essential for functionality and security.
Why the Application Layer is Critical for Enterprises
According to Karp, and echoed by Mahindra, enterprise AI demands safeguards and functionalities that foundational models alone cannot provide. The application layer makes AI “safe and precise” by addressing crucial business and governmental concerns:
- Data ownership and storage location.
- Prompt security.
- External information transfer protocols.
- Auditability and governance of AI-driven decisions.
These capabilities are particularly vital for highly regulated sectors such as defense, banking, healthcare, and classified government programs.
India's Strategic Imperative: Beyond Models Alone
While affirming support for India's efforts to develop sovereign frontier AI models, Mahindra suggested a broader strategy. If AI models become increasingly commoditized, with numerous open and proprietary options available, India's focus should not be exclusive to models. Instead, the country must foster an integrated AI ecosystem where compute, models, and applications evolve synergistically. This holistic approach would prevent over-reliance on any single layer and build a more resilient national AI strategy.
Advantage for AI Service Companies
Mahindra believes that AI service firms are uniquely positioned to gain a durable competitive advantage. With decades of enterprise workflow expertise, these companies can remain model-agnostic, selecting the best fit from various available models. Their strength lies in building secure, industry-specific applications that preserve business knowledge, governance, and operational control, ultimately enabling enterprises to maintain their 'alpha' or competitive edge.