# Palantir CTO Shyam Sankar on AI Diffusion and Industrial Strategy An expert analysis on why AI must move beyond models to industrial workflows. - **Source:** [Stratechery](https://stratechery.com) - **Published:** 2026-10-29 - **Topic:** [Business](https://brifing.si/c/business) - **Original article:** - **On Brifing:** https://brifing.si/s/099b3f73-c787-4ca6-a9c8-5e7893e2ff0a ## The AI Diffusion Challenge Shyam Sankar argues that the current AI industry over-invests in model generation while neglecting the critical process of diffusion. Much like electricity required new machine tools to transform the economy, AI needs integrated workflows to create real value. Current focus remains too heavily skewed toward smarter models rather than practical application. ## Redefining Enterprise Productivity The atomic unit of the enterprise is the workflow, not the individual knowledge worker. Sankar notes that improving a single analyst's output often yields zero incremental ROI. True value capture requires optimizing entire supply-chain processes. While coding benefits from individual productivity gains, most real-world economic sectors require broader, systemic operational improvements. ## The Globalization Paradox Sankar critiques the historical assumption that nations can outsource production while retaining innovation. He argues that innovation is a direct consequence of productivity. Once a country stops manufacturing physical goods, it loses the stimulus required to drive technological research. This decoupling has led to significant gaps in domestic industrial capabilities. ## Meritocracy and American Agency Growing up in Lagos and Orlando shaped Sankar’s belief in the American Dream and individual agency. He views meritocracy as a fundamental precondition for societal progress. For Sankar, being American involves a profound commitment to improving the union, contrasting this with the cynicism often found in other global cultural contexts. ## The Limits of Models Despite the technical brilliance of tools like AlphaFold, Sankar points out that zero drugs have reached the market because of it. He warns against the industry becoming satisfied with 'cool' technology that lacks tangible real-world impact. AI must transition from a theoretical research pursuit to a mechanism for solving concrete problems. ## Systems Engineering vs. Safety Sankar expresses frustration regarding the current approach to AI safety. He advocates for a rigorous systems engineering methodology rather than the abstract, lab-centric debates currently dominating the field. By failing to apply engineering principles to safety, the industry risks missing the opportunity to build robust, reliable, and truly beneficial AI systems. ## Wage Growth and Stability The relationship between GDP growth and wage growth broke in 1971, leading to decades of flat inflation-adjusted earnings. Sankar believes this disconnect is the root cause of modern societal instability. He views AI as a potential tool to reverse this trend, provided it is deployed to enhance productivity across the economy. ## The Role of Forward-Deployed Engineers Palantir utilizes forward-deployed engineers to bridge the gap between complex software and real-world operational challenges. These engineers work directly within client environments to ensure technology solves specific, high-stakes problems. This hands-on approach is essential for diffusing intelligence into sectors that are traditionally resistant to rapid digital transformation and technological adoption. ## The Religious War in Tech Sankar describes the current state of the technology industry as a 'religious war' regarding the future of AI. This conflict pits those focused on pure model development against those advocating for practical, industrial-scale diffusion. The outcome of this debate will determine whether AI becomes a transformative economic force or a niche. ## Preserving Alpha in Business Companies that focus on top-line growth tend to be more successful in adopting AI than those focused solely on cost-cutting. Sankar emphasizes that organizations must learn how to preserve 'alpha'—their unique competitive advantage—rather than settling for median performance. AI should be used to amplify these distinct institutional strengths and capabilities. ## The Necessity of Production Innovation is not a standalone activity but a byproduct of engaging with physical production. Sankar uses the example of Google Translate, which was improved by 3% to solve a specific, prosaic problem. This small, practical goal eventually led to the profound research breakthroughs that define the modern generative AI landscape today. ## Future of AI Diffusion The ultimate success of AI depends on its ability to inflect GDP and improve societal outcomes. If the industry fails to move beyond convenient AI assistants, it risks losing its license to operate. Sankar’s perspective underscores the urgent need for a shift toward systems engineering to ensure AI delivers lasting value. ## Sources covering this story - https://stratechery.com/2026/an-interview-with-palantir-cto-shyam-sankar-about-the-ai-religious-war/ — --- _Summary written by Brifing from the original reporting. Copyright in the original article belongs to its publisher; read the full piece at the link above._