Deccan AI Secures $25M to Challenge Mercor, Leveraging India's AI Talent
The demand for sophisticated AI model training and refinement is surging, and Deccan AI is poised to capitalize on this growth. The startup, specializing in post-training data and evaluation, has recently secured $25 million in its first major funding round. A significant portion of this work is powered by a highly skilled workforce based in India, positioning Deccan AI as a key player challenging established companies like Mercor in the AI services landscape.
Deccan AI: Bridging the Gap Between Model Development and Real-World Reliability
While leading AI labs such as OpenAI and Anthropic focus on building core models internally, the crucial post-training phase – encompassing data generation, rigorous evaluation, and reinforcement learning – is increasingly being outsourced. This trend is driven by the need to ensure AI systems are reliable and perform effectively in real-world applications. Deccan AI is emerging as a frontrunner in providing these essential services, offering a comprehensive suite of solutions to address the complexities of post-training AI.
Founded in October 2024, Deccan AI’s services span a wide range of capabilities. These include enhancing models’ coding and agent skills, training systems to seamlessly interact with external tools via APIs, and generating expert feedback. The company also provides advanced evaluation suites and automation platforms, catering to both frontier AI labs and enterprise clients.
Key Services Offered by Deccan AI
- Data Generation: Creating high-quality, diverse datasets for model training.
- Model Evaluation: Assessing model performance across various metrics and scenarios.
- Reinforcement Learning: Developing environments and algorithms to optimize model behavior.
- API Integration: Training models to effectively utilize external software systems.
- Helix Evaluation Suite: A product designed for comprehensive model assessment.
- Operations Automation Platform: Streamlining and automating post-training workflows.
A Growing Client Base and Impressive Revenue Growth
Deccan AI boasts a growing roster of prominent clients, including Google DeepMind and Snowflake. The company currently serves approximately 10 customers and manages around two dozen active projects simultaneously. Founder Rukesh Reddy highlights the company’s rapid growth, stating they’ve achieved a 10x increase in size over the past year, reaching a double-digit million-dollar revenue run rate. Approximately 80% of their revenue is generated by their top five clients, reflecting the concentrated nature of the cutting-edge AI market.
The Power of the Indian AI Workforce
Deccan AI’s strategic decision to concentrate its operations in India is a key differentiator. The company employs around 125 people directly and leverages a network of over 1 million contributors, including students, domain experts, and PhD holders. Approximately 5,000 to 10,000 contributors are actively engaged each month. This access to a vast and skilled talent pool allows Deccan AI to deliver high-quality services at scale.
Around 10% of Deccan’s contributor base holds advanced degrees (Master’s and PhDs), with the proportion increasing for projects requiring specialized expertise. This emphasis on quality control sets Deccan AI apart from competitors who often rely on broader, less focused contributor networks.
Why India? A Strategic Advantage
Reddy explains that focusing on a single country allows for more effective quality management. “Many of our competitors go to 100-plus countries to find the experts,” he says. “If you have operations in just one country, it becomes far easier to maintain quality.” This approach underscores India’s growing importance as a global hub for AI talent and training data, though the development of frontier AI models remains largely concentrated in the U.S. and China.
While primarily focused on India, Deccan AI is beginning to source talent from other markets, including the U.S., to address niche expertise in areas like geospatial data and semiconductor design.
Navigating the Competitive Landscape: Deccan AI vs. Mercor and Others
The AI training services market is becoming increasingly competitive, with companies like Scale AI (owned by Meta) and Surge AI, alongside startups Turing and Mercor, vying for market share. Deccan AI differentiates itself by focusing on higher-skill work from the outset, positioning itself as a “born GenAI” company, unlike traditional data labeling firms that initially concentrated on computer vision tasks.
Reddy emphasizes the critical importance of quality in post-training AI. “Quality remains an unsolved problem,” he states, adding that the tolerance for errors is “close to zero” due to the direct impact on model performance. This necessitates highly accurate, domain-specific data, which is challenging to scale.
Addressing Concerns About Working Conditions and Pay
The AI training data sector has faced scrutiny regarding working conditions and compensation for gig workers. Deccan AI addresses these concerns by offering competitive earnings on its platform, ranging from approximately $10 to $700 per hour, with top contributors earning up to $7,000 per month. This commitment to fair compensation aims to attract and retain a highly skilled workforce.
The Future of AI Training: Speed, Accuracy, and the Rise of “World Models”
The demand for AI training services is not only growing but also evolving. Reddy highlights the need for speed and accuracy, noting that AI labs often require large volumes of high-quality data within days. Furthermore, the focus is shifting beyond text-based models to “world models” that can better understand physical environments, opening up new opportunities in robotics and vision systems.
Deccan AI is well-positioned to capitalize on these trends, leveraging its skilled workforce, robust platform, and commitment to quality. The company’s success underscores the growing importance of India in the global AI value chain and its potential to become a leading force in the development and deployment of artificial intelligence.
GearTech Disrupt 2026 & Founder Summit
Stay ahead of the curve in the rapidly evolving tech landscape. Join 10,000+ founders, investors, and tech leaders at GearTech Disrupt 2026 (San Francisco, CA | October 13-15, 2026) for three days of tactical sessions, powerful introductions, and market-defining innovation. Register now to save up to $400!
For focused growth and scaling strategies, attend the GearTech Founder Summit 2026, bringing together 1,000+ founders and investors. Learn from industry leaders and connect with peers navigating similar challenges. Offer ends March 13. Save up to $300 or 30%!