Home Made CobbleDB Replaces DynamoDB at Perplexity to Cut Query Latency 5x and Reduce Cloud Storage

Billy Odell Tucker-Robinson Founder & Host — Banking With Billy Network • Intelligence Network • Data Science • AI Research • World News
Published: 2026-09-25T14:47:20.408Z • Permanent link

Perplexity, a San Francisco-based startup, has made a significant move in its search infrastructure by migrating from Amazon DynamoDB to an internally developed key-value store called CobbleDB. The change, which took place in the past quarter, was led by Perplexity's CEO, Julian Graham. Graham, who has been instrumental in shaping the company's technology direction, has stated that the move was necessary to address the limitations of DynamoDB in handling large document batches.

According to data released by Perplexity, the new CobbleDB architecture has resulted in a substantial reduction in query latency, with the company reporting a 5x decrease in processing time. The company has also seen significant cost savings, with estimates suggesting that the new system will reduce storage costs by at least 30%. These results have been confirmed by industry analysts, who have praised Perplexity's decision to develop its own database solution.

Perplexity's decision to develop its own database solution has been met with interest from the broader research community, with several institutions expressing a desire to learn from the company's experience. The company has already begun sharing its knowledge and expertise with other organizations, and has announced plans to release its CobbleDB codebase as an open-source project. This move is expected to have a significant impact on the development of search infrastructure, and could potentially lead to a new wave of innovation in the field.

The decision by Perplexity to migrate from DynamoDB to CobbleDB has significant implications for companies and researchers working with large datasets. For Perplexity's competitors, such as search engine companies and data analytics firms, the move represents a major challenge to their existing database solutions. According to a recent report by Gartner, the global database market is expected to grow by 10% in the next year, driven by increasing demand for big data analytics and search capabilities.

The impact of Perplexity's decision will also be felt in the research community, where scientists and researchers are increasingly relying on large datasets to drive their discoveries. The company's new database solution is expected to make it easier for researchers to access and analyze these datasets, potentially leading to new breakthroughs in fields such as medicine, finance, and climate science. Furthermore, the open-source nature of CobbleDB is expected to drive innovation and collaboration in the field of search infrastructure, with other companies and researchers potentially adopting similar approaches to develop their own database solutions.

Perplexity's decision to develop its own database solution is part of a larger trend in the tech industry, where companies are increasingly turning to in-house solutions to meet their specific needs. This approach has been successful for companies such as Google, which has developed its own database solutions to support its search engine and other applications. However, the move by Perplexity also reflects a broader shift in the industry, as companies begin to prioritize customization and control over off-the-shelf solutions.