AI Forecasting and Predictions are a Matter of Time With InfluxData

Time is a critical element in business decision-making. To make decisions at the speed of business, it is fundamentally important that enterprises have access to relevant data in a timely manner. It is also essential, however, that data is processed and analyzed in the correct time sequence. In order to decide when to buy or sell, a trader needs to be sure that the price data they are analyzing is not only up to date but also presented in the correct order. Despite the relevance of time to all business events, time-series databases—which are optimized for the storage and processing of time-stamped data—occupy a relatively small niche of the wider data platforms landscape. That could be changing. Time-series database specialists such as InfluxData are experiencing increased interest thanks in part to artificial intelligence (AI) use cases that rely on timely and accurate data to provide forecasting, predictions and anomaly detection.

InfluxData was founded in 2012 by Chief Technology Officer Paul Dix and Todd Person (now CTO at real-time log processing provider Hydrolix) to create the open-source InfluxDB time-series database. The company has been led since early 2016 by Chief Executive Evan Kaplan, the founder and former CEO and chairman of communications company Aventail. InfluxData has attracted more than 2,600 enterprise customers in industries such as manufacturing, aerospace, energy and utilities, financial services and telecommunications. Key use cases include network and infrastructure monitoring, Internet of Things (IoT) analytics, industrial systems and predictive maintenance, as well as machine learning (ML) and AI. The company has raised $171 million in funding from a variety of investors, including Battery Ventures, Mayfield Fund and Sapphire Ventures. InfluxDB is available for self-managed deployment using the open-source InfluxDB 3 Core or InfluxDB 3 Enterprise, which provides additional security, compliance, performance and reliability capabilities. Additionally, InfluxData also offers the InfluxDB Cloud fully managed service with serverless and dedicated cloud options.

InfluxDB was initially released over a decade ago, but was significantly enhanced and updated in 2025 with the launch of InfluxDB 3. This new version of the database features a core processing engine rewritten using the Rust programming language and was re-engineered for column-oriented storage to deliver improved compression and query performance. InfluxDB 3 is based on a combination of multiple open-source projects, including the Apache Parquet column-oriented data storage format, the Apache Arrow columnar memory format, the Apache Arrow Flight SQL interface and the Apache DataFusion in-memory query planning, optimization and execution framework. InfluxDB 3 was also designed on the assumption that the data lakehouse would be the primary data storage architecture, with expanded support for the Apache Iceberg table format to take advantage of its optimized query performance, metadata management, transactional consistency and schema evolution capabilities. I assert that by 2027, more than 8 in 10 enterprises using data lakehouses will adopt open table formats to deliver support for ACID transactions and CRUD operations on data stored in object storage. Integration with data lakehouse environments is a key component of the company’s approach to AI, enabling enterprises to access time-series data in InfluxDB from any Iceberg-compatible lakehouse for AI/ML model development and training.

More recently, InfluxData updated its products with the launch of InfluxDB 3.7, which added enhanced performance and utilization monitoring with version 1.5 of the company’s Explorer user interface. This followed the previous beta launch of Ask AI for natural-language querying and support for executing custom Python code to automate workflows and transform data. In July 2025, InfluxData introduced the InfluxDB MCP server to connect InfluxDB 3 to AI tools using Model Context Protocol (MCP). In October 2025, InfluxData announced that InfluxDB 3 is now available for users of Amazon Timestream for InfluxDB, thanks to the company’s partnership with Amazon Web Services.

The reason time-series databases represent a small subset of the overall data platforms market is due in part to the widespread adoption of general-purpose relational databases, which can also be used to store and process temporal data. Whether the temporal capabilities of a relational database are sufficient for an individual use case depends on whether time is a primary or secondary attribute of the recorded metric, as well as the rate at which data is recorded and analyzed. This has significant implications for the volume of data that needs to be stored, as well as database scalability and performance requirements.

Time-series databases are optimized for the storage and processing of data for which time is the primary attribute, as well as for monitoring and measuring changes in metrics as they occur. Time-series databases are also optimized for queries that span large volumes of data to enable temporal analysis of historical trends and facilitate predictive analytics for potential future events. A key example comes from industrial IoT, where sensor telemetry data can be processed in real time for continuous asset monitoring and combined with ML models trained on historical data, as well as anomaly detection and forecasting capabilities to detect performance degradation and facilitate predictive maintenance. While these use cases can theoretically be achieved using a relational database, performance and scalability requirements favor a specialized time-series database. As such, I recommend that enterprises that have not already done so evaluate time-series databases and InfluxData for use cases that involve time-sensitive forecasting, predictions and anomaly detection.

Regards,

Matt Aslett