From Model-Centric Hype to Data-Centric Reality
Enterprise data infrastructure is the collection of platforms, databases, and governance tools that store, process, and operationalise data so that AI systems, analytics, and business applications can use it reliably at scale. That infrastructure is now the centre of gravity for enterprise AI spending. While foundation models still draw headlines, companies are learning that models without clean, well-governed data do not reach production or deliver value. ClickHouse and Databricks sit squarely in this “AI data platforms” layer, and their soaring traction shows where budgets are moving. Investors are rewarding companies that can make both structured and unstructured data usable for AI agents, automation, and analytics, rather than those focused only on training bigger models. The new question inside boardrooms is less “Which model?” and more “Which platform will turn our data into a production-grade AI asset?”.
ClickHouse: Real-Time Analytics Becomes an AI Bottleneck
ClickHouse’s growth highlights how real-time analytics is becoming a key bottleneck for AI in production. The company reports an annualised revenue run rate above USD 250 million (approx. RM1,150 million), tripling year over year, and expects this to reach the high nine digits by year end. Its open-source, columnar database is tuned for massive, high-speed analytical workloads, which are essential for AI agents that must query recent events, logs, and user behaviour. Managed cloud services now generate most of its commercial revenue, as more than 4,000 customers, including Anthropic, Meta, and Capital One, embed ClickHouse into AI-heavy workflows. Recent acquisitions such as Langfuse, which tracks and evaluates AI agent performance, show the company stepping further into AI observability. By targeting startups that extend its core database, ClickHouse is building an enterprise data infrastructure stack aimed at making AI answers both fast and measurable.
Databricks’ Valuation Shows Confidence in AI Data Platforms
Databricks’ latest funding round puts its data infrastructure valuation into rare territory and sends a signal about where investors think enterprise AI value sits. According to Inc., the company raised more than USD 7 billion (approx. RM32,200 million), including USD 5 billion (approx. RM23,000 million) in equity financing, at a USD 134 billion (approx. RM616,400 million) valuation. Databricks has positioned itself as an AI data platform built around its “Data Lakehouse”, combining structured and unstructured data in a single environment. Its annual revenue run rate reached USD 5.4 billion (approx. RM24,840 million), with fourth-quarter revenue up more than 65% year over year. That pace, especially at this scale, suggests enterprises are consolidating data pipelines, governance, and AI workloads on fewer platforms. Rather than buying separate tools for data lakes, warehouses, and AI pipelines, buyers are gravitating toward unified enterprise data infrastructure that can support agents, analytics, and automation together.
Data Infrastructure as the Critical AI Production Constraint
Both ClickHouse and Databricks frame data infrastructure as the critical constraint on AI value inside large organisations. Databricks competes with Snowflake to control the layer where enterprise data is prepared, governed, and exposed to AI models, ERP systems, and agentic workflows. This competition goes far beyond storage or SQL queries; it covers unstructured documents, transaction feeds, and model-customisation pipelines needed for AI to act inside finance, supply chain, HR, and customer operations. ClickHouse, meanwhile, attacks the latency and scale problems of real-time analytics, providing the high-throughput queries AI systems need to respond to fast-changing data. Their acquisitions of AI and data-management startups show that the contest now includes observability, model evaluation, and transaction-level connectivity. In effect, AI models have become modular, while the messy work of data readiness remains the true differentiator for production AI.
Enterprise AI Spending Shifts to Data-Centric Foundations
The valuations of ClickHouse and Databricks are part of a wider shift in enterprise AI spending patterns. Budgets that once chased headline-grabbing models are now moving toward AI-ready data foundations that can support many models over time. For ERP and application leaders, this means long-term commitments to AI data platforms, not one-off model experiments. Databricks and Snowflake are racing to become the default enterprise data infrastructure layer, each acquiring AI and data startups to widen their reach into governance, model training, and real-time transaction data. ClickHouse’s rise shows that specialised databases for high-speed analytics also have a place in this stack, especially for AI agents that depend on current, queryable data. Together, these trends signal a model-agnostic future: enterprises will plug different models into a stable, governed data platform, with spending dominated by the systems that keep data reliable, accessible, and ready for AI.