Crunchbase today released a Model Context Protocol server that carries its private company data and predictive signals directly into large language model answers, letting investors, analysts and revenue teams query the material through plain-English questions inside the AI tools they already run. The company said the server draws on more than 39 billion live private market signals and 15 million predictions, and that its prediction models have accurately anticipated 84 percent of real-world funding events before they occurred.

The launch, dated July 21, 2026, arrives at a moment of heavy capital movement in private markets. According to Crunchbase, $510 billion was invested in startups during the first half of 2026 alone. That figure frames the pitch behind the release: as competing firms gain access to the same general-purpose AI models, the company argues, the data feeding those models becomes the point of differentiation rather than the models themselves.

What the server exposes

The Crunchbase MCP is built on the Model Context Protocol, an open standard Anthropic introduced in November 2024 to define how AI systems connect to external data sources and tools. Rather than requiring each customer to build and maintain a bespoke integration, the server routes Crunchbase's intelligence layer into any MCP-compatible environment through a single standard interface. The company listed Claude, ChatGPT, Gemini, Cursor and other AI tools and agents among the compatible destinations.

The underlying data set is substantial. According to Crunchbase, the server draws on more than 39 billion live private market signals and 790 data fields per company, alongside 15 million predictions. Those figures rest on live activity from more than 80 million users, information pulled from 1,000 public sources, and contributions from 600,000 data team members, venture capital partners and subject matter experts. The 84 percent accuracy claim attaches specifically to the prediction models that anticipate funding events.

Crunchbase drew a distinction between its approach and simpler protocol implementations. Where some MCP servers return raw search results or static lists, the company said its server identifies the filters, fields and relationships a user might not have thought to request. A user can ask which companies within a defined industry and time frame show the strongest growth signals, or which accounts are most likely to raise capital within the next year, and the server determines the relevant companies, metrics and signals to include in the response. The stated aim is to let professionals reach the depth of the data without technical fluency in query construction.

Ketaki Rao, chief product officer at Crunchbase, tied the release to the competitive logic the company has repeatedly emphasised. "When every enterprise has access to the same AI tools, the underlying data becomes the competitive edge," she said, according to the announcement. "The Crunchbase MCP provides direct AI access to our live private market data and predictive signals so decision-makers can ask questions naturally, receive trusted responses and ultimately, act with greater speed and conviction."

Three named workflows

Crunchbase described the server as built to support work across financial analysis, investing, revenue and product teams, and it grouped the use cases into three areas.

For investment teams, the company positioned the server as a way to identify high-growth companies, run diligence in compressed time frames and act on private market opportunities. For revenue teams, the framing centred on finding and evaluating targets quickly and prioritising accounts by buying power, using growth signals, funding data, exit predictions and leadership data. For product teams, Crunchbase presented the server as a route to embed private market intelligence into a product without the engineering work such integrations have historically required, with access governed by enterprise-grade security controls from the outset.

The server is available starting today, according to the company. Crunchbase framed the release as the first time the private market intelligence used by its investment and revenue customers becomes reachable through a plain-English question, without code, an integration build, or switching between tabs.

Partner reactions

The announcement carried statements from several companies building on or alongside the server, all of them presented as named individuals rather than corporate boilerplate.

Will Bryk, CEO at Exa AI, framed the partnership in terms of data quality for autonomous systems. "AI agents depend on data," he said. "Exa partners with Crunchbase to enable the highest quality company data in a new agentic search economy."

Strib Walker, head of product at Rogo, drew a line between raw information and applied context. "Financial professionals need more than information," he said. "They need relevant context they can apply directly to investment and advisory decisions. Bringing Crunchbase intelligence into Rogo through MCP helps teams surface the right company and market signals, combine them with their firm's own knowledge and workflows and move from research to execution faster."

Zubin Irani, vice president of partnerships at Glean, made a related point about combining external signals with internal knowledge. "Private market intelligence is most valuable when teams can connect it to what they already know about their customers, prospects and strategy," he said. "By bringing Crunchbase's live market signals into Glean through MCP, organizations can combine external intelligence with their own enterprise context to identify the right opportunities and act faster."

Mo Ladha, CEO at Semaverse, described the pairing of Crunchbase data with a firm's own investment thesis. "The quality of Crunchbase's private market data raises what Semaverse can do for a deal team," he said. "We hold the investment context, a firm's thesis and how it judges a deal. Bring Crunchbase's company intelligence and predictive signals into that, and the two compound rather than add. A firm assesses the entire market against its own strategy and moves on the deals that fit, with an edge neither data nor context delivers alone."

Context: a protocol that has spread across marketing technology

The release does not stand alone. It slots into an adoption pattern that has moved the Model Context Protocol from a developer-facing tool into connective tissue across advertising and marketing technology infrastructure over roughly eighteen months, a trajectory PPC Land has tracked closely.

The pattern in customer data systems began early. HubSpot launched the first CRM connector for Anthropic's Claude on July 29, 2025, letting customers query relationship management data through natural language, generate visualisations and surface insights without exporting records to a separate tool. The retail media side followed. Amazon Ads moved its own MCP server into open beta on February 2, 2026, at the IAB Annual Leadership Meeting, connecting Claude, ChatGPT and Gemini to advertising API workflows through conversational queries after a closed beta that opened on November 13, 2025.

The largest platforms extended the pattern into campaign management. Meta opened its advertising infrastructure to Claude and ChatGPT on April 29, 2026, through connectors built on its ads MCP server, and notably supported write access from day one, allowing agents to create and edit campaigns rather than only read reports. Governance moved in parallel with capability. Usercentrics acquired MCP Manager on January 14, 2026, extending consent frameworks into agentic workflows as EU AI Act enforcement began, a signal that the question of how AI systems reach business data had become a compliance concern and not only a technical one. The industry has also built advertising-specific extensions on top of the base standard, including the Ad Context Protocol, which launched in October 2025 to establish a shared vocabulary for AI agents negotiating and verifying ad placements.

Read against that backdrop, Crunchbase's server is a data provider entering the same standardised layer that platforms, CRM vendors and measurement firms have been populating. The distinguishing element is the category of data. Where most of the prior implementations expose campaign performance, analytics or customer records, Crunchbase is routing forward-looking private market predictions into the same conversational surface.

Why it matters for the marketing community

For business-to-business marketers, the release touches account targeting directly. The company positioned the server around identifying accounts most likely to raise capital and prioritising targets by buying power, using growth and funding signals. Crunchbase's private company data already feeds B2B advertising infrastructure through other channels: the company was among the initial partners when Bombora launched Curated Ecosystem Audiences on October 30, 2025, converting proprietary data from specialised B2B platforms into addressable advertising segments across programmatic and social channels. An MCP server that surfaces the same funding and growth signals inside an analyst's chat interface represents a second, more conversational route to the intelligence that already shapes some ecosystem audience definitions.

The timing also aligns with measurable demand for the underlying product. The launch follows a period of commercial growth that Crunchbase has documented. When the company introduced its Market Insights product on June 3, 2026, mapping private markets into more than 5,100 micro-industries, it disclosed that average contract value for its Predictions and Insights offering had grown 900 percent year-over-year in Q1 2026, with demand concentrated among financial services firms, investment organisations and software and AI companies seeking trusted private market data for AI-driven workflows. The company also said its two largest deals in its history had closed in the preceding two quarters. The MCP server extends that same predictive platform into a delivery format that reduces the friction of reaching it.

There is a broader structural read as well. Crunchbase's framing rests on a claim that as AI models commoditise, proprietary data becomes the durable advantage. That argument has surfaced elsewhere in coverage of the sector's funding dynamics. Crunchbase's own venture data, reported in a study covered when AI sales and marketing startups raised $3.7 billion globally in 2026, showed investors concentrating bets on companies deploying autonomous agents for marketing and customer service tasks. A data provider positioning its proprietary signals as the differentiating input for those agents is, in effect, selling into the demand its own research measures.

The near-term practical consequence is narrower. Teams that already pay for Crunchbase's Predictions and Insights can now reach the material without leaving Claude, ChatGPT, Gemini or Cursor, and product teams can embed it without the engineering lift that custom integrations demanded. Whether that convenience translates into the "competitive edge" the company describes depends on how many competing firms adopt the same standardised access to the same data, a question the adoption curve across the rest of the marketing technology stack has already started to answer.

Timeline

  • November 2024 - Anthropic introduces the Model Context Protocol as an open standard for connecting AI systems to external data sources and tools.
  • July 29, 2025 - HubSpot launches the first CRM connector for Anthropic's Claude, enabling natural language queries against CRM data.
  • October 30, 2025 - Bombora launches Curated Ecosystem Audiences with Crunchbase among the initial partners, converting private company data into addressable B2B advertising segments.
  • November 13, 2025 - Amazon Ads opens a closed beta for its MCP server, supporting Claude, ChatGPT, Amazon Q and Amazon Bedrock.
  • January 14, 2026 - Usercentrics acquires MCP Manager to govern AI data access through the Model Context Protocol as EU AI Act enforcement begins.
  • February 2, 2026 - Amazon Ads moves its MCP server into open beta at the IAB Annual Leadership Meeting, connecting Claude, ChatGPT and Gemini to advertising workflows.
  • April 29, 2026 - Meta opens its advertising infrastructure to Claude and ChatGPT with write access through connectors built on its ads MCP server.
  • June 3, 2026 - Crunchbase launches Market Insights, mapping private markets into more than 5,100 micro-industries, and discloses 900 percent year-over-year contract value growth for its Predictions and Insights product in Q1 2026.
  • July 21, 2026 - Crunchbase releases its MCP server, drawing on more than 39 billion live private market signals and 15 million predictions, and available starting the same day.

Summary

Who: Crunchbase, the private company intelligence provider headquartered in San Francisco, with named commentary from chief product officer Ketaki Rao and from partner executives at Exa AI, Rogo, Glean and Semaverse.

What: A Model Context Protocol server that carries Crunchbase's private market data, 790 data fields per company, and 15 million predictions into MCP-compatible AI tools including Claude, ChatGPT, Gemini and Cursor, drawing on more than 39 billion live signals and prediction models the company says have anticipated 84 percent of real-world funding events before they happened.

When: Announced and available July 21, 2026.

Where: Available inside any MCP-compatible AI environment through a single standard interface, without custom integration work.

Why: Crunchbase argues that as enterprises gain access to the same general-purpose AI models, proprietary data becomes the competitive differentiator, and the server routes its predictive private market intelligence into conversational AI surfaces at a moment when $510 billion was invested in startups in the first half of 2026 and demand for the company's predictive product grew 900 percent year-over-year in Q1 2026.