Hex Raises $100M to Rebuild Analytics Around AI Agents

Hex is making room for an unusual customer journey: the analysis can begin in another company's software. Its newly introduced Hex 4.0 gives external AI agents a way to build against a company's data while keeping the resulting work available for inspection and sharing inside Hex. For an analytics supplier, that puts commercial importance on what happens after someone asks the question.

On October 27, 2026, Hex announced a $100M financing at a $1.2B valuation. Peak XV Partners and 137 Ventures led, with Amplify Partners, Avra, a16z, Redpoint, Sequoia and Zero Prime participating. CEO Barry McCardel says the capital will support team growth and go-to-market expansion as the company develops its rebuilt platform.

The Hex 4.0 documentation identifies the product as a private beta. The distinction matters for buyers: Hex is introducing a new foundation while continuing to serve an existing business, rather than reporting that every customer has already adopted the new architecture. The broader proposition is to give data teams a manageable relationship with analysis created by agents, including agents those teams did not build.

Why Hex is opening its workflow to external agents

The customer need has a familiar shape. A business user wants an answer, an analyst needs a flexible environment, and a data leader needs to understand how the organization reached its conclusion. Giving the first person a faster way to generate a chart can increase the work waiting for the other two. That is the operating problem Hex is attempting to bring inside its product.

Hex's documentation describes projects whose logic and presentation live in code files. External agents can use the workspace's context, create projects on connected data, check their work and publish apps within the workspace's permissions. A person working in Claude, Codex or Cursor can therefore reach Hex without beginning the task in its own interface. The company is treating a preferred agent as a route into the platform.

Commercially, this broadens the definition of useful analytics software. A supplier can provide value through governed execution, shared context and a place to examine results, even when it does not control the first screen. That is an interpretation of Hex's product design, rather than a forecast about which software category wins. It gives a prospective buyer a concrete question to explore: does this arrangement help a team keep track of the analysis people are already creating?

The review experience becomes part of the product

In its Hex 4.0 product announcement, Hex describes an inspection interface that connects an app element to its logic, data sources and lineage. Reviewers can move from a summary toward the code behind a result. That creates a practical route for discussing an answer without requiring every recipient to start by reading a notebook.

Consider a team deciding whether a sales chart supports an expansion plan. The chart's appearance tells the reviewer little about the customer definition, time period or filtering decisions beneath it. An inspection workflow can expose the material needed for that conversation. Someone still has to decide whether the reasoning fits the business question, but the product can make that decision easier to locate and explain.

The product announcement also describes a rebuilt execution engine, including parallel execution of independent work, streamed results and selective reruns of affected dependencies. Hex presents these as improvements to interactive apps; its internal examples should be read as company demonstrations, rather than universal performance benchmarks. Faster interaction is useful because reviewing an analysis often involves changing an assumption and seeing what moves.

Hex identifies StubHub and Ro as alpha testers in the same announcement. Those relationships offer evidence of early customer involvement, with the usual boundary around a vendor-published account: they establish named participation, rather than independent proof of a result across the wider customer base. The work of turning a promising review interface into a routine business practice continues through the beta.

A new architecture inside an established company

Hex's founding problem predates the current wave of agents. In New Relic's 2022 interview with the founders, Caitlin Colgrove and Glen Takahashi described the effort to make technical analysis easier to collaborate on and share. Barry McCardel, Caitlin Colgrove and Glen Takahashi founded the company; Hex dates its start to 2019. The tools around the problem have changed considerably during that period.

Caitlin Colgrove remains Hex's co-founder and CTO, as confirmed by the company's Prompt 2026 agenda. That engineering continuity gives the rebuild a particular character. The team replacing the platform's assumptions is also connected to the work that produced them, carrying customer knowledge into a different product architecture.

The financing history shows earlier investment in that broader analytics ambition. Hex announced a $52M Series B in March 2022 and a $70M Series C in May 2025. Those dated rounds provide context for the latest financing without requiring a guessed cumulative funding figure. The new capital arrives at a point when the company has chosen to revisit its foundations, rather than simply extend its feature list.

Where the capital meets customer adoption

Hex's careers page lists work across engineering, product, sales, partnerships and customer functions, with offices in San Francisco and New York and remote roles. Those openings fit the CEO's stated intention to expand the team and go-to-market effort. They also show that commercializing the rebuilt platform involves people who help customers evaluate, adopt and operate it.

Private beta creates a practical purchasing conversation. The current documentation lists capabilities that are not yet supported in 4.0, including scheduled runs and app notifications, writeback cells and several export or collaboration functions. An organization evaluating the new architecture will need to compare that scope with its actual workflow, while existing Hex projects continue to operate. The funding announcement supplies resources for that transition, without setting a verified adoption timetable.

Hex's proposition puts the data team's review work close to the business user's freedom to explore. A colleague may arrive with an app created in an external agent, a question about its numbers and an urgent decision to make. The relationship Hex is trying to build with that organization now extends through the explanation of how the answer came to exist, and the shared work of deciding what to do with it.