Hex announced on 27 October 2026 that it has raised $100 million at a $1.2 billion valuation, alongside a from-scratch rewrite of its data platform for the era of AI agents. The round was led by Peak XV and 137 Ventures, with participation from Amplify, Avra, Andreessen Horowitz, Next Legacy, Redpoint, Sequoia and Zero Prime. Hex 4.0, the company says, rebuilds the product around a single assumption, that everything is code, after six years on an older architecture that was never designed for agents.
Why Hex Started Over
Hex was founded in 2019, before the current AI wave, and spent four years adding AI features to a product whose core assumptions were set in 2020 and 2021. That history, chief executive Barry McCardel argues, eventually became a liability.
The company’s agent harness grew bloated with hundreds of tools needed to teach agents how to use Hex, external agents could only reach the platform through a keyhole via the Model Context Protocol, and the compute engine was not built for the high-speed, high-concurrency, iterative workflows agents prefer. Customers, meanwhile, were working around their legacy tools.
People use Cursor, Codex, Cognition, Claude and custom agents to build dashboards and reports, then share the results as HTML files, without permissions, provenance, context or observability. McCardel calls that outcome a mess, and argues the previous version of Hex could not solve it in the way the company wanted, because its data model and architecture were baked in before agents existed.
What Hex 4.0 Is
Hex 4.0 is built on plain code, not a proprietary format, with the context, controls, guidance and guardrails needed to make data work trusted wherever it happens. The idea is that teams need flexibility and agent-nativity, but also the ability to trace where a number came from, who was allowed to see it, and whether it can be relied on.
That combination, trust plus code, is the product’s central bet, and the company says the new version is rolling out over the coming weeks. The rewrite was, by the company’s own account, accelerated by using agents to write much of the code, guided by decades of combined experience building scaled data platforms.
That detail is worth noting, because it mirrors the product’s pitch: agents are not an add-on to Hex 4.0, but the primary way data work will be done inside it. Whether customers trust agent-written analytics as readily as human-built dashboards will be tested in the rollout.
The Funding, And The SaaS Argument
The $100 million round arrives after what Hex says was its best-ever quarter, with more than 2,500 customers globally, both company claims rather than independently verified figures. McCardel frames the raise against a fear that gripped software investors earlier in the year, that most software categories would die and only a few companies would remain.
His counter-argument is that many categories will combine and be reborn in new shapes, and that data work is one of them. That argument connects to a wider shift in developer tools, where agents are moving from answering questions to doing structured work across repositories and systems, as covered in Copilot Projects.
Hex is betting that the same shift will remake analytics: not a chatbot beside a dashboard, but agents that build, check and share data work directly, with governance built in rather than bolted on.
What Agents Mean For Data Teams
For data teams, the practical question is less about the valuation than about migration. A from-scratch rewrite means existing workflows, integrations and habits must move, and the company will be judged on how smoothly that move goes.
The promise is a platform where agents can work natively, without hundreds of bespoke tools, and where a result produced by an agent carries the same provenance as one produced by an analyst. That promise also appears in agent platforms for other professions, including the platform that puts accountants and AI on one financial system, as covered in Intuit’s agent platform.
The pattern is consistent across sectors: the agent era is pushing software vendors to rebuild for agents, not merely to add an assistant to an old interface, and the vendors that rebuild first are asking customers to move with them.
Why It Matters
Hex’s rewrite is a useful test case for the whole software industry, because it makes the agent transition explicit. Most vendors are adding AI features to existing products and hoping the architecture holds; Hex decided it would not, and chose the cost and risk of starting over.
If the rewrite succeeds, it will encourage other vendors to follow, and if it stumbles, it will warn them that rebuilding is harder than announcing a rebuild. The funding matters too, because it shows investors are willing to back that bet at a $1.2 billion valuation even while other software companies struggle to raise.
Model-independent platforms, in Peak XV’s apparent view, have a future precisely because they sit above any single model. Whether that future belongs to Hex, or to a rival that rebuilds faster, is now a race the company has publicly entered.
Background
Hex was founded in 2019 and built data tools before the current AI wave, launching a text-to-SQL prototype using GPT-3 before ChatGPT, Magic features in February 2023, its first data agent last year, and generative apps earlier this year.
Source: Hex announcement
0 Comments