Exchange CEO's "Titanic" Quip Signals Market Data Reckoning as AI Reshapes Finance Workflows
Euronext CEO Stéphane Boujnah delivered an unusually blunt assessment of the financial data industry this month, suggesting the much-hyped "data revolution" may have been oversold. "Maybe this data boat was a Titanic boat, that we missed the Titanic boat," Boujnah told investors in February 2026, pushing back against years of criticism that exchanges weren't capitalizing on their information assets.
The comment arrives as AI coding tools begin penetrating finance workflows in unexpected ways, raising questions about which data moats actually matter. Marc Rubinstein, a veteran finance newsletter writer, spent last week using Anthropic's Claude Code to build his first software application in four decades—a system that indexes his five-year archive of financial commentary and surfaces relevant news stories each morning.
The experiment highlights a shift that should concern CFOs: AI tools are making bespoke data analysis accessible to non-engineers, potentially undermining the value of standardized market data products that exchanges and terminal providers have sold for premium prices. Rubinstein's application required no traditional programming skills, yet it learned his editorial patterns from 250-plus archived issues and now delivers curated story suggestions "remarkably well attuned" to his coverage areas.
"For a newsletter writer, I may be late to Claude Code," Rubinstein wrote, noting tech-focused publications adopted the tool earlier this year. But adoption metrics suggest the technology remains nascent even among early users. Anthropic reports the median Claude Code session lasts just 45 seconds, with only 0.1% of users maintaining sessions longer than 40 minutes. The company may have no more than one million active users total.
Within regulated financial institutions, uptake appears slower still. Rubinstein cited a survey of 150 quantitative analysts and research professionals—though the source document cuts off before revealing results—suggesting compliance and data security concerns may be dampening enthusiasm in finance departments.
The technology's current limitations are telling. While Claude Code successfully indexed Rubinstein's public archive and matched news to his editorial style, it cannot access private information repositories or replicate his writing voice. "Could I add a module to get Claude to write the whole thing for me? Perhaps," he wrote, before noting the AI lacks access to personal experiences, network contacts, or the ability to draw connections using proprietary insights.
Those constraints matter for finance applications. A CFO evaluating AI coding assistants must consider whether the tool can access internal financial systems, comply with data governance policies, and integrate with existing analytics workflows—questions that extend beyond pure technical capability.
Boujnah's "Titanic" metaphor suggests exchange executives are rethinking their data strategies as AI tools democratize analysis. If finance professionals can build custom applications that process public information more efficiently than standardized terminals, the premium pricing for market data feeds faces pressure. The CEO's comments imply Euronext may have dodged a bullet by not over-investing in data products that AI could soon commoditize.
For finance leaders, the immediate question is whether to experiment with AI coding tools for department-specific applications or wait for enterprise-grade solutions with proper controls. Rubinstein's 45-second median session statistic suggests most users are still testing rather than deploying these systems for critical workflows.
The broader uncertainty: whether Boujnah's data skepticism proves prescient, or whether he's simply rationalizing Euronext's historical underinvestment in information products just as AI makes proprietary datasets more valuable than ever.
Responses (0 )