The Fall of FiscalNote: How AI Disrupted a Policy Intelligence Giant

From NYSE Darling to OTC Trading: A Cautionary Tale

In a move that speaks volumes about the rapidly shifting landscape of technology and data-driven business models, FiscalNote Holdings found itself unceremoniously booted from the New York Stock Exchange on March 25. The AI-powered policy intelligence company, which once positioned itself as an essential bridge between complex government regulations and the businesses that need to understand them, couldn’t maintain the minimum $1.00 average share price required for NYSE listing over a 30-day trading period. What makes this particularly painful is that the company had already burned through its one allowed remedy—a reverse stock split executed within the previous twelve months—leaving no lifeline available when shares tumbled again.

The delisting represents more than just a technical failure to meet exchange requirements; it’s a stark illustration of how quickly the ground can shift beneath even seemingly innovative companies in the age of artificial intelligence. FiscalNote built its entire business on being the essential middleman between mountains of public policy data and the organizations that needed it processed, analyzed, and served up in digestible formats. But as large language models became increasingly sophisticated and accessible, that carefully constructed moat evaporated almost overnight. The company that once promised to democratize access to policy intelligence found itself disrupted by technology that truly democratized it—by making the interpretation of policy documents essentially free for anyone with access to ChatGPT or similar tools.

The Ironic Timing of Success and Failure

There’s a bitter irony in the timing of FiscalNote’s delisting that perfectly captures the company’s predicament. On the very same day the NYSE pulled the plug on its listing, FiscalNote announced a seemingly major win: its PolicyNote MCP server had been approved for the OpenAI App Store. This integration would theoretically give ChatGPT’s staggering 700 million weekly active users direct access to structured policy data covering Congress, all fifty U.S. states, and more than 100 countries worldwide. For most companies, landing a spot in OpenAI’s ecosystem would be cause for celebration and likely a boost to stock prices. For FiscalNote, it coincided with the ultimate market rejection.

This juxtaposition tells the entire story of FiscalNote’s strategic confusion. The company seemed caught between two incompatible visions of its future: remaining a premium subscription service that charges enterprises hefty fees for policy analysis, or becoming infrastructure that powers AI tools. The OpenAI integration represents a tacit admission that FiscalNote’s human-curated analysis couldn’t compete with what AI models can generate instantly from raw data. By pivoting to provide data feeds for AI systems rather than finished analysis for human customers, FiscalNote essentially admitted that its original value proposition had been commoditized. The company tried to spin the delisting as “the start of a new phase of health and opportunity,” pointing to workforce reductions of 25% and operational cost cuts of 19%. Management projected positive free cash flow beginning in April 2026—a timeline that asks investors to maintain faith for another year while the company bleeds cash and searches for a sustainable business model.

Understanding the SaaSpocalypse: When AI Eats Your Lunch

FiscalNote’s struggles fit neatly into a broader pattern that market analysts have dubbed the “SaaSpocalypse”—a term that would be funny if it weren’t describing the destruction of billions of dollars in shareholder value. Since early 2026, the software-as-a-service sector has lost approximately $1 trillion in market capitalization as enterprises fundamentally rethink their technology spending. The culprit isn’t a recession or a scandal; it’s the rapid rise of AI agents that can perform many of the same functions as specialized software subscriptions, but faster, cheaper, and often better.

FiscalNote’s business model depended on a specific type of information asymmetry that the internet age had somehow left intact. Government policy documents have always been public—transparency is supposedly a cornerstone of democracy—but actually collecting those documents from hundreds of different sources, structuring the data consistently, tracking changes over time, and interpreting what it all means was genuinely difficult and expensive. FiscalNote invested in building databases, hiring policy experts, developing tracking systems, and creating user interfaces that made this information accessible to corporate compliance teams, government affairs professionals, and lobbying firms. Companies paid substantial subscription fees because the alternative—doing it themselves—would have cost even more.

Then large language models arrived and demolished this entire value chain. A compliance officer can now copy the text of a pending bill, paste it into ChatGPT, and get a detailed summary, impact analysis, and list of relevant stakeholders in seconds—all for the marginal cost of a few cents in compute time. The same transformation happened to FiscalNote that happened to travel agents when Expedia launched, or to encyclopedia salespeople when Wikipedia went live. The difference is the speed: where those earlier disruptions unfolded over years, the LLM revolution has compressed similar transformations into months. FiscalNote went from viable business model to existential crisis in essentially one product cycle.

Desperate Pivots: When a Company Tries Everything

Watching FiscalNote’s strategic pivots over the past year feels like watching someone throw everything at the wall to see what sticks, guided more by whatever’s trending in tech headlines than by any coherent vision. In June 2025, the company announced it was evaluating stablecoins as a payment option for international customers—a move that seemed to address a problem no one knew FiscalNote had. By September, the company had jumped on the corporate Bitcoin treasury bandwagon, announcing exploration of Bitcoin, Ethereum, and Solana as strategic reserve assets. This followed the playbook popularized by MicroStrategy, whose CEO Michael Saylor had turned corporate treasury management into a leveraged bet on cryptocurrency.

Then came the prediction markets phase. In February 2026, FiscalNote launched PoliticalPredictions.com and signed a non-binding memorandum of understanding with 365Prediction, bringing on a former Sportradar executive as strategic advisor. Most recently, the company signed another MOU with a Korean law firm to distribute U.S. policy intelligence across Asian markets. Each of these moves taps into legitimate trends—corporate crypto adoption is real, prediction markets are booming, and there’s genuine demand for understanding U.S. policy in international markets. But the scattershot approach suggests a company in panic mode rather than executing a thoughtful transformation.

None of these initiatives has generated meaningful revenue, at least not enough to stabilize the stock price or convince investors that FiscalNote has found its next act. The crypto treasury move looks particularly questionable in hindsight, given that it was announced when digital assets were near cyclical peaks—meaning FiscalNote may have bought high during its brief window of available cash. The stablecoin payment processing angle solves a problem that established financial infrastructure already handles perfectly well. And while the prediction markets push has more strategic logic (which we’ll explore), it’s unclear whether FiscalNote has the resources, expertise, or time to build a competitive position in that space.

The Prediction Markets Angle: Promising Concept, Uncertain Execution

Of all FiscalNote’s recent strategic experiments, the prediction markets initiative actually makes the most intuitive sense, even if the execution remains questionable. Prediction markets have exploded in popularity and legitimacy, with monthly trading volumes now reaching approximately $10 billion. Kalshi has emerged as the market leader with roughly 66% market share, overtaking Polymarket and bringing prediction markets into the mainstream of financial and political analysis. Major media outlets now regularly cite prediction market odds alongside traditional polling, and sophisticated investors use these markets as real-time probability engines for scenario planning.

The conceptual fit with FiscalNote’s expertise is obvious. Companies subscribing to policy intelligence services don’t fundamentally need to know what a particular bill or regulation says—that text is publicly available. What they actually need is forward-looking probability assessment: Will this bill pass? When? In what form? What are the chances this regulatory agency actually enforces this provision? These are inherently probabilistic questions, and prediction markets are arguably the best mechanism humans have invented for aggregating dispersed information into probability estimates. FiscalNote’s decade of experience tracking legislative processes and understanding policy dynamics could theoretically provide edge in structuring and analyzing prediction markets focused on regulatory outcomes.

But there’s a structural problem that may prove insurmountable. Prediction markets require liquidity, and liquidity follows attention. High-profile events like presidential elections, Federal Reserve interest rate decisions, or the outbreak of wars attract enough interested participants to create deep, efficient markets. The obscure regulatory questions where FiscalNote’s specialized knowledge would be most valuable—”Will the EPA finalize updated PFAS regulations by Q3?” or “What’s the probability the EU’s AI Act Article 6 high-risk classification criteria expand to include large language models?”—are precisely the ones too niche to attract sufficient betting volume. Without liquidity, prediction markets fail to perform their core function of price discovery, becoming just another illiquid derivatives market with wide spreads and unreliable signals. FiscalNote’s board has indicated it’s reviewing all strategic options, including selling off non-core assets, but it’s unclear what exactly counts as “core” for a company that’s tried to reinvent itself multiple times in twelve months.

The Broader Lesson: Information Middlemen in the Age of AI

FiscalNote’s journey from SPAC merger darling in 2021 to OTC-traded penny stock in 2025 offers a sobering case study in how artificial intelligence is reshaping the economics of information. For decades, successful information businesses followed a reliable playbook: identify a domain where valuable data exists but is scattered, unstructured, or hard to access; invest in aggregating and organizing that information; build tools that make it searchable and analyzable; then charge subscription fees to customers who need it. This model worked brilliantly across countless industries—financial data (Bloomberg), business information (Dun & Bradstreet), real estate (CoStar), and policy intelligence (FiscalNote).

Large language models have fundamentally undermined this entire category of business by collapsing the cost of the interpretation layer to nearly zero. The expensive part used to be taking raw information and turning it into insight. Now, AI can perform that transformation instantly and cheaply for anyone with access to the underlying data. This doesn’t eliminate the value of having good data—garbage in, garbage out still applies—but it dramatically reduces what customers will pay for it. Why pay thousands per month for a service to summarize and analyze policy documents when you can paste those same documents into ChatGPT and get similar analysis for pennies? The companies that survive this transition will be those that either own truly proprietary data that AI companies can’t easily replicate, or that successfully move up the value chain to provide services that AI can’t yet match. FiscalNote tried to do the latter with its various pivots, but none have gained traction quickly enough to offset the collapse of its core subscription business. Whether the company can complete a transformation from the weakened position of OTC trading, with constrained access to capital and damaged credibility, remains highly uncertain. What’s certain is that FiscalNote won’t be the last information middleman to discover that the moat they spent years building can be drained in months by sufficiently powerful AI. The delisting is just one data point in a much larger trend that’s still unfolding across the entire information economy.