The financial AI company Rogo has completed a $1.6 billion Series D round, with its valuation surging from $7.5 billion to $20 billion in just three months—nearly tripling. Rogo positions itself as an AI analyst for investment banks, private equity firms, and asset management institutions, capable of performing tasks such as company research, data retrieval, valuation modeling, and client presentation preparation. By integrating foundational models from OpenAI, Anthropic, and other providers with financial data from Capital IQ, PitchBook, and similar sources, and connecting seamlessly with productivity tools like Excel and PowerPoint, Rogo automates workflows previously scattered across disparate systems. As of August 2026, Rogo has been adopted by over 300 financial institutions and serves 40,000 users. Its core competitive advantages lie in compliance infrastructure, model neutrality, and ecosystem integration, positioning it as a central workflow hub for financial institutions.

Article author and source: Wall Street Journal

AI has finally entered Wall Street.

In April this year, the financial AI company Rogo completed a $1.6 billion Series D round, valuing it at $2 billion. Just three months earlier, its valuation was $750 million.

Three months, nearly doubled.

What exactly is Rogo?

In simple terms, it’s like giving investment banks, private equity firms, and asset management institutions an “AI analyst.”

It can take over the extensive desk work that previously required junior analysts to work late nights—researching companies, finding data, building valuation models, editing Excel files, and preparing client materials.

But the interesting part is that Rogo doesn't seem to have anything particularly exclusive.

The underlying models are from OpenAI and Anthropic, financial data comes from Capital IQ and PitchBook, and the final outputs are in Excel and PowerPoint, both from Microsoft.

In other words, the most critical elements—from models and data to office software—are not inherently their own. What Rogo does is more like standing in the middle, connecting all these components together.

Logically, this type of company is most easily bypassed by上下游.

But the reality is exactly the opposite.

As of August 2026, Rogo has been adopted by over 300 financial institutions and is used by more than 40,000 financial professionals, with its valuation rising to $2 billion.

That’s interesting. Why has a seemingly mere “organizer” middle layer become one of Wall Street’s most sought-after AI companies?

Today, Silicon-based Guy will talk about Rogo and how AI began taking over the jobs of Wall Street analysts.

Complete the homework for Wall Street analysts

Investment bankers are likely the highest-paid group in the world who also perform the most repetitive manual labor.

Although they appear in suits and tie, frequenting Wall Street, they actually spend very little time sitting down with clients to conduct business. According to Wall Street Mastermind, an investment banking training institution, investment bankers spend about 90% of their time on desk work—such as gathering data, writing research reports, updating Excel models, and creating PowerPoint presentations.

Previously, Forbes interviewed the founder of Rogo and mentioned that some junior investment banking employees work up to 90 hours per week—even if they work seven days a week, that averages nearly 13 hours per day.

The reason is simple. Traditional investment banks largely operate as a "labor-intensive" business: when projects increase, they hire more analysts; when workload rises, they rely on overtime to cope.

Rogo wants to use AI to change this model.

The first step is to automate "finding data."

Investment banks conduct projects by checking financial data in Capital IQ and FactSet, reviewing funding information in PitchBook, and examining M&A transactions in LSEG. These databases have long been purchased by investment banks, but analysts previously had to manually search, filter, and copy the data.

Rogo’s approach is straightforward: they partner with these data providers to grant agents access to customers’ existing data permissions.

After integrating with Capital IQ at the end of 2024, the agents in Rogo can directly access fundamental data, market expectations, and other information to continue their analysis.

In other words, Rogo transformed previously "data for human inspection" into "production resources for agent invocation."

The second step is to deliver a usable file directly.

Felix Agent for Rogo can directly access the actual Excel models used by investment banks to update data, roll forward forecast years, refresh charts, and adjust comparable companies.

This may seem simple, but it’s actually very difficult. A real investment banking model could have dozens of worksheets filled with cross-references, historical data, and forecast assumptions. In the past, analysts had to manually pull data, update years, and verify formulas every time they made a change.

Felix can now take over directly.

For example, if the free cash flow for 2027 in the model suddenly turns negative, it can trace back through the formulas to identify which specific assumptions have changed, and then provide recommendations for adjustments.

For particularly complex models, it will even convert part of the computation logic into Python and rerun it to verify the results.

In other words, Felix began to directly deliver the outcomes of investment banking work.

Step three is to chain individual tasks into a complete transaction.

Investment banking deals may seem complex, but once broken down, they simply involve repetitive tasks such as comparable companies, precedent transactions, DCF analysis, CIMs, teasers, and screening buyers.

Rogo has encapsulated these actions into individual Agents. It engaged over 75 professionals with real-world experience in investment banking, private equity, credit, and research to help build them, and the shared Agents have now been executed more than 430,000 times.

But the standard agent is not enough, because each investment bank has different templates, guidelines, and processes.

Therefore, Rogo will continue sending teams into clients’ organizations to further customize the Agents. Just in the first week of July 2026, Rogo collaborated with clients to build over 2,450 customized Agents.

More importantly, these agents can be chained together.

One agent drafts the CIM and teaser, while another proceeds with comparable company analysis and DCF modeling, then identifies potential buyers. Tasks once scattered across different analysts are now being stitched together into an automated transaction workflow. This is where Rogo truly shines with its imagination.

Just because software can do it doesn't mean investment banks will actually use it. That's why Rogo's business model is also very unique.

It charges per seat and is essentially SaaS. As of February 2026, it has over 25,000 users and 150 institutional clients.

But the implementation is very hands-on. Rogo sends former bankers and engineers directly into clients’ organizations to help them build workflows, provide training, and modify agents.

For example, at Baird, Rogo launched approximately 600 seats. Through on-site training and ongoing follow-up, user engagement reached 95% within a few months, with around 10,000 automated tasks completed weekly, totaling over 250,000 tasks.

In other words, although Rogo sells software, its implementation is more like consulting.

If Rogo’s model proves successful, the same number of people could cover more projects in the future. For an industry that has relied on manual labor to generate computing power for decades, this is where AI’s greatest value lies.

Logically, a company like Rogo is quite risky.

Above are model providers such as OpenAI and Anthropic; below are data providers like Capital IQ and PitchBook, with clients including major financial institutions such as Goldman Sachs and Lazard.

It’s caught in the middle—any step forward by either side could bypass it.

But the reality is exactly the opposite.

As of August this year, Rogo has 40,000 users and over 300 institutional clients, including established investment banks such as Rothschild and Lazard.

Why does a seemingly easiest-to-replace middle layer become deeper over time?

The key is that Rogo has positioned itself at the three most critical areas of financial AI.

First is compliance.

Rogo was founded around 2022, but large-scale commercialization didn't occur until two years later.

Over the past two years, its main achievement has been convincing financial institutions to trust AI with their data.

Investment banks handle merger and acquisition information, client data, and internal research daily. If AI leaks MNPI or miscalculates a critical number, the consequences could directly result in compliance and reputational incidents.

Therefore, Rogo has long implemented data isolation, MNPI logical separation, granular permissions, and end-to-end audit trails, and has progressively obtained SOC 2 Type I/II, ISO 27001, and the 2025 ISO 42001 AI governance certification.

These tasks aren’t glamorous, but they’re extremely time-consuming. As a result, once commercialization kicked off, growth accelerated rapidly: we reached seven-figure ARR in less than five months, and revenue exceeded $15 million in 2025.

Second, it does not bet on models.

The biggest risk for many AI application companies is betting on the wrong underlying model.

Rogo simply doesn't lock in. It has built a Model Broker that integrates with models from OpenAI, Anthropic, Google, and others, automatically selecting the most suitable one for each financial task.

Why do this? Because there is no one-size-fits-all model for financial tasks.

Rogo created Big Finance Bench using 928 real financial tasks and found that GPT-5.5 excels in capital structure and M&A, Claude Sonnet 4.6 is better suited for financial statement analysis, and Claude Opus 4.7 performs better in private equity and forecasting.

By simply routing based on task type and data source, the overall performance can be 4.5 percentage points higher than the best single model.

The cost difference is greater. For the same task, using a top-tier model might cost $1.26, while a more suitable lightweight model can cost as little as $0.02.

This architecture means one important thing: the underlying model can be swapped out at any time, but customer data, agents, and workflows remain on Rogo.

However, what truly determines Rogo's ceiling is the third layer.

It began to squeeze its way into the core software of financial institutions.

Investment banking workflows are highly fragmented. Data resides in Capital IQ and PitchBook, internal documents are stored on SharePoint, client relationships are managed in Salesforce, emails are in Outlook, models are built in Excel, and transaction workpapers are kept in the Data Room.

Previously, analysts spent a significant amount of time moving data back and forth between these software tools.

Rogo began piecing them together.

In September 2025, we acquired Subset to embed the Agent directly into Excel, enabling it to modify complex models composed of dozens of worksheets. In March of this year, we acquired Offset to give the Agent “memory,” allowing it to understand how a model was previously built and modified.

At the same time, it integrates with SharePoint, Outlook, Salesforce, Box, and Intralinks.

By August this year, Rogo further acquired Rivanna, expanding its capabilities into M&A due diligence. The agent can process tens of thousands of transaction documents at once, providing original text citations, identifying risks, and tracking the entire due diligence process.

Individually, these are just features. But when you connect them together, things change.

Rogo is gradually consolidating the workflow of a financial transaction, which was previously scattered across a dozen different software systems, into his own layer.

If this puzzle ultimately comes full circle, Rogo will control more than just an AI tool.

It will become the gateway through which financial institutions operate daily. This is the real reason capital is valuing it at $2 billion.