NextFin News - Rogo Technologies, the AI startup founded by former junior bankers to automate the very work they used to do, is setting up its first Asia-Pacific office in Singapore - a move that puts finance's most exposed workforce at the center of the industry's next geographic shift. The New York-based company, valued at roughly $2 billion after a $160 million Series D in April, is expanding to Singapore as its first APAC base and hiring across the region, CEO Gabriel Stengel announced while promoting a July 30 gathering with Singapore's banking and private-equity community.
The expansion matters because it is not a sales outpost in the usual sense. Rogo's product is a generative-AI platform built to perform the research, modeling, and slide-deck work that has defined the bottom rungs of the banking pyramid for decades - and it is now taking that product into one of the fastest-growing deal regions on the map. The combination is a live stress test of a claim Wall Street is still arguing over: that artificial intelligence will not just assist junior bankers, but permanently absorb a large share of the work that once justified hiring them in the first place.
The News: A First APAC Base, Built on a $2 Billion Bet
Rogo was founded in late 2021 around a kitchen table in a Manhattan apartment by Gabriel Stengel, who left his investment-banking analyst job at Lazard, and John Willett, a former JPMorgan banker and fellow Princeton computer-science graduate. They were later joined by Tumas Rackaitis, a software engineer and former trader at a New York hedge fund, as chief technology officer. The founding premise was blunt: the people doing the grunt work were the ones best positioned to code its replacement.
That premise has been repriced by venture capital at an escalating rate. In March 2025 the company raised $50 million at a $350 million valuation. In January 2026 it took $75 million in a Series C led by Sequoia Capital and opened its first international office in London. On April 29, 2026, Rogo announced a $160 million Series D led by Kleiner Perkins, with participation from Sequoia, Thrive Capital, Khosla Ventures, J.P. Morgan Growth Equity Partners, BoxGroup, Mantis VC, Jack Altman, Evantic and Positive Sum, bringing total funding to more than $300 million. The round landed at an approximately $2 billion valuation.
The Singapore move follows that capital. Stengel described it in a company post as Rogo's "first office in APAC," and the company has been posting Singapore-based roles - including Customer Support Manager for APAC, Solutions Architect for APAC, and Enterprise Account Executive for APAC - through mid-August 2026. An early APAC hire wrote that he was "launching the APAC business" and that the company was hiring in Singapore across nearly every function.
The timing is deliberate. Asia-Pacific dealmaking has been recovering, and Singapore has positioned itself as the regional hub for wealth, private markets, and technology - the exact client mix Rogo targets. The company says it is now used by more than 50,000 financial professionals at over 350 institutions, with more than 150,000 queries sent through the platform daily. In April, when the Series D closed, those figures stood at more than 35,000 users and over 250 institutions, including Rothschild & Co, Jefferies, Lazard, Moelis and Nomura. The customer list has also included JPMorgan, Bank of America and Wells Fargo.
What those customers are buying is a platform that sits inside a bank's own systems - SharePoint, CRM, market-data terminals, filings and research - and produces what Rogo calls "institutional-grade outputs": auditable Excel models, investment memos, diligence materials and pitch decks. In April the company introduced Felix, an agentic AI that it says can execute multi-step financial processes autonomously, from deal screening and confidential-information-memorandum generation to buyer outreach and data-room diligence.
"The world's most sophisticated financial institutions are fundamentally reshaping how they operate using AI, and they're choosing to do it with Rogo," Stengel said when announcing the Series D. "The institutions at the forefront are rapidly moving beyond automating tasks to becoming AI-native firms, with agentic systems that work across the firm and get smarter with every deal."
The Singapore office is the geographic expression of that pitch. It is also the latest data point in a labor story that has moved from speculation to observable fact.
Why Junior Bankers Are the Front Line of the AI Shift
To understand why a tool like Rogo lands hardest on junior bankers, start with what those bankers actually do. An analyst's week is dominated by tasks that are data-heavy, repetitive, and governed by strict formatting conventions: pulling comparable-company figures from filings, building and updating valuation models, formatting pitch books, taking notes on calls, and assembling diligence materials. These are not the relationship or judgment tasks that senior bankers sell to clients. They are the apprenticeship tax - the thousands of hours juniors pay in exchange for learning the craft and, eventually, moving up the pyramid.
Generative AI is unusually well matched to that work. It reads documents, extracts structured numbers, drafts text in a house style, and can be constrained to produce outputs in formats that already exist. That is why the specific automation estimates point squarely at the bottom of the pyramid. Goldman Sachs research has suggested AI could automate about a quarter of work hours, with junior banking tasks - financial modeling, note-taking, spreadsheet analysis and deck formatting - the "most exposed." A Citigroup report found that 54% of banking jobs have high automation potential, the highest of any sector the bank analyzed.
The counter-pressure is real, and it is where the debate lives. In EY's latest Global Financial Services CEO Outlook Survey, 60% of the 240 financial-services CEOs surveyed said they believe investment in AI will maintain or increase headcount in 2026, compared with 28% who expect AI to drive a reduction. The argument behind that number is that productivity gains create capacity for more deals, that compliance and client work still require humans, and that someone has to check the machine's output. Then there is the apprenticeship problem: if banks stop giving juniors the grunt work, how do they ever become seniors who can exercise judgment?
Debasish Patnaik, senior partner and leader of QuantumBlack, McKinsey & Co.'s AI consulting arm, put the tension in one line: "Banking is an apprenticeship business. Today's junior analysts become tomorrow's managing directors. Senior judgment cannot be manufactured laterally." He also noted that banks are cutting junior analyst classes by as much as two-thirds while sourcing roughly 62% of their AI talent from those same cohorts - shrinking the very pipeline that feeds their AI transformation.
Rogo's founders acknowledge the anxiety and reframe it. Their prediction is that junior bankers will benefit by being freed from grunt work so they can take on more meaningful roles earlier in their careers. Stengel has gone further, predicting the technology will spawn "AI-first" investment banks, where staff focus on higher-value work from the start. In other words, the pyramid does not just shrink - it changes shape.
That is the crux of the cyclical-versus-structural question, and it deserves a clean answer. A cyclical read would say today's smaller analyst classes are a reaction to a slow deal market and cost pressure, and that hiring will rebound when fees recover. The structural read - which the evidence increasingly supports - is that the tasks being automated are not coming back regardless of the deal cycle. A model that builds comps does not need to be re-hired when M&A volumes pick up. The cost of producing a pitch book has fallen, and in a competitive industry, lower costs do not stay with the firm; they get passed to clients as fee pressure and to employees as a smaller base.
The Singapore expansion sharpens this because Asia is where banks have been most eager to grow headcount. If the automation thesis were merely cyclical, banks would be hiring armies of analysts in Singapore right now. If it is structural, a tool like Rogo lets them grow coverage without growing the analyst ranks proportionally. The company's own hiring pattern is a tell: the Singapore roles are weighted toward support, solutions architecture, and go-to-market - the functions that scale a software business - rather than a large local army of finance analysts.
The Second-Order Battle: Who Owns the Intelligence Layer
The first-order story is headcount. The second-order story - the one that matters more for where the money goes - is about who owns the system of record for financial knowledge. When a bank's analysts build models and memos, the institutional knowledge lives in the bank: in the analysts' heads, in the files on the server, in the relationships. When that work moves into a platform like Rogo, a portion of that knowledge migrates into the platform's workflows, its fine-tuned finance models, and its logs of what good output looks like.
This is why the funding math is so aggressive. Kleiner Perkins partner Mamoon Hamid said of Rogo: "When a platform becomes the operating system for an entire industry, the opportunity is generational." That sentence is not about saving analyst hours. It is about becoming the layer through which financial work is done - the place where the firm's data, its reasoning, and its output all pass. Operating systems capture more value than the applications that run on them, and they are notoriously hard to dislodge once embedded.
The moat, in this framing, is not the frontier model. Rogo itself says it is exploring a range of AI models, and CTO Tumas Rackaitis has said the company is "investing heavily in reinforcement learning which I think is going to be a major differentiator for us." The real barriers are distribution inside conservative institutions, deep integrations with a bank's internal systems, compliance certifications - the company lists SOC 2, ISO 27001, GDPR and EU AI Act alignment - and domain-specific tuning that a general-purpose model does not have out of the box.
That is also why the competitive landscape is more crowded than a single startup story suggests. OpenAI has assembled more than 100 former investment bankers to help train its systems on building financial models for transactions including restructurings and initial public offerings - a direct attempt to move down the stack and own the same workflows. The frontier labs have the models; the vertical specialists have the workflows, the trust, and the integrations. The question for banks is whether they want their core analytical engine supplied by a horizontal AI giant or by a specialist whose entire business depends on finance.
There is a third actor in this chain: the banks themselves. Several large institutions are building their own internal AI tooling, and some are strategic investors in the very startups that could displace their juniors. J.P. Morgan participated in Rogo's Series C, and its growth-equity arm joined the Series D. That creates an unusual alignment: the bank is simultaneously a customer, a distribution partner, and a potential competitor to its vendors. It also means the banks are hedging - buying equity in the likely winners while building in-house options.
"Rogo has built an AI platform that the most demanding institutions in finance trust with their most critical workflows," Hamid said. "Their combination of technical depth, proprietary data integrations, and genuine domain expertise is why Rogo is pulling away from the field."
The Singapore base is a small piece of this larger positioning game. APAC is the growth market for private markets and wealth, and the firms Rogo sells to - investment banks, private-equity houses, asset managers - are all expanding their Asia presence. Establishing a local office, hiring local client-facing and support staff, and running events with the banking and private-equity community is how a software vendor converts a U.S. success story into a regional default.
The Strongest Case Against the Structural Read
The bear case against the "end of the junior banker" thesis is not weak, and it deserves weight. First, the apprenticeship argument is substantive: judgment in finance is learned through repetition of the very tasks being automated. If juniors never build a comp set by hand, they may not learn to spot when the model is wrong. Banks that gut their analyst programs could, in this view, be optimizing away their own future partners.
Second, regulation and liability constrain how far automation can go. Financial institutions operate under strict rules about who is responsible for the advice and analysis they produce. An AI-generated model that feeds a client pitch or a diligence memo still needs a human to stand behind it. That human has to understand the work well enough to defend it - which requires doing some version of the work. This is why many compliance officers treat AI as a drafting assistant rather than an autonomous decision-maker, and why adoption may be slower than the usage numbers suggest.
Third, the headcount data is not yet a clean confirmation. The EY finding that 60% of financial-services CEOs expect AI to maintain or increase headcount in 2026 cuts against the displacement narrative. So does the fact that Rogo's own growth is measured in users and queries, not in documented reductions of client payrolls. Usage can rise because the tool makes existing bankers more productive - which is consistent with stable or even rising employment.
These points are strongest in the short run. In the medium term, however, they describe friction, not a reversal. Apprenticeship can be redesigned - firms can create structured training that uses AI output as the teaching material rather than banning the tool. Compliance regimes adapt; the EU AI Act and sector rules are being written now, and they will shape the market rather than close it. And headcount is a lagging indicator: firms typically absorb productivity gains quietly before they cut hiring visibly, especially in a relationship business where over-hiring in a recovery is costly.
The falsifying signal is specific. If, through 2027 and 2028, bulge-bracket banks and elite boutiques report rising junior-analyst headcount and larger analyst classes while Rogo's own usage growth flattens - user counts and daily queries plateauing for multiple quarters - then the structural thesis is wrong and the displacement was cyclical cost-cutting after all. Conversely, if analyst classes stay compressed even as deal volumes recover, and Rogo's query volume keeps compounding, the regime-shift read is confirmed.
What to Watch: Three Time Horizons
Short term (6-12 months): Expect more geographic announcements like Singapore, and more hiring on the software side - solutions architects, support, enterprise sales - rather than finance analyst hiring. Watch Rogo's own metrics: the company currently reports more than 50,000 users, over 150,000 daily queries, and more than 350 institutions. Quarter-over-quarter movement in those numbers is the cleanest real-time gauge of adoption.
Medium term (1-3 years): The key signal is bank hiring. If analyst class sizes remain at the reduced levels Patnaik described - cuts of up to two-thirds at some firms - while revenue per banker rises, the productivity story has landed. Also watch whether banks begin to publicly attribute margin improvement to AI tooling in earnings commentary, and whether fee pressure emerges as firms compete on the lower cost base.
Long term (3+ years): The structural outcome is the "AI-first" institution Stengel describes - a firm whose workflows are built around agentic systems from day one, with a flatter pyramid and a smaller, more senior-skewed workforce. The winners in that world are the platforms that became the operating system: the ones embedded in the firm's data and processes, not the ones that sold a chatbot on top. Consolidation is likely; a handful of vertical platforms will capture most of the value, and general-purpose model providers will be suppliers rather than owners of the workflow.
The base case is that both forces operate at once: banks keep more humans than the most alarmist forecasts imply, because relationships and judgment remain human, but the base of the pyramid is permanently smaller and the work that remains is more analytical and less mechanical. The upside case for the platforms is faster-than-expected agentic adoption, where systems like Felix move from drafting to executing multi-step processes - deal screening, CIM generation, buyer outreach, data-room diligence - with minimal human intervention. The downside case is a compliance or accuracy shock: a high-profile error in AI-generated analysis that forces banks to pull back, re-impose manual controls, and slow the rollout.
For Singapore specifically, the question is whether the city-state becomes the regional showcase for AI-native finance - a place where the new model of banking work is adopted fastest, precisely because it is building new franchises rather than retrofitting old ones. That would make Rogo's first APAC office more than a real-estate decision. It would make it a marker of where the industry's center of gravity is moving: not just east, but into a different way of organizing the work itself.
The central judgment: Rogo's move into Singapore is not primarily a story about a startup opening an office. It is a story about the financial labor pyramid being rewritten in real time, with the base tasks that once justified armies of junior bankers being absorbed into software - and with the value of that work migrating from payrolls to platform equity. The apprenticeship model will survive in some redesigned form, and headcount will not vanish overnight. But the direction of travel is set, and the firms that treat AI as a cost line to manage rather than an operating model to rebuild are the ones most likely to find themselves on the wrong side of the shift.
Explore more exclusive insights at nextfin.ai.
