Gabe Stengel - Building Investing Superintelligence

Anthropic

Across the shows: 4 speakers in other episodes, 4 bullish and 10 bearish calls, past 30 days

• Gabe Stengel · RogoBearish

  Anthropic won't pursue vertical finance AI niches like Rogo's; they'd be a penny-pick for a lab chasing $1T revenue.

  Stengel argues Anthropic is not a threat to vertical finance AI. Finance is a collection of messy niches (data sets, regulatory requirements, data rooms, systems of record) that the labs, which are aiming to grow from $100B to $1T in revenue, will treat as 'picking up a penny.' Rogo can reach roughly $5B in revenue by going deep on these workflows. The stance is bearish on the lab competing in this vertical, while Anthropic remains Rogo's model supplier.

  “That for Anthropic would kinda be like stopping on the side of the road to pick up a penny because they're on the pathway of trying to go from a 100,000,000,000 in revenue to a trillion in revenue.”at 18:32

• Gabe Stengel · RogoBullish

  Anthropic's Claude Code and Cowork beat ChatGPT through better harness and presentation, despite similar underlying models.

  Stengel credits Anthropic's Claude Code and Cowork with better harness and presentation than OpenAI's ChatGPT even though the underlying models were similar. The harness let the models exercise long-running capabilities, driving a run-up in usage and expansion. He uses this to argue the infrastructure around a model matters as much as the model.

  “The models were actually fairly similar but the harness and the way that it was presented from Claude was far better and it just allowed the models to exercise more of their long running capabilities and that's why they had a run up in usage and a huge amount of expansion.”at 20:10

• Gabe Stengel · RogoBullish

  Anthropic's token-consumption model let Rogo's spend grow exponentially with no human sales loop, showing usage-based AI scales cheaply.

  Stengel notes that Anthropic sold to Rogo with ease and that Rogo's spending on Anthropic has risen exponentially without a human sales loop, because of the token consumption model. This illustrates how usage-based AI revenue can scale rapidly with minimal go-to-market effort.

  “I look at how hard it was for Anthropic to sell to us very easy and the amount that we pay them has risen exponentially without a human loop because it's a token consumption model.”at 25:44

OpenAI

Across the shows: 6 speakers in other episodes, 3 bullish and 9 bearish calls, past 30 days

• Gabe Stengel · RogoBearish

  OpenAI and Anthropic won't build deep finance infrastructure like data rooms and transaction venues, leaving room for vertical AI.

  Stengel says frontier labs like OpenAI and Anthropic will not build the deep, perpendicular infrastructure finance requires, such as data rooms, compliance plumbing, transaction venues and systems of record. That leaves room for vertical AI companies, so he expects limited encroachment into this niche.

  “I don't think OpenAir and Thropic will ever wanna build a data room business. And if you actually wanna be the exchange for all of high finance and all of capital markets, you not just need to own the intelligence, you need to own the transaction venue”at 19:20

JPMorgan Chase & Co.

• Gabe Stengel · RogoBullish

  JPMorgan is using AI productivity to enter SMB M&A, enabled by single-banker deal teams, a segment once uneconomical.

  Stengel highlights JP Morgan's announced push into M&A for small and mid-sized businesses as an example of using AI productivity strategically. Deal fees in that segment were previously too small to justify large staffing, but a banker acting as a 'deal team of one' makes it viable.

  “JP Morgan just announced that they're gonna try and do a lot more m and a work for SMBs. For parts of the market that before they didn't think it made sense to go out and serve cause the deal fees were probably too small”at 33:27

Enterprise AI adoption & AI software

• Gabe Stengel · Rogo

  Whether AI can replace top investors' judgment is unresolved; proprietary data and expert relationships remain the human edge for now.

  Stengel says it's unresolved whether AI can replace the judgment of the best investors. If a 'move 37' moment happens in public equities, what matters would change dramatically. Meanwhile, durable human advantage lies in gathering proprietary data and expert relationships that feed models. Over a two-to-five-year horizon, top investors will be busy reinventing their firms around AI.

  “Jury's still out on whether or not that is something that AI can eventually replace. When move 37 happened... If that starts to happen in public equities, yeah it's gonna really change what matters.”at 13:13

• Gabe Stengel · Rogo

  Individual AI productivity gains in investment banking are large, but converting them into measurable firm-level gains is the open challenge.

  Bankers report huge individual productivity gains, for example MDs producing in ten minutes what took an analyst three days. The open challenge is converting individual productivity into measurable firm-level gains: winning more deals, cutting costs, or serving new market segments. He cites JP Morgan targeting SMB M&A that previously wasn't economical, enabled by a 'deal team of one.'

  “The majority of firms are seeing a huge amount of individual productivity and they're trying to figure out how do we parlay that individual productivity into firm productivity that we can measure.”at 32:34

• Gabe Stengel · Rogo

  As finance AI shifts from copilots to autopilots, auditability matters more than accuracy for trust, debugging and regulatory compliance.

  As tools move from copilots for information retrieval to autopilots that make investment decisions or send emails, auditability matters more than raw accuracy. Users need to see assumptions and data sources to trust and debug outputs. Regulators in capital markets will also require explanations for decisions, so unauditable systems cannot be deployed. Accuracy still matters, but a mostly accurate, easily checked system is actionable.

  “I think it's actually more important to be auditable than it is to be accurate.”at 34:41

Private equity

• Gabe Stengel · Rogo

  Private equity, hedge funds and banks face their first innovator's dilemma from AI-native rivals, and must encode talent into systems.

  After two decades of easy money for capital allocators, PE firms, hedge funds and banks face their first real innovator's dilemma. AI-native investment firms and banks can attack incumbent business models, forcing incumbents to reinvent everything. Firms that fail to put their best people's knowledge into owned software, data and systems risk being disrupted, since enterprise value is likely to shift from people to systems over a decade.

  “There has not been an innovator's dilemma for private equity firms, for hedge funds, for investment banks in a long time.”at 52:33

• Gabe Stengel · Rogo

  AI will standardize unstructured private-markets data, boosting liquidity and potentially fully automating sponsor-owned, secondaries and private credit transactions.

  Private markets have resisted standardization because of unstructured data, which limits transaction volume and liquidity. AI should fix that and make assets easier to track and trade. He thinks sponsor-owned businesses, secondaries, private credit and GP-LP secondaries could be fully automated. The timing is uncertain, depending on regulatory or market forces for standardization and on how fast small business owners get comfortable selling without a handshake.

  “Private markets have been immune to standardization because there's so much unstructured data. AI should fix that.”at 57:38

Banks

• Gabe Stengel · Rogo

  Top investment banks will hold 90% of enterprise value in software, data and systems rather than people within a decade.

  Banks face a shock where AI-native investment banks can attack their model, and they must rethink what they do. He expects the world's best banks to hold 90% of enterprise value in software, data and systems rather than people within ten years. Banks that don't encode their best bankers' knowledge into systems, or that rely on being merely better-read, will see that advantage obviated by AI.

  “In ten years, the world's best investment firms, best banks will have 90% of their enterprise value, not in people, but in software and data and systems.”at 54:59