The Trillion-Dollar Crime Scene: Where Value Really Vanished Between Giants and Startups

The Hook: Blood on the Boardroom Carpet

Tuesday, 08:17 a.m., quarterly board review.

On the screen: a beautiful strategy slide about “synergies between incumbents and startups in financial services, retail, health, mobility, education and media.” On the table: a term sheet for a late‑stage startup that burns more cash in a quarter than a Tier‑2 bank branch network in a year.

Then someone asks the only question that matters:

“If startups are so superior… why do more than 80% of them still lose money, while the so‑called ‘dinosaurs’ keep printing cash?”

Silence.

I’m the one in the room paid to be uncomfortable. I don’t care about narratives, I care about multi‑billion shifts. So I treat this like what it is: a crime scene. Massive value has gone missing. The suspects: traditional industry and the startup ecosystem, across finance, retail, health, mobility, education, entertainment.

The pitch decks say incumbents are obsolete. The quarterly reports say otherwise. Latin America’s startup ecosystem alone shrank 45% in 2025, while “shadow banking” quietly grew to 51% of global financial assets and pushed total financial assets to a record $503.7 trillion. Banks are “dead,” but the money went somewhere.

This report is the forensic audit of that disappearance—and the trillion‑dollar bet hiding in the blind spots.

The Genesis: How the Perfect Story Hid the Perfect Crime

During a decade of cheap money, we collectively bought a fairy tale:

• Incumbents: slow, bureaucratic, doomed.
• Startups: agile, data‑driven, cloud‑native winners.

We wrapped it in jargon: open banking, telemedicine, e‑commerce, microservices, embedded finance. The story was simple: traditional industries were “legacy”; startups were “the future.”

Reality is noisier:

• In Spain, only 18% of startups have positive EBITDA, and that figure has been flat since 2017.
• Latin America’s startup ecosystem lost 45% of its value in 2025, versus a 23% global drop.
• At the same time, shadow banking already moves more money than traditional banking (51% of global financial assets) and total financial assets have reached $503.7 trillion.
• In financial services, generative AI is already improving productivity by 20% in software development, customer service, and IT.

The giants didn’t die. They shed their skin while everyone was watching “disruption” conferences. Startups didn’t “take power”; they became an outsourced—and often fungible—R&D force.

The official story—“win‑win collaboration between corporates and startups”—sounds great on a panel. But the numbers don’t add up. Where did the value go?

The Invisible Conflict: The Missing Value We Don’t Want to See

There’s a conflict almost no one names: the business model that wins on the pitch is not the one that wins in the regulatory, tax, and balance‑sheet Excel.

Seen like a forensic auditor:

• Traditional industry: optimizes value capture—regulation, brand, physical distribution, long‑term contracts, balance‑sheet control.
• Startups: optimize value creation—technology, UX, data, experimentation.

Both sides believe the other will do the dirty work:

• The bank expects the fintech to build its “digital front end” without eroding margins or regulatory hegemony.
• The fintech expects the bank to grant cheap access to licenses, balance sheet, and customers.

The value disappears in the crack between those two illusions.

This conflict repeats, sector by sector.

Evidence & Insights: Sector by Sector, Cut Open on the Table

I’ll treat each industry as a different scene of the same crime. Same pattern, different instruments.

1. Financial Services: The Balance‑Sheet Machine vs. The UX Factory

Business model

• Incumbents (banks, insurers): interest income, fees, charges. Heavy structure: branches, regulatory compliance, capital. Margins protected by licenses and regulation.
• Fintech (Revolut, N26, Stripe, Fintonic, Crediverso): premium subscriptions, transaction fees, interchange, revenue‑sharing. Low marginal costs, but brutal dependence on volume and external funding.

While fintechs promise to “kill” banks, the reality is that shadow banking is the real predator: 51% of global financial assets outside traditional banking. The crime here is conceptual: the narrative equated disruption with apps, while real power shifted to balance‑sheet structures and investment vehicles.

Technology and architecture

• Banks: legacy cores, monoliths, long release cycles. But huge IT budgets: global banking IT spend hit around $652 billion in 2023.
• Fintech: cloud‑native, microservices, open APIs, data‑driven. Heavy integration with third parties via open banking, payment gateways, banking‑as‑a‑service.

Generative AI already gives ~20% productivity gains in development and support. That’s not a “nice to have,” it’s a massive multiplier combined with cheap capital and a stable customer base.

User experience and product

• Fintech: onboarding in minutes, polished mobile UX, almost transparent pricing, real‑time notifications.
• Banks: heavy onboarding, KYC/AML friction, fragmented channels.

Crediverso illustrates another part of the crime: the value is not only in the financial product, but in education and trusted intermediation for a specific audience (the Hispanic community in the US), with impartial comparisons and content in Spanish.

Organization, culture, governance

• Banks: hierarchies, reputational risk aversion, regulator dependence. High risk aversion, dense corporate governance.
• Fintech: aggressive growth, high tolerance for experimentation, but often immature governance.

Competitive dynamics and scenarios (5–10 years)

• BBVA and others push corporate venture capital, acquisitions, alliances.
• Stripe and peers rewrite payment infrastructure without needing a full banking license.
• Fintonic evolves from a friendly app into a scoring and financing player sharing risk with partners.

Where the trillion hides: in embedded finance. It’s not the banking app that wins, but the layers of credit, payments, and insurance embedded in retail, mobility, education, etc. The crime: many founders limited themselves to “building another neobank” while the big value flows shifted to cross‑sector platforms.

2. Retail / E‑commerce: The World’s Stomach vs. The Software Layer

Business model

• Traditional retail (Walmart, big chains): volume, turnover, inventory optimization. Tiny margins, huge scale. Stable cash flow.
• Startups / platforms (Shopify, niche marketplaces): SaaS subscriptions, take rates on sales, value‑added services (logistics, seller financing).

Walmart and peers are cash factories but prisoners of physical assets. Shopify owns no inventory but captures margin across thousands of micro‑merchants.

Technology and architecture

• Incumbents: on‑prem ERP and inventory systems, complex integrations, technical debt.
• Startups: cloud platforms, APIs, plug‑and‑play integrations with payments, shipping, marketing.

User experience and product

• Ecommerce startups obsess over friction: one‑click checkout, recommendations, real‑time tracking.
• Retailers struggle with omnichannel: physical stores, outdated websites, half‑abandoned apps.

Organization and culture

• Traditional retail: margins so thin that any risk is seen as a threat.
• Startups: incentives toward fast growth, constant testing, without the operational constraints of thousands of physical stores.

Dynamics and scenarios

• Amazon proved that logistics + data + subscription (Prime) is a nuclear weapon.
• Retailers respond with startup acquisitions, their own marketplaces, and logistics alliances.

Where the trillion hides: in platforms that combine ecommerce + financing + behavioral data. Embedded finance appears again: point‑of‑sale credit, insurance, subscriptions. The crime: many retailers still see themselves as “sellers of stuff,” not as owners of a monetizable behavior graph.

3. Health: The Regulatory Cathedral vs. The Digital Lab

Business model

• Hospitals / traditional systems: billing insurers, fee‑for‑service, agreements with public systems. High revenue per episode, dependence on physical capacity.
• Startups (Zocdoc, Teladoc and peers): intermediation fees, telemedicine subscriptions, B2B2C models with insurers and employers.

Technology and architecture

• Incumbents: legacy electronic health records, proprietary systems, poor integration.
• Startups: AI for assisted diagnosis, telemedicine, remote monitoring apps.

User experience

• Startups: book an appointment in minutes, video consult, reminders.
• Hospitals: phone calls, waiting rooms, duplicated processes.

Organization and governance

• Traditional health: heavy regulation, professional associations, rigid protocols.
• Startups: agile, but clash with regulatory and trust barriers.

Dynamics and scenarios

• Collaboration grows: hospitals integrate telemedicine; startups become “channel providers.”
• Teladoc and peers redefine what it means to “go to the doctor.”

Where the trillion hides: in systems that integrate clinical + financial + behavioral data for long‑term risk management. Legal friction slows this now, but the vector is unstoppable. The crime: overinvestment in “yet another appointment app” and very little in the deep back‑end of health risk and outcomes.

4. Mobility / Transport: Infrastructure That Cannot Fail vs. High‑Burn Toys

In mobility, the story got distorted: some companies perceived as “startups” (Uber, Lyft) are already incumbents in their own category.

Business model

• Traditional transport (fleet operators, public transport): public contracts, regulated fares, relatively predictable revenues.
• Mobility and micromobility platforms (Uber, Lyft, Lime, Bird): per‑use pricing, subscriptions, take rates on drivers or users. In micromobility, asset‑intensive (scooters, bikes) with high maintenance costs.

Technology and architecture

• Traditional: older fleet management systems, little real‑time optimization.
• Platforms: matching algorithms, dynamic pricing, mobile apps, maps.

Experience

• Uber and peers: slick UX—ETAs, invisible payments, in‑app support.
• Traditional transport: fixed schedules, physical tickets, low visibility.

Organization and governance

• Traditional operators: deeply tied to municipal regulation and unions.
• Platforms: regulatory battles, labor conflicts, constant reputational risk.

Dynamics and scenarios

• Many micromobility startups have struggled due to capital‑intensive models and weak unit economics.
• Cities are forcing integration: shared data, limited licenses.

Where the trillion hides: in orchestrating the full flow of people and goods + payments + insurance, not in launching another scooter app. The crime: confusing “transport app” with “mobility infrastructure.” Very few are building the transactional and risk layer at city‑ or country‑scale.

5. Education: The Credential System vs. The Attention Economy

Business model

• Universities and traditional centers: tuition, public subsidies, research, official degrees.
• Edtech: subscriptions, paid courses, content marketplaces, B2B deals with companies for continuous training.

Technology and architecture

• Traditional: legacy LMS, physical campuses, stale content.
• Startups: cloud platforms, micro‑learning, AI for personalized learning paths.

User experience

• Edtech optimizes flexibility: on‑demand, mobile, modular.
• Universities remain tied to semester calendars and in‑person requirements.

Organization and governance

• Formal education: curricular rigidity, accreditation, bureaucracy.
• Startups: fast iteration, but little institutional legitimacy.

Dynamics and scenarios

• Hybrid programs grow: universities using edtech platforms; companies ignoring degrees and betting on skill‑based hiring.

Where the trillion hides: in platforms that connect learning, assessment, and employability with longitudinal skills data. The crime: treating edtech as “yet another Netflix for courses,” ignoring the signaling layer for the labor market where real power lies.

6. Entertainment / Media: The IP Factory vs. The Attention Algorithm

Business model

• Traditional media: advertising, subscriptions, content licensing.
• Content startups and platforms: subscription, programmatic ads, revenue‑share with creators.

Technology and architecture

• Traditional: proprietary broadcast systems, custom CMS.
• Startups: cloud platforms, OTT distribution, recommendation algorithms.

User experience

• Platforms enable extreme personalization and cross‑device consumption.
• Linear media lose relevance among younger audiences.

Where the trillion hides: in ownership and orchestration of IP + behavioral data + direct monetization mechanisms (subscription, micropayments, fan economy). The crime: many media companies adopt modern distribution tech, but their business model remains trapped in 20th‑century mass‑audience logic.

The Cross‑Cutting Framework: The Crime Matrix

To see the pattern, here’s the forensic matrix I use internally.

Table 1 – The Cross‑Sector Evidence Sheet

The Strategic Shift: How to Rewrite the Case Before the Prosecutor Shows Up

If this were just description, it would be useless. The game is to capture the “missing value” before others do. Here’s the plan, stripped down.

For incumbents: stop being the crime scene and become the prosecutor

1. Build “balance‑sheet platforms,” not just pretty apps
   Use your licenses, capital, and compliance as offensive assets. In finance, that means creating banking‑as‑a‑service and embedded‑finance rails for third parties. In health, insurance, and mobility, it means packaging data and risk into API‑first products.

2. Industrialize the relationship with startups
   Stop doing marketing POCs. Set up:

   • Clear theses by sector and business‑model type.
   • Standard technical integration processes, with defined timelines and SLAs.
   • Commercial agreement templates that scale to real business, not eternal pilots.

3. Use generative AI where it hurts most, not where it looks coolest
   In finance, it already yields 20% productivity in development and support. Apply it to:

   • Legacy code modernization.
   • Back‑office and compliance automation.
   • Internal assistants for branch staff, call centers, doctors, agents.

4. Redesign internal incentives to accept controlled disruption
   Tie bonuses to:

   • Revenue generated via external/startup platforms.
   • Reduced integration times.
   • Percentage of new products launched in collaboration with third parties.

5. Master data as a sovereign asset
   The real crime would be handing it to external aggregators:

   • Design clear data‑sharing and data‑monetization policies.
   • Build your own customer graphs that can span sectors (e.g., retail + finance + mobility).

For startups: from pitch heroes to value surgeons

1. Pick a side: be a “critical layer” or die as a “feature”

   • Critical layer: financial infra (Stripe), scoring (Fintonic), unique channels (Crediverso for US Hispanics), logistics rails, health‑risk systems.
   • Nice‑to‑have feature: another onboarding UI for banks, another course marketplace, another appointment app without access to clinical data.

2. Design business models that survive without capital as a drug
   Remember the numbers: only 18% of Spanish startups are EBITDA‑positive; LatAm lost 45% of value. The market has shown what it does with models that only work with zero rates.

3. Understand where legacy truly hurts
   The crime isn’t that banks have ugly apps, but that:

   • Credit processes are slow.
   • Risk management ignores alternative data.
   • Systems block integrated products (e.g., insurance + credit + payments).

4. Negotiate from data, not from narrative
   Bring hard evidence:

   • Documented uplift in conversion, churn reduction, scoring improvement.
   • Measured impact on cost and risk for the incumbent.

5. Build governance before it’s imposed on you
   Especially in finance and health:

   • Internal risk and compliance committees.
   • Data policies aligned with regulators from day one.

The Winners vs. Losers Scorecard: Who’s Actually Armed for the Next Cycle

Table 2 – The Winners vs. Losers Scorecard (2026–2035)

The Big Picture: The Trillion Dollar Bet

If I had to compress this forensic audit into a single bet, it would be:

The next trillion won’t be created “vs. the giants,” but “between the giants,” in the infrastructure and data layers that connect industries.

This is not an epic David vs. Goliath story. It’s a cold game between:

• Those who control licenses, balance sheet, and regulated access.
• Those who control data, experience, and reconfigurable tech layers.

Previous narratives talked about disruption wars, tribes, strategy kitchens, future collapses. This angle is different: a crime‑scene report. The bodies are already there: bloated regional ecosystems that then contract 45%, whole cohorts of zombie neobanks and edtechs, inflated IT budgets that don’t move the needle on customer experience.

My job as a VC is not to save anyone, it’s to read the scene correctly:

• If you’re an incumbent and still believe the value lies in “protecting your core,” you’ll end up as a cheap license provider for someone smarter.
• If you’re a startup and still believe that “another pretty app” entitles you to a slice of those $503.7 trillion in global financial assets, you’ll end up as a footnote in a liquidation report.

The real trillion‑dollar bet is simple to state and brutal to execute:

• Bet on companies that operate at the intersection of sectors: fintech embedded in retail, health financed as a financial asset, education measured as human credit‑risk.
• Demand models that can show real EBITDA on a reasonable horizon, without the anesthesia of zero rates.
• Penalize any actor—incumbent or startup—without a clear data and platform governance plan.

In a crime, you always follow the money. This decade, you follow the data and the license too. Where those three meet, that’s where the next trillion will be. Everything else is conference noise.

References

1. “Empresa emergente”, definition and characteristics of startups, Wikipedia in Spanish.
2. “Suscripción” and “Freemium”, business models, Wikipedia in Spanish.
3. Fundación Persan, “Lean StartUp Canvas – Business Model”, introduction to business models and the business model canvas.
4. MBA Asturias, explanation of the Marketplace business model.
5. El País – Negocios, article on franchises as a business model.
6. Xataka, “La trampa del gigantismo: qué hace más innovadoras a las startups que a las grandes empresas.”
7. Cinco Días (El País), “Las startups mejoran su edad media pero se estancan en beneficios e ingresos”, 18% positive EBITDA figure for Spain.
8. Simalco, “El ecosistema de startups de América Latina se contrae 45% en 2025.”
9. Bain & Company, press note on 20% productivity improvement in financial services thanks to generative AI.
10. El País (branded content), “Las finanzas integradas revolucionan los servicios financieros.”
11. Cinco Días (El País), “Fintonic cierra su reestructuración y ficha como CEO al ex jefe de seguros del Santander.”
12. Cinco Días (El País), “La banca en la sombra mueve ya más dinero que la banca tradicional…”, shadow‑banking and global financial‑asset data.
13. AS USA, coverage of Crediverso as the first fully Spanish‑language financial platform in the US.
14. TI Inside, Gartner quote on global IT spending in banking and financial services ($652 billion in 2023).