New Wine Meets Old Wineskins

Executive Summary

The 2025 Nobel Prize in Economic Sciences provides SaaS founders with both a warning and a roadmap. The mathematical models of Aghion and Howitt — demonstrating that innovation raises quality from A to γA (where γ > 1) and that incumbent value is discounted by displacement risk λz — are now unfolding in real time through dramatic case studies.

Creative destruction is no longer a theoretical abstraction; it is the defining and accelerating force of software capitalism. Ancient wisdom and Nobel Prize–winning economic theory converge on the same conclusion: when AI is the new wine, it cannot be contained within the old wineskins of legacy SaaS thinking.

AI and SaaS: When New Wine Meets Old Wineskins

"Neither do people pour new wine into old wineskins. If they do, the skins will burst."

— Book of Matthew  ·  New Wine Meets Old Wineskins  ·  Thor ThunderScan Thesis 2026

Part I: The Mathematics of Creative Destruction in Software

The Aghion-Howitt Model Applied to SaaS

The 2025 Nobel Prize winners' work provides a precise mathematical framework for understanding what is happening in software markets right now. Unlike previous economic frameworks that modeled innovation as gradual, the Aghion-Howitt model captures the step-function nature of quality displacement — exactly what we are witnessing with AI-native entrants.

🏅

Sveriges Riksbank Prize in Economic Sciences in Memory of Alfred Nobel 2025 Nobel Prize in Economic Sciences

Joel Mokyr

½ Prize Share

Northwestern University, Evanston, IL, USA

"for having identified the prerequisites for sustained growth through technological progress"

Philippe Aghion

¼ Prize Share

Collège de France · INSEAD · London School of Economics

"for the theory of sustained growth through creative destruction"

Peter Howitt

¼ Prize Share

Brown University, Providence, RI, USA

"for the theory of sustained growth through creative destruction"

📎 nobelprize.org/prizes/economic-sciences/2025 · Awarded October 2025 · Ceremony December 10 2025, Stockholm

The Innovation Ladder: When an AI-native product enters with quality γA, it does not merely improve on the incumbent — it makes the incumbent's entire value proposition obsolete at that rung of the ladder. In software, γ values of 3–10x are already in production: order-of-magnitude improvements in task throughput, accuracy, and cost per outcome.

Knowledge Spillovers and the Asymmetric Advantage

The Aghion-Howitt model also captures knowledge spillovers — the mechanism by which prior innovations reduce the cost of future ones. In the AI era, these spillovers are catastrophically asymmetric:

• AI-native entrants benefit immediately from foundation model research (GPT, Claude, Gemini), open-weight models, and rapidly commoditizing AI infrastructure — they stand on the shoulders of $100B+ in R&D they did not fund.
• Legacy SaaS incumbents must retrofit decades of technical debt: monolithic architectures, rigid data schemas, synchronous workflow logic, and seat-based pricing models that are structurally incompatible with agentic computing.

This asymmetry creates a widening compounding gap. Each quarter an incumbent does not restructure around AI, the entrant climbs another rung of the quality ladder using cheaper, better tooling. The window for incumbents to act is not infinite — it is being actively consumed.

Part II: Critical Case Studies

Part III: The Survival Data & Market Reaction

The SaaS-Pocalypse by Numbers

The market data is unambiguous. This is not a cyclical correction — it is creative destruction pricing itself into public and private valuations simultaneously:

The Creative Destruction Timeline

The decisive observation from MIT's research on enterprise GenAI failure: companies avoid the necessary friction of transformation. The friction — restructuring data architecture, rebuilding pricing models, retraining GTM teams — is exactly where value is created or destroyed. Avoiding friction means ceding the quality ladder to those willing to bear it.

Part IV: Strategic Framework for SaaS Founders

The Three Paths: Reinvent, Enhance, or Exit

Based on the research, every SaaS founder faces one of three strategic realities. The choice is not optional — market forces will make it for you if you don't make it deliberately.

The critical diagnostic question: Do you have proprietary data that gets better as AI uses it? If yes, Reinvent or Enhance. If no — if your product is primarily workflow software without a compounding data flywheel — the λz discount will accelerate until it reaches zero.

The Vertical SaaS AI Playbook

The Horizontal SaaS AI Playbook

For horizontal SaaS companies, the transition requires a different emphasis — moving from systems of record to systems of action:

• Move from Storage to Work: Your product should not merely store data — it should perform work using that data. The shift from Salesforce-as-CRM to Salesforce-as-Agentforce is the horizontal model.
• Redesign Pricing for AI: Move from seat-based to outcome-based or usage-based pricing for AI features. Implement hybrid models: legacy subscription for the core product, consumption pricing for AI work performed.
• Invest in Telemetry: Build robust customer-facing usage tracking and spend prediction tools. AI cost unpredictability is the #1 enterprise objection — solve this with transparency tooling.

The Late-Stage SaaS Transformation Playbook (The Intercom Model)

For mature, plateauing SaaS businesses with the Intercom profile — slowing growth, comfortable but vulnerable margins, legacy architecture — the playbook is more aggressive:

Part V: The Venture Capital Reality

The capital allocation data for 2025 is the clearest signal of where value is being created and destroyed:

What institutional investors are actually underwriting — the four signals that drive AI premium valuations:

• Data depth and infrastructure: Clean, connected, proprietary data that compounds with AI usage. The cleaner your data, the higher your γ multiplier.
• System-wide intelligence: Not one-off tools or point features, but AI woven into the core workflow loop. The distinction investors make: "Is AI table stakes or is it the whole table?"
• Speed of execution: AI adoption cycles measured in quarters, not years. Boards are watching how fast the CEO can ship AI product iterations.
• Measurable outcomes: Improvements in GRR, NRR, profit margins, and time-to-value that can be directly attributed to AI. Story without metrics is marketing — metrics without story is a dashboard.

Part VI: Actionable Recommendations for SaaS Founders

Next 6 Months — Build the Foundation

6–24 Month Vision — Own a Rung

• Build the investor narrative now: Develop a metrics-backed story about how AI drives your next phase of value. Investors who haven't heard it will price in maximum λz uncertainty — don't let silence set your valuation.
• Develop proprietary models: If you hold specialized domain data — healthcare records, legal contracts, financial transactions — start fine-tuning. Your data is an asset foundation models cannot replicate; proprietary weights compound that advantage.
• Pick your layer in the stack: Position as Systems of Record (base layer), Agent OS (orchestration tier), or Outcome Interfaces (top layer). The highest-value position is Agent OS — owning the layer between raw AI capability and customer outcomes. The endgame for Vertical SaaS is not AI-enhanced software; it is becoming the intelligent infrastructure an entire industry runs on.

Final Thoughts: The Essential Fact About Software Capitalism

Four irreducible truths emerge from this synthesis. Creative destruction is the essential fact about capitalism — new technologies displace old ones, and the process is often sudden and painful in ways that comfortable incumbents systematically underestimate.

The deep research reveals four irreducible truths:

1. The pace of creative destruction is accelerating. AI is compressing innovation cycles from years to quarters. The γ multiplier is not 1.1 or 1.5 — it is 3–10x, and it is already here in production deployments.
2. Value is shifting from access to outcomes. The AIaaS business model charges for work performed, not software accessed. Every pricing model built on seats is a ticking clock.
3. Incumbents can win if they move fast enough. The Aghion-Howitt model shows that incumbents can maintain ladder position if they innovate faster than entrants. SaaS companies have real advantages in data, distribution, and enterprise trust — but only if activated decisively and immediately.
4. The feedback loop between knowledge and practice is essential. Mokyr's insight — that sustained growth requires a self-generating process of innovation — applies directly. The companies that will lead the AIaaS era are those that make learning-from-AI-deployment a core organizational capability, not a one-time initiative.

The case studies of Intercom, Salesforce Agentforce, and Vertical SaaS evolution demonstrate that the transition is possible — but it requires existential commitment, brutal speed, and the willingness to reinvent everything from pricing to organizational structure. This is not merely a technology upgrade. It is a refounding of the company, a re-architecting of the product, and a redesign of the business model.

The companies that treat this as an incremental feature addition will be the case studies in the next Nobel Prize lecture about creative destruction. The companies that treat it as an existential transformation will write the next chapter of software capitalism.