Session Report — SaaSpocalypse Axiom Engine

The Sorting

A structural analysis of publicly traded software companies, scored against seven axioms of durability in a world where AI has collapsed the cost of building software to near-zero.

August 3, 2026 7 Axioms
I — Thesis

When software was expensive — thin, horizontal, best-of-breed software stacks extracted rents across every business. Now that software is cheap — value moves to vertically integrated businesses that deliver opinionated end-to-end experiences.

— Naval Ravikant, July 2026

The SaaSpocalypse narrative says AI kills SaaS. The axiom engine says that's half right.

AI kills commodity SaaS — thin horizontal layers extracting rent from features any competent founder can now rebuild in a single Claude session. But it strengthens specialized infrastructure — payments, identity, compliance engines, financial exchanges — because the integration friction that was their bottleneck has evaporated. Every new vibe-coded app is a new customer for Stripe, Cloudflare, and Twilio.

The SaaSpocalypse is not an extinction event. It is a sorting event. This report applies the sort to the entire public software universe.

II — Methodology

The Seven Axioms

Each company is scored 0 or 1 on seven binary axioms of structural durability. A score of 1 indicates the company possesses a moat that AI cannot dissolve. The axiom total (0–7) determines the verdict.

1

Regulatory Gate

Does the product require certifications, licenses, or regulatory approvals that take more than 12 months to obtain?

2

Counterparty Network

Does the product's value derive from pre-negotiated bilateral relationships with external counterparties?

3

Liability Transfer

Does the customer use the product to move legal or financial liability off their own books?

4

Data Gravity

Does the product accumulate proprietary data that compounds in value and cannot be replicated?

5

Patent Protection

Is the core technology protected by patents that would block a clean-room rebuild?

6

The Session Test

Can a competent engineer with Claude rebuild 80% of the product in a single session? (Inverted: 1 = NOT rebuildable)

7

Vibe Coder Adoption

Are AI-augmented developers actively using this company's API as infrastructure in their own projects?

III — Results

Distribution

IV — Complete Universe

All Scored Companies

Ticker ▲ Company ▲ Verdict ▲ Score ▲ Reg ▲ Net ▲ Liab ▲ Data ▲ Pat ▲ Sess ▲ Vibe ▲ Mkt Cap ▲ P/S ▲ 52W DD ▲ Price ▲ 1M% ▲ 3M% ▲ 6M% ▲
V — Recommendations

Highest Conviction Signals

Structurally durable companies punished by the market (long targets) and structurally doomed companies the market has not yet repriced (short targets). Filtered to companies with >$200M market cap.

Loading…
VI — Extended Methodology: Axiom Scoring Prompts

How the Binary Decisions Were Made

Each company was scored by Claude Sonnet 4.6 via the Anthropic API, invoked through a Supabase Edge Function. The model received a system prompt defining the seven axioms with precise scoring criteria, followed by a user prompt containing the company's ticker, name, SIC classification, and business description extracted from SEC EDGAR filings.

The system prompt was engineered to prevent false positives — each axiom includes explicit guidance on what does not qualify. The full system prompt, reproduced verbatim below, is the analytical engine that produced every score in this report.

System Prompt — Axiom Scorer
You are an expert software industry analyst scoring public software companies against the "SaaSpocalypse Axioms" — a framework for determining whether a software company's core product is structurally durable or structurally vulnerable in a world where AI has collapsed the cost of building software to near-zero. The thesis: AI kills commodity SaaS (thin horizontal layers extracting rent from features any competent founder can now rebuild in a session). But AI strengthens specialized infrastructure providers (payments, identity, compliance) because the integration friction that was their bottleneck has evaporated. The SaaSpocalypse is not an extinction event — it is a sorting event. Score each of the 7 axioms as 0 or 1 with 2-3 sentences of reasoning each.
Axiom 1 — Regulatory Gate
Score 1 only if the product itself requires regulatory approval to operate (PCI-DSS, money transmitter licenses, FedRAMP, FDA clearance, SOC 2 Type II as contractual requirement). The VENDOR must need regulatory blessing, not just the customer. Generic software that happens to serve regulated industries scores 0 — the question is whether the VENDOR needs regulatory blessing, not the customer.
Axiom 2 — Counterparty Network
Score 1 only if the product's core value requires pre-negotiated bilateral agreements with external counterparties that take months/years to establish. Stripe's bank and card network partnerships. Twilio's carrier agreements. Plaid's bank-by-bank integrations. A two-sided marketplace where users self-serve onto the platform does NOT qualify — the question is whether the VENDOR had to negotiate institutional partnerships.
Axiom 3 — Liability Transfer
Score 1 only if customers specifically choose this product to move legal or financial liability off their books onto the vendor. Stripe assumes PCI liability. Cloudflare assumes DDoS mitigation liability. A cybersecurity vendor whose product reduces risk but does not contractually assume it scores 0.
Axiom 4 — Data Gravity
Score 1 only if the product accumulates proprietary data that (a) grows more valuable over time, (b) creates switching costs independent of the software, and (c) could not be replicated by a well-funded new entrant within 2 years. A CRM where the customer owns their own data and can export it does NOT — even if migration is painful, the data is the customer's, not the vendor's.
Axiom 5 — Patent Protection
Score 1 only if the company holds patents on the core technology within its most successful product that would legally block a competitor from building equivalent functionality. Must be patents on the technology itself, not design patents, business method patents, or defensive patent portfolios that exist but don't cover the core product mechanism.
Axiom 6 — The Session Test (Inverted)
Score 1 if a competent engineer with Claude CANNOT rebuild 80% of the core value proposition in a single working session (approximately 3 hours). Score 0 if they CAN. Consider: does the value come from the code itself (rebuildable) or from what surrounds the code — data, relationships, regulatory status, network effects? A project management tool is rebuildable. A real-time fraud detection engine trained on billions of transactions is not.
Axiom 7 — Vibe Coder Adoption
Score 1 if there is evidence that AI-augmented developers are actively using this company's API/SDK as infrastructure in their own projects — appearing in Claude Code sessions, vibe-coding tutorials, indie hacker stacks, YC company tech stacks. This signals that the company is becoming embedded in the AI-native developer workflow, which means cheaper software actually drives MORE usage of their product. Score 0 if the product is primarily used through its own UI by end users rather than called as infrastructure via API.
User Prompt Template — Per Company
Score this company: COMPANY: {company_name} ({ticker}) SIC DESCRIPTION: {sic_description} PRODUCT CATEGORY: {product_category} PRIMARY PRODUCT: {primary_product} BUSINESS DESCRIPTION: {business_description}
Response Schema
{ "axiom_1_regulatory_gate": 0 or 1, "axiom_2_counterparty_network": 0 or 1, "axiom_3_liability_transfer": 0 or 1, "axiom_4_data_gravity": 0 or 1, "axiom_5_patent_protection": 0 or 1, "axiom_6_session_test": 0 or 1, "axiom_7_vibe_coder_adoption": 0 or 1, "axiom_total": sum of all axioms, "reasoning": { "axiom_1": "2-3 sentence reasoning", "axiom_2": "2-3 sentence reasoning", ... } }

The verdict classification follows directly from the axiom total: Strong Long (5–7), Long (3–4), Neutral (2), Short (1), Strong Short (0). Company data was sourced from SEC EDGAR (SIC codes, CIK, business descriptions), Yahoo Finance (market cap, revenue, 52-week high/low, price history), and the Claude API (axiom scoring). Total API cost for 582 companies: approximately $15.

Disclaimer. This is an analytical framework, not investment advice. Axiom scores are generated by an LLM (Claude Sonnet 4.6) based on public information and may contain errors. The framework does not account for valuation multiples, earnings quality, management execution, macroeconomic conditions, or any of the other factors that determine whether a stock is a sound investment. Use as one input among many. Do your own diligence. Past performance is not indicative of future results.