The 0.5% SaaS Trap: AI’s “Pixar Moment”
In the early 1980s, Pixar wasn’t a film studio. It was a struggling hardware company born inside Lucasfilm and backed by Steve Jobs. They built the Pixar Image Computer — a sleek, powerful machine designed for medical imaging and scientific visualization.
Nobody wanted it. No hospitals, no research labs, not even George Lucas. As Pixar bled cash, they had one odd advantage: a tiny in-house team making short animated clips just to demonstrate what the hardware could do. The hardware flopped, but the clips blew people’s minds.
So, Pixar made a radical creative leap: they stopped selling the silicon and started selling the story. They stopped optimizing a tool and reimagined its ultimate purpose. The result was Toy Story, and it rewrote the economics of Hollywood.
Today, AI founders are facing their own “Pixar moment.” Too many are obsessing over the underlying technology, the tool itself, while missing the grander narrative of the market: delivering undeniable, real-world outcomes.
This fixation has led to what can be called the 0.5% SaaS trap: founders are optimizing for the easiest layer of AI adoption while completely missing the massive 99.5% opportunity waiting in the real economy.
The Canary in the SaaS Coal Mine
Goldman Sachs recently published a sobering metric for the tech ecosystem: software services (SaaS) represent less than 0.5% of global GDP. The remaining 99.5% belongs to the physical world — manufacturing, energy, logistics, healthcare, construction, and robotics.
For the past year, we’ve watched software companies face a massive reality check. The ten largest holdings in a major software ETF collectively shed nearly $800 billion in market cap. This isn’t because software is dying; it’s because AI is breaking the traditional SaaS business model.
The old assumptions — that pricing should be based on user seats, that more users equal more value, or that adding a basic wrapper creates a defensive moat — are collapsing. As autonomous AI agents begin to do the actual work, value is shifting away from software access and moving directly toward trusted, physical outcomes.
The capital is already moving. Look at Prometheus, the AI engineering startup backed by Jeff Bezos that recently raised $12 billion at a $41 billion valuation. They aren’t building a better chatbot to write marketing emails; they are building an AI engineer to design physical jet engines and complex industrial systems.
The real battleground is the remaining 99.5% of the economy. But conquering it requires playing by an entirely different set of rules.
The 5 Laws of Industrial AI
In software, an AI hallucination wastes tokens. A confident but wrong AI answer isn’t just a product bug. In industrial settings, it can quickly turn into a trust and liability problem. Goldman Sachs highlights five capabilities that may separate industrial-AI leaders from commoditized competitors:
- Physics-Based Architecture: AI must understand materials, motion, temperature and mechanical constraints — not just language.
- Proprietary Operational Data: The strongest moats will come from private deployment data, including failures, edge cases and real-world operating patterns.
- Edge Deployment: Critical systems cannot depend entirely on cloud connectivity. Intelligence must often run locally for speed, reliability and safety.
- Certifiability: In regulated environments, technical performance is insufficient. The system must be demonstrably predictable, controllable and safe.
- Workflow Integration: Customers will not rebuild their operations around a new tool. The highest-value AI becomes a seamless capability layer inside existing workflows.
The Hidden Risk: Is Your Scenario Mature?
Goldman Sachs maps the technical requirements for the real economy. But for founders and investors, technical readiness is only half the battle. The greater risk is commercial readiness.
This is where the Scenario Maturity Assessment Framework (SMAF) becomes essential.
A technology can perform brilliantly and still fail to scale because the commercial environment around it is not ready. When industrial AI stalls, the problem is often not the product itself, but the scenario in which it is being applied. Three questions expose the gap:
- The Buyer & Budget: Are you selling to a specific operational buyer with a P&L, or pitching “abstract efficiency” that no single department actually owns?
- The Integration Friction: Can the product fit into existing operations, or must the customer redesign its workflow before receiving value?
- The Value Capture: Are you selling software access to an industrial buyer who only cares about measurable operational and financial outcomes?
Remember Pixar? Their Image Computer was technically brilliant. But the scenario was immature. They were applying a world-class capability to the struggling hardware scenario — until they pivoted to a more valuable outcome: computer-generated storytelling.
The next generation of AI winners will not necessarily have the most sophisticated models. They will be the companies that translate AI into outcomes that are trusted, workflow-integrated and commercially measurable.
When the technology works but adoption and revenue stall, do not immediately assume the product has failed. Audit the scenario.
At Mans International, we help deep-tech founders and investors identify where commercialization is breaking down—whether the answer is stronger workflow integration, a better value-capture model or a more mature market application.