---
id: "contrarian-off-the-shelf-over-proprietary"
type: "contrarian-insight"
source_timestamps: ["§ Training for a Marathon When You Can Only See Two Feet Ahead"]
tags: ["contrarian-insight", "llm-strategy", "build-vs-buy"]
related: ["claim-proprietary-models-not-competitive-advantage", "concept-ai-orchestration-layer"]
challenges: "The belief that data-rich legacy enterprises must build proprietary foundation models to maintain a competitive moat in the AI era."
sources: ["execution"]
sourceVaultSlug: "hbr-seg-execution"
originDay: 8
articleStem: "hbr-cl-93-legacy-financial-all-in-genai"
sourceUrl: "https://hbr.org/2025/03/how-a-legacy-financial-institution-went-all-in-on-gen-ai"
sourceTitle: "How a Legacy Financial Institution Went All In on Gen AI"
---
# Commercial LLMs Outvalue Proprietary Foundation Models

## Contrarian Insight: Commercial LLMs Outvalue Proprietary Foundation Models

**Challenges:** the belief that data-rich legacy enterprises must build proprietary foundation models to maintain a competitive moat in the AI era.

Many large enterprises with massive proprietary datasets assume their competitive advantage lies in **training their own foundation models**. [[entity-moodys|Moody's]] explicitly rejected this, deciding that models are merely *'ready-to-use'* commodity tools (see [[quote-ready-to-use-tools]]). The true advantage lies in the **speed of applying** these off-the-shelf models to proprietary data via an [[concept-ai-orchestration-layer|orchestration layer]].

### Connections
- The formal claim (attributed to [[entity-steve-tulenko|Steve Tulenko]]): [[claim-proprietary-models-not-competitive-advantage]].
- The architecture that enables it: [[concept-ai-orchestration-layer]].
- Requires deep domain knowledge: [[prereq-domain-expertise]].

### Steelman / counter-perspective (from enrichment)
Building on top of commercial models may **increase vendor dependence**. The orchestration layer reduces some lock-in, but it does not eliminate exposure to cloud/provider pricing, model-policy changes, or access constraints — the vault's own [[question-long-term-vendor-lock-in|open question on autonomy]] is a real strategic critique. Adjacent literature: this fits the broader **build-vs-buy** view that domain data, workflow integration, and distribution matter more than model ownership for most enterprises.


## Related across articles
- [[claim-public-llms-low-value]]
- [[action-use-proprietary-slms]]
- [[concept-ai-orchestration-layer]]