Harvey, an early OpenAI investment, is now creating proprietary models on open-source foundations. This signals a major trend where vertical AI companies are protecting their valuable training data and avoiding dependency on platforms that could become competitors.

Enterprise customers are restricting use of frontier models from Anthropic and OpenAI due to data privacy concerns. Competitors like Microsoft and NVIDIA are exploiting these fears, promoting solutions like running models locally on a customer's own hardware as the only truly secure alternative.

Major customers of frontier AI labs, such as voice AI company Eleven Labs, are actively working on proprietary models. This trend of verticalized model development signals a desire to escape data leakage concerns and dependence on potential future competitors.

A growing number of companies, especially in regulated industries like finance and healthcare, are opting for open-source AI models they can run on-premise. This trend is driven by concerns over data leakage, IP security, and national data sovereignty, creating a distinct market need for more domestic, controllable AI solutions separate from frontier models.

Relying on third-party LLMs is a temporary phase. The ultimate advantage will come from companies training and owning their own models, potentially on physical hardware in their office. This transforms AI from a rented tool into a core, defensible intellectual property.

Frontier models from giants like OpenAI force enterprises to share sensitive data, creating platform risk. The future of corporate AI lies in private, fine-tuned, open-source models that keep a company's "intelligence" in-house, preventing it from training potential competitors.

A coming "emperor has no clothes moment" will see enterprises reject cloud-based AI over data privacy fears. Fed up with providers scraping and training on their sensitive data, companies will increasingly buy their own hardware (like NVIDIA's DGX Spark) to run AI models in a secure, ring-fenced environment.

Large enterprises are avoiding commitment to a single AI provider like OpenAI or Anthropic. Instead, they're building control planes and abstraction layers that allow them to hot-swap the underlying models, mitigating technology risk and preventing dependence on one provider's terms of service.

Regulatory uncertainty and delayed access to top-tier models from labs like OpenAI and Anthropic are pushing enterprises to adopt open-source alternatives like GLM 5.2. This shift allows companies to secure their own computing resources and train proprietary models, gaining data sovereignty and cost control.