Using public AI models leaks sensitive corporate data, as prompts and agent traces are sent to model providers. To protect proprietary information and maintain control, enterprises may revert to costly but secure on-premise infrastructure, reversing a 20-year trend of cloud migration.
Related Insights
To meet strict enterprise security and governance requirements, Snowflake's strategy is to "bring AI to the data." Through partnerships with cloud and model providers, inference is run inside the Snowflake security boundary, preventing sensitive data from being moved.
Enterprise SaaS companies (the 'henhouse') should be cautious when partnering with foundation model providers (the 'fox'). While offering powerful features, these models have a core incentive to consume proprietary data for training, potentially compromising customer trust, data privacy, and the incumbent's long-term competitive moat.
Strict regulations prohibit sending sensitive data to external APIs, creating a compliance nightmare for cloud-based AI. Small, on-premise models solve this by keeping data within the enterprise boundary, eliminating third-party processor risks and simplifying audits for regulated industries like healthcare and finance.
Despite public hype around powerful consumer AI, many product managers in large companies are forbidden from using them. Strict IT constraints against uploading internal documents to external tools create a significant barrier, slowing adoption until secure, sandboxed enterprise solutions are implemented.
Recent security breaches (e.g., Gainsight/Drift on Salesforce) signal a shift. As AI agents access more data, incumbents can leverage security concerns to block third-party apps and promote their own integrated solutions, effectively using security as a competitive weapon.
The "agentic revolution" will be powered by small, specialized models. Businesses and public sector agencies don't need a cloud-based AI that can do 1,000 tasks; they need an on-premise model fine-tuned for 10-20 specific use cases, driven by cost, privacy, and control requirements.
The high cost and data privacy concerns of cloud-based AI APIs are driving a return to on-premise hardware. A single powerful machine like a Mac Studio can run multiple local AI models, offering a faster ROI and greater data control than relying on third-party services.
Companies are becoming wary of feeding their unique data and customer queries into third-party LLMs like ChatGPT. The fear is that this trains a potential future competitor. The trend will shift towards running private, open-source models on their own cloud instances to maintain a competitive moat and ensure data privacy.
When companies don't provide sanctioned AI tools, employees turn to unsecured public versions like ChatGPT. This exposes proprietary data like sales playbooks, creating a significant security vulnerability and expanding the company's digital "attack surface."
The primary driver for running AI models on local hardware isn't cost savings or privacy, but maintaining control over your proprietary data and models. This avoids vendor lock-in and prevents a third-party company from owning your organization's 'brain'.
