The Californer - Federal AI-Privilege Ruling Puts Confidential Work on Notice; eRacks Points to the On-Premise Answer

Trending...

First Partner Siebel Newsom, Governor Newsom, and #HalfTheStory launch California's first Teen Tech Council - 104
synseer Nominated for Best Startup in 2026 Prix Galien USA Awards Following Record Breaking Applicant Pool - 103
Phinge Exposes Massive AI Security Risks, Claiming Its Patented Hardware-Verified Architecture Is The Only Safeguard Against Surveillance Capitalism

In US v. Heppner, a federal court held that material a defendant prepared with a consumer AI platform was not protected by privilege or the work-product doctrine. The architecture lesson is direct, and eRacks builds it: private AI servers where prom

CAMPBELL, Calif. - Californer -- eRacks Open Source Systems, a builder of open-source rackmount servers since 1999, today highlighted the architecture implications of United States v. Heppner (No. 25-cr-00503-JSR, Southern District of New York), the February 10, 2026 ruling in which Judge Jed Rakoff held that roughly thirty-one documents a defendant had prepared using the consumer version of a generative AI platform were not protected by attorney-client privilege or the work-product doctrine.

The court's reasoning applies well beyond one case: information entered into a public AI platform is disclosure to a third party, and disclosure is how confidentiality protections are waived. In a separate matter, a federal court ordered roughly 20 million ChatGPT conversation logs produced in the consolidated copyright litigation against OpenAI, over its privacy objections.

More on The Californer

The Heppner holding was expressly tied to its facts: a public, non-enterprise AI platform, used without direction of counsel. Legal commentators analyzing the ruling have drawn the practical implication that tools which contractually or architecturally guarantee confidentiality can support a different analysis. On-premise AI is the strongest form of that guarantee, because no third party ever holds the data at all.

"The court drew the line exactly where the architecture draws it," said Joseph Wolff, Founder and CTO of eRacks Systems. "On a rented AI service, your prompts are records on someone else's server, under someone else's retention policy. On a private AI server, the model runs inside your walls and the only logs are yours. That is not a feature toggle. It is a different machine."

eRacks private AI servers ship with the full open-source AI stack pre-installed free, running current open models including DeepSeek, Llama, Qwen, and Mistral. Systems are configured air-gapped or egress-controlled, so data physically cannot leave the premises. The private AI line starts with the 2U AILSA at $7,695, and every configured price on eracks.com is a real, orderable number. The company's AI Provisioning and Setup service is $1,795 flat for end-to-end deployment, including model selection, tuning on the customer's hardware, and 30 days of support.

More on The Californer

The company notes that it builds architecture, not legal opinions: how the Heppner analysis applies to a given practice is a question for counsel. A detailed breakdown for legal practices is at https://eracks.com/law-firm-ai-server/ and the full private AI line is at https://eracks.com/products/ai-rackmount-servers/