Compliance
NVIDIA TensorRT-LLM in India
Structured issue-spotting for deploying NVIDIA TensorRT-LLM in India, read against the rules for that jurisdiction.
- Source
- Rules engine over a generic brief — no anchored quote on this page
- Verified
- Evidence not verified
- Confidence
- Low
Structured issue-spotting to support your own review — not legal advice. Verify against the cited primary sources and your counsel.
01What this reading assumes
- Jurisdiction
- India
- Principal framework
- The Digital Personal Data Protection Act, 2023 received assent in August 2023 and comes into force provision by provision, on dates the Central Government appoints. It is consent-based, with notice, security safeguards, breach notification, additional duties for Significant Data Fiduciaries, extraterritorial reach over processing connected with offering goods or services to people in India, and penalties up to 250 crore rupees. The DPDP Rules, 2025 supply the operative detail on a phased timetable: rules 1, 2 and 17 to 21 from publication in November 2025, the consent-manager rule one year later, and rules 3 and 5 to 16 — notice content, security safeguards, breach reporting, children’s verifiable consent, retention, Significant Data Fiduciary duties and cross-border transfer — eighteen months after publication. Until then the 2011 SPDI Rules and sectoral regimes remain the operative privacy law.
- Delivery assessed
- Self-hosted
- Data leaves the network
- no
- Vendor home jurisdiction
- United States
- Verified vendor positions
- none — every vendor position below is a question, not an assurance
- Rules evaluated
- 35
- Rules fired
- 10
Assumptions about use
- An internal deployment used by employees, not a public-facing product.
- A person reads the output before acting on it — but that is not recorded, so the engine reports it as a gap rather than assuming it.
- No significant automated decision is taken about a person by the system alone.
02Issues to work through
Cross Cutting
Self-hosting moves the security obligation to you
Keeping data on your own hardware answers the transfer question and creates an operations question. Patching, backups, key management, monitoring and incident response are now yours, and an unpatched inference server on the office network is a worse outcome than a well-run vendor.
- Required checks
- Name the person responsible for patching each component, and the cadence.
- Confirm backups exist, are encrypted, and have been restored at least once.
- Confirm there is an incident response path that includes this system.
- Technical controls
- Encrypt at rest and in transit, including between the application and the inference server.
- Centralise authentication through the existing identity provider rather than local accounts.
- Keep an audit log of who queried what, and protect it from the people it records.
- Subscribe to security advisories for each component and track upgrade lag.
Being able to reconstruct a decision months later
The question that arrives after a complaint is what the system was shown and what it produced on a particular day. Models change, prompts change, and indexes are rebuilt, so the answer has to be recorded at the time. Without it, the only available response is that the output cannot be reproduced.
- Required checks
- Decide what is recorded per interaction: model and version, prompt template version, retrieved document ids, output, reviewer and outcome.
- Set how long those records are kept, balanced against the retention duties that also apply to them.
- Vendor questions
- Does the vendor pin model versions, and how much notice is given before a model is retired or changed?
- Technical controls
- Version prompt templates in source control and log the version used.
- Log the model identifier and version returned by the provider, not the one you requested.
Confidentiality duties bind independently of data protection law
Material can be entirely free of personal data and still be the material a contract stops you disclosing. Client retainers, non-disclosure agreements, supplier contracts and common-law duties are the usual sources, and several of them require consent before a third party processes the material at all — which a model API call is.
- Required checks
- Review the confidentiality clauses in the contracts covering the material going into the system.
- Identify any contract requiring notice or consent before a subcontractor processes the material.
- Decide whether the deployment needs a confidentiality carve-out negotiated into new contracts.
- Vendor questions
- Will the vendor accept a confidentiality undertaking beyond its standard terms?
- Which staff at the vendor can access customer content, under what controls?
- Technical controls
- Segregate the most sensitive corpora into an index that the general assistant cannot reach.
Human review required — take this to your counsel
We were not told whether a person reviews the output
Where output influences a decision about a person, the reviewer has to be able to disagree with it. That needs three things a rubber-stamp review lacks: enough information to judge, enough time to judge, and an override that is used often enough to be real. Design it before the volume makes it impossible.
- Required checks
- Name the role that reviews the output and what they see when they do.
- Decide what evidence is retained about each review, so the practice can be shown to exist.
- Set a threshold below which the system must not act without review.
- Vendor questions
- Does the product expose the retrieved context and the confidence behind a suggestion, or only the answer?
- Technical controls
- Show the reviewer the retrieved sources next to the suggestion, not the suggestion alone.
- Record the reviewer’s decision, including overrides, as part of the audit trail.
Human review required — take this to your counsel
An AI deployment creates new copies of the data
Vector indexes, prompt logs, completion caches, evaluation datasets, fine-tuning checkpoints and backups are all copies of the source material in places the existing retention schedule does not mention. Deletion requests are the moment this is discovered, because deleting the source document does not delete its embedding.
- Required checks
- List every store the deployment creates and add each to the retention schedule.
- Establish how a deletion request propagates to the index, the caches and the logs.
- Establish how long backups keep material that has been deleted from the live system.
- Vendor questions
- What does the vendor retain, where, and for how long after we delete our copy?
- Technical controls
- Store the source document id with every embedding so deletion can cascade.
- Set time-to-live on prompt and completion logs rather than relying on manual cleanup.
Who can reach the model, the index and the weights
A self-hosted stack has three access surfaces that are easy to leave open: the inference endpoint, the vector index, and the weights on disk. Retrieval also carries an authorisation problem an ordinary application does not have — the index must not return a document to someone who could not open it in the source system.
- Required checks
- Confirm the inference endpoint is not reachable from outside the network and requires authentication.
- Confirm retrieval filters by the requesting user’s permissions, not only by relevance.
- Confirm who can read the model files and the index volume at the operating-system level.
- Technical controls
- Bind the inference server to a private interface and put an authenticating proxy in front of it.
- Carry document-level access control into the index and enforce it at query time.
- Encrypt the volume holding the weights and the index, and restrict it to the service account.
- Rotate API keys and keep them out of client-side code and container images.
What ends up in a prompt, and where it goes next
Even with inference inside the network, prompts and retrieved context accumulate in logs, traces and caches, and system prompts can often be extracted from the output. The leak path is internal rather than external, but it is still a copy of the source material in a new place.
- Required checks
- Write down which categories of information may be entered into a prompt, and tell people.
- Establish what the system prompt contains and whether disclosing it would matter.
- Establish which shadow tools staff are already using; the policy has to name the permitted ones.
- Vendor questions
- Are prompts and completions retained, for how long, and can retention be set to zero?
- Are prompts used for abuse monitoring, and if so who can read them and for how long?
- Technical controls
- Redact or block high-risk patterns before the prompt leaves the application.
- Keep prompt and completion logs out of general-purpose observability tools.
- Set an explicit retention period on prompt logs and enforce it.
India
The six-hour incident clock and in-India log residency apply today
India’s national cyber-incident directions require listed incidents to be reported within six hours of noticing them or being told of them, with no confirmation threshold, and require ICT logs to be enabled and kept for a rolling 180 days within Indian jurisdiction. Both have applied since 2022, which makes them the sharpest live constraint on an Indian deployment — and a managed AI service that logs only abroad does not meet the second.
- Required checks
- Confirm where the AI platform’s logs are stored, because logs held only outside India do not satisfy the residency requirement.
- Rehearse a six-hour report: the clock starts on noticing, not on confirming.
- Check the incident list against what an AI deployment can actually suffer, including data exposure through the model.
- Vendor questions
- Can logs relating to Indian operations be retained inside India for 180 days, and can you evidence that?
- What is your incident notification time to us, and does it fit inside a six-hour regulatory clock?
- Technical controls
- Mirror the relevant logs into an Indian store, and confirm the mirror is complete rather than sampled.
Human review required — take this to your counsel
DPDP is notified and mostly not yet in force — build to it, do not claim it applies
The DPDP Rules commence in stages: the framework rules on publication, the consent-manager rule a year later, and rules 3 and 5 to 16 eighteen months after publication. Notice content, security safeguards, breach reporting, children’s verifiable consent, retention, Significant Data Fiduciary duties and the transfer rule are all in that last bucket. Until then, the earlier sectoral privacy regime is what applies.
- Required checks
- Name which DPDP rule you are designing against and when it commences, in the assessment itself.
- Do not switch off the existing sectoral privacy controls in anticipation — they are the law until the rules commence.
- Assess whether the organisation is likely to be designated a Significant Data Fiduciary, because those duties change the design.
- Vendor questions
- Will your contract terms be updated to reflect DPDP before its substantive rules commence?
- Technical controls
- Build the consent and notice records now, since retrofitting a consent history is not possible after the fact.
India’s AI governance framework is voluntary, and says so in its own principles
The India AI Governance Guidelines are built on seven sutras — Trust, People First, Innovation over Restraint, Fairness and Equity, Accountability, Understandable by Design, and Safety, Resilience and Sustainability — designed to be technology-agnostic and applicable across sectors. "Innovation over Restraint" is a deliberate policy stance: the framework is voluntary and there is no Indian AI statute behind it.
- Required checks
- Use the sutras as a review checklist rather than treating them as a compliance regime.
- Do not tell a stakeholder that India requires an AI risk assessment — it does not.
Structured issue-spotting to support your own review — not legal advice. Verify against the cited primary sources and your counsel.
03What this reading does not know
- Whether any of the data falls into a special or sensitive category.
- Whether any material is covered by legal professional privilege.
- Whether a person reviews the output before it is acted on.
04Instruments these issues point at
- statuteDPDP ActIndia’s data-protection statute: consent-based processing, notice, security safeguards, breach notification, Significant Data Fiduciary duties, extraterritorial reach over services offered to people in India, and penalties up to 250 crore rupees. It commences provision by provision, and the operative detail sits in the DPDP Rules.
- regulationDPDP RulesThe operative rules under the DPDP Act, phased over eighteen months. They cover notice content, consent managers, security safeguards, breach reporting, verifiable consent for children, retention, Significant Data Fiduciary duties and cross-border transfers.
- guidanceIndia AI Governance GuidelinesMeitY’s voluntary AI governance framework under the IndiaAI Mission. Seven sutras — Trust, People First, Innovation over Restraint, Fairness and Equity, Accountability, Understandable by Design, and Safety, Resilience and Sustainability — with six pillars across enablement, regulation and oversight.
05Vendor documents being watched
- Data processing agreementhttps://www.nvidia.com/en-us/agreements/data-processing-addendum/nvidia-cloud-services-data-processing-addendum/not yet fetched
- Privacy policyhttps://www.nvidia.com/en-us/about-nvidia/privacy-policy/not yet fetched
- Security pagehttps://www.nvidia.com/privacy-center/not yet fetched
- Terms of servicehttps://developer.nvidia.com/legal/termsnot yet fetched
06Ask about your own deployment
This page reads the rules against a generic organisation. Your size, industry, data and existing contracts change which of these issues matter and which fall away.