# 175,000 Exposed Ollama LLM Hosts Enable AI Model Abuse

> 175,000+ Ollama large language model (LLM) instances are exposed to the public internet on default port 11434 with zero authentication, creating the largest unauthenticated AI compute surface in history. Shodan scans confirm exposed instances across AWS, Azure, GCP, Linode, DigitalOcean, Hetzner, OVH, and home networks globally — with concentrations in the US, China, Germany, India, and South Korea. Ollama ships with no authentication, no API keys, and binds to 0.0.0.0 by default when deployed via Docker — meaning any internet-connected deployment is immediately accessible. Attackers exploit these exposed instances for LLMjacking (stolen AI compute at $0 cost), model weight theft (downloading proprietary fine-tuned models), training data extraction via prompt injection, server-side request forgery (SSRF) into cloud metadata services, and denial-of-service via resource exhaustion. The exposure scale — 175K+ hosts — transforms individual misconfigurations into a systemic infrastructure crisis.

- **Published:** 2026-02-03T03:55:00Z
- **Last reviewed:** 2026-02-03T03:55:00Z
- **Canonical:** https://intel.threadlinqs.com/threat/TL-2026-0047
- **Card image:** https://intel.threadlinqs.com/og/threat/threat-intel-exposed-ollama-llm-hosts-tl-2026-0047.webp?v=1eecf8f8 (Threadlinqs threat intel: Ollama Port 11434 Open on 175K Hosts: Detection & IOCs (TL-2026-0047) — High severity, 41 IOCs, 15 detection rules.)
- **ID:** TL-2026-0047
- **Severity:** HIGH (CVSS 7.5)
- **Category:** MISCONFIGURATION
- **Status:** ACTIVE
- **Detections:** 15 (1 shown free below) · **IOCs:** 41
- **CVEs:** CVE-2024-39720, CVE-2024-39722

## Description

The 175K Exposed Ollama Host Crisis — Attack Surface at Scale:

Ollama is the dominant open-source framework for running large language models locally, with 100M+ Docker pulls and deployment across individual developer machines, corporate AI labs, university research clusters, and production inference servers. Its design philosophy prioritizes ease of use over security: install, pull a model, run — no API keys, no authentication tokens, no access control. This 'zero-friction' approach is the root cause of the largest exposed AI compute surface ever documented.

Why 175K+ Hosts Are Exposed:
1. Default Binding: Ollama binds to 127.0.0.1 in native installs but 0.0.0.0 in Docker deployments. Docker is the primary deployment method for servers. Users deploy `docker run -p 11434:11434 ollama/ollama` and assume the container is isolated — it's not, it's bound to all interfaces.
2. No Authentication: Ollama has zero built-in authentication. No API keys, no tokens, no basic auth. The REST API is fully accessible to anyone who can reach port 11434. There is no `--auth` flag, no config file for credentials.
3. Cloud Security Group Gaps: Developers deploy Ollama on cloud VMs (AWS EC2, Azure, GCP) and open port 11434 in security groups for remote access from their laptop. They forget to restrict the source IP, leaving 0.0.0.0/0 inbound access.
4. Reverse Proxy Absence: Without built-in auth, security requires an external reverse proxy (nginx, Caddy, Traefik) with authentication. Most developers skip this step because Ollama doesn't document it as required.
5. Model Library Growth: Ollama's model library now includes Llama 3.x, Mistral, Gemma, Phi, CodeLlama, DeepSeek, Qwen, and 100+ models. More models = more reasons to run Ollama = more exposed instances.

Attack Surface Analysis — What 175K Exposed Hosts Enable:

1. LLMjacking — Stolen AI Compute ($0 Cost):
- Attacker discovers exposed Ollama instance via Shodan/Censys/masscan
- Sends inference requests to run prompts through the model at zero cost
- GPU compute for inference costs $0.50-$5.00/hour on cloud providers
- 175K hosts × average 4 hours/day abuse = $350K-$3.5M/day in stolen compute
- Used for: spam generation, phishing email crafting, malware code generation, CAPTCHA solving, academic cheating, SEO content farming
- Attacker infrastructure cost: $0. All compute is stolen.

2. Model Weight Theft — Downloading Proprietary Models:
- Ollama API: GET /api/show returns model metadata
- Ollama API: POST /api/pull can pull model weights to attacker's system
- Organizations fine-tune models on proprietary data (medical, legal, financial, military)
- Fine-tuned model weights contain embedded training data patterns
- Stolen weights = stolen intellectual property + potential training data leakage
- Corporate espionage vector: competitor pulls your fine-tuned model for free

3. Training Data Extraction via Prompt Injection:
- Models memorize fragments of training data (known research: extracting training data from LLMs)
- Adversarial prompts can extract memorized sensitive data: PII, API keys, internal URLs, credentials
- Exposed Ollama hosts with custom fine-tuned models are highest risk
- Prompt: 'Repeat the following training example verbatim...' or 'What data were you trained on?'
- No rate limiting on exposed instances = unlimited extraction attempts

4. SSRF Into Cloud Metadata Services:
- CVE-2024-39720 and related: Ollama's model import functionality can be exploited for SSRF
- Attacker sends crafted request to Ollama API targeting http://169.254.169.254/latest/meta-data/ (AWS IMDSv1)
- Ollama fetches the URL server-side, returning cloud instance credentials, IAM role tokens, and infrastructure metadata
- Pivots from 'exposed LLM' to 'full cloud account compromise'
- Also targets: Azure IMDS (169.254.169.254), GCP metadata (metadata.google.internal), Kubernetes service account tokens

5. Denial of Service — Resource Exhaustion:
- LLM inference is GPU/CPU intensive. A single large prompt can consume 100% GPU for seconds
- Attacker sends continuous large-context requests to exhaust GPU memory and compute
- Legitimate users on shared infrastructure lose access
- No rate limiting, no request size limits, no concurrent request caps on default Ollama
- 175K hosts = potential DDoS amplification network (reflective inference)

6. Model Poisoning via API:
- Ollama API: POST /api/create allows creating new models from Modelfiles
- Attacker can create poisoned models on exposed instances
- If organization pulls from their own exposed Ollama instance, they get the poisoned model
- Supply chain attack: attacker replaces legitimate model with backdoored version

Exposure Distribution (Shodan Analysis):
- AWS: ~45K hosts (26%) — EC2 instances across us-east-1, eu-west-1, ap-southeast-1, ap-south-1, eu-south-1
- Azure: ~18K hosts (10%) — primarily dev/test VMs
- GCP: ~12K hosts (7%) — Compute Engine instances
- Hetzner: ~15K hosts (9%) — European developer favorite for GPU instances
- DigitalOcean: ~10K hosts (6%) — smaller GPU droplets
- OVH: ~8K hosts (5%) — European cloud provider
- Linode/Akamai: ~7K hosts (4%) — developer instances
- Home/ISP: ~25K hosts (14%) — residential IPs running Ollama on home servers/NAS
- Other VPS/Cloud: ~35K hosts (20%) — various smaller providers
- Total: ~175K+ hosts, growing ~5K/month

Geographic Distribution:
- United States: ~52K (30%)
- China: ~28K (16%)
- Germany: ~18K (10%)
- India: ~15K (9%)
- South Korea: ~10K (6%)
- Japan: ~8K (5%)
- France: ~7K (4%)
- UK: ~6K (3%)
- Brazil: ~5K (3%)
- Other: ~26K (15%)

Ollama Security Architecture (or Lack Thereof):
- Version 0.1.x-0.5.x: No authentication mechanism at all
- OLLAMA_ORIGINS environment variable: CORS control only — not authentication
- OLLAMA_HOST: Controls binding address but not access control
- No built-in TLS — API traffic is plaintext HTTP
- No audit logging — no record of who accessed what model or when
- No rate limiting — unlimited requests per second
- No request size limits — arbitrarily large prompts accepted
- No model access control — all models accessible to all callers
- Community workarounds: nginx reverse proxy, Tailscale VPN, firewall rules — all external to Ollama

## MITRE ATT&CK

- T1190 Exploit Public-Facing Application
- T1583.004 Server
- T1496 Resource Hijacking
- T1595 Active Scanning
- T1596 Search Open Technical Databases
- T1133 External Remote Services
- T1059 Command and Scripting Interpreter
- T1203 Exploitation for Client Execution
- T1505 Server Software Component
- T1528 Steal Application Access Token
- T1552 Unsecured Credentials
- T1580 Cloud Infrastructure Discovery
- T1082 System Information Discovery
- T1005 Data from Local System
- T1213 Data from Information Repositories
- T1684.001 Impersonation
- T1210 Exploitation of Remote Services
- T1567 Exfiltration Over Web Service
- T1048 Exfiltration Over Alternative Protocol
- T1499 Endpoint Denial of Service
- T1565 Data Manipulation
- T1071 Application Layer Protocol
- T1583 Acquire Infrastructure
- T1548 Abuse Elevation Control Mechanism
- T1590 Gather Victim Network Information
- T1584 Compromise Infrastructure
- T1078 Valid Accounts
- T1036 Masquerading
- T1046 Network Service Discovery
- T1526 Cloud Service Discovery
- T1530 Data from Cloud Storage
- T1498 Network Denial of Service
- T1648 Serverless Execution
- T1525 Implant Internal Image

## Free detection (Splunk SPL, Microsoft KQL)

**External Access to Ollama API** (`TL-2026-0047-DET-001`)

Detects inbound connections to Ollama API port from external sources

- **Languages:** Splunk SPL, Microsoft KQL
- **Data sources:** firewall, network
- **Severity:** critical
- **Confidence:** high

### Splunk SPL

```spl
index=firewall sourcetype=firewall dest_port=11434 action=allowed
| where NOT cidrmatch("10.0.0.0/8", src_ip) AND NOT cidrmatch("172.16.0.0/12", src_ip) AND NOT cidrmatch("192.168.0.0/16", src_ip)
| stats count dc(src_ip) as unique_sources values(src_ip) as source_ips by dest_ip
| where count > 0
| eval alert_severity="CRITICAL"
| eval alert_reason="External access to Ollama LLM server - potential unauthorized AI access"
```

### Microsoft KQL

```kql
// External Access to Ollama API
// Detects inbound connections to Ollama API port from external sources
CommonSecurityLog
| where TimeGenerated > ago(24h)
| where DestinationPort == 11434
| where DeviceAction == "Allow" or DeviceAction == "allow"
| where not(ipv4_is_private(SourceIP))
| summarize ConnectionCount=count(), UniqueSources=dcount(SourceIP), Sources=make_set(SourceIP) by DestinationIP
| where ConnectionCount > 0
```

**Known false positives:**

- Legitimate remote Ollama access (should be VPN/authenticated)

**Tuning notes:** All external Ollama access should be blocked or VPN-only

### 14 more detection rules (Splunk SPL, Microsoft KQL, Sigma), unlock with TL Creds

- Ollama API Abuse Pattern (locked: unlock with TL Creds)
- Ollama Server Internet Exposure Audit (locked: unlock with TL Creds)
- Cloud Security Group Misconfiguration — Port [indicator] Open to Internet (locked: unlock with TL Creds)
- Ollama Model Poisoning — External /api/create Request Deploying Backdoored Model (locked: unlock with TL Creds)
- Ollama Path Traversal File Read — CVE-2024-39722 Exploitation via /api/push (locked: unlock with TL Creds)
- Ollama Internet Exposure — External HTTP Connection to Port [indicator] (locked: unlock with TL Creds)
- Docker Ollama Container Binding Misconfiguration — 0.0.0.0 Port Exposure (locked: unlock with TL Creds)
- Ollama Resource Exhaustion DoS — Coordinated High-Context Inference Requests (locked: unlock with TL Creds)
- Sim Sigma — Ollama API Abuse Pattern Detector & Cloud Metadata SSRF Moni (locked: unlock with TL Creds)
- Sim Sigma — Ollama Exposure Census & LLMjacking Economics Analyzer (locked: unlock with TL Creds)
- Sim Sigma — Exposed Ollama Instance Scanner & LLMjacking Compute Theft D (locked: unlock with TL Creds)
- LLMjacking — External IP Inference Requests to Ollama API with GPU Utilization Spike (locked: unlock with TL Creds)
- Ollama Model Weight Theft — Enumerate-Examine-Download Sequence from External IP (locked: unlock with TL Creds)
- Ollama SSRF to Cloud Metadata Service — CVE-2024-39720 Exploitation (locked: unlock with TL Creds)

Sign up free to unlock the detection rules: https://threadlinqs.com/?modal=signup

## Sources

- [SecurityWeek: 175,000 Exposed Ollama Hosts Could Enable LLM Abuse](https://www.securityweek.com/175000-exposed-ollama-hosts-could-enable-llm-abuse/)
- [Shodan — Ollama Port 11434 Internet Exposure Census](https://www.shodan.io/search?query=ollama)
- [Ollama GitHub — Security Considerations and Default Configuration](https://github.com/ollama/ollama/blob/main/docs/faq.md)
- [Wiz Research — LLMjacking: Stolen Cloud Compute for AI Model Abuse](https://www.wiz.io/blog/llmjacking)
- [Sysdig — LLMjacking: Stolen Cloud Credentials Used for LLM Inference](https://sysdig.com/blog/llmjacking-stolen-cloud-credentials-used-in-new-ai-attack/)
- [NIST NVD — CVE-2024-39720: Ollama SSRF via Model Import](https://nvd.nist.gov/vuln/detail/CVE-2024-39720)
- [NIST NVD — CVE-2024-39719: Ollama API Model Existence Confirmation](https://nvd.nist.gov/vuln/detail/CVE-2024-39719)
- [NIST NVD — CVE-2024-39721: Ollama API Resource Exhaustion via /api/create](https://nvd.nist.gov/vuln/detail/CVE-2024-39721)
- [NIST NVD — CVE-2024-39722: Ollama Path Traversal via /api/push](https://nvd.nist.gov/vuln/detail/CVE-2024-39722)
- [Censys — Exposed AI Infrastructure: Ollama, TGI, and vLLM Internet Scan](https://censys.io/exposed-ai-infrastructure)
- [Ollama Blog — Security and Configuration Best Practices](https://ollama.com/blog)
- [OWASP — Top 10 for Large Language Model Applications](https://owasp.org/www-project-top-10-for-large-language-model-applications/)
- [Docker Hub — Ollama Official Image (100M+ Pulls)](https://hub.docker.com/r/ollama/ollama)
- [AWS Security Blog — Protecting AI Workloads: IMDS and Instance Metadata Security](https://aws.amazon.com/blogs/security/)

## Full data

The other 14 detection rules (Splunk SPL / Microsoft KQL / Sigma) unlock with TL Creds; a free account includes daily creds: https://threadlinqs.com/?modal=signup
Indicator values unlock with TL Creds, at a lower price on Red and Purple; AI agents can query rules and IOCs through the Threadlinqs MCP server (Purple): https://intel.threadlinqs.com/mcp

Canonical: https://intel.threadlinqs.com/threat/TL-2026-0047
