https://shipboardai.com Sovereign AI infrastructure, deployed at sea. Commentary and analysis on maritime AI, edge computing, and the hardware that keeps working when the link drops. en-us Fri, 10 Apr 2026 16:25:35 GMT https://shipboardai.com/blog/on-vessel-gpu-selection-guide-h100-l40s-a100 https://shipboardai.com/blog/on-vessel-gpu-selection-guide-h100-l40s-a100 Sat, 11 Apr 2026 00:00:00 GMT A practical buyer's guide for on-vessel GPU hardware. When to pick H100, when L40S is the right sweet spot, when A100 makes sense, and the power and cooling realities nobody tells you before you sign the PO. hello@shipboardai.com (James Calder) infrastructure https://shipboardai.com/blog/claude-mythos-project-glasswing-vessel-security https://shipboardai.com/blog/claude-mythos-project-glasswing-vessel-security Fri, 10 Apr 2026 00:00:00 GMT Project Glasswing used Claude Mythos to find thousands of zero-day vulnerabilities in partner infrastructure, many of them decades old. Here's what that news means for vessel fleet cybersecurity, and why the defensive use case is the smaller story. hello@shipboardai.com (Marcus Hale) security-compliance https://shipboardai.com/blog/satellite-connectivity-ai-workloads-at-sea https://shipboardai.com/blog/satellite-connectivity-ai-workloads-at-sea Thu, 09 Apr 2026 00:00:00 GMT Starlink promised 200 Mbps. You're getting 11. Here's what satellite connectivity actually looks like on a vessel, and how to build AI systems that work regardless. hello@shipboardai.com (James Calder) networking https://shipboardai.com/blog/starlink-gigabit-ambitions-wont-fix-vessel-ai https://shipboardai.com/blog/starlink-gigabit-ambitions-wont-fix-vessel-ai Thu, 09 Apr 2026 00:00:00 GMT SpaceX is talking about gigabit-class speeds for Starlink in 2026. Here's why the throughput number is the wrong thing to fixate on if you care about AI that works on your vessel. hello@shipboardai.com (James Calder) resilience https://shipboardai.com/blog/claude-sonnet-4-6-vessel-context-windows https://shipboardai.com/blog/claude-sonnet-4-6-vessel-context-windows Wed, 08 Apr 2026 00:00:00 GMT Anthropic's new Sonnet model ships with a 1M token context window. For vessels running sovereign AI, that changes what you can do without RAG, without the cloud, and without the link. hello@shipboardai.com (Ethan Marsh) edge-ai https://shipboardai.com/blog/ai-guest-concierge-systems-at-sea https://shipboardai.com/blog/ai-guest-concierge-systems-at-sea Wed, 01 Apr 2026 00:00:00 GMT Cloud-based chatbots fail the moment your ship leaves port. Here's what an on-vessel AI assistant actually looks like, and why your guests won't know the difference from shore-side service. hello@shipboardai.com (James Calder) edge-ai https://shipboardai.com/blog/ai-crew-management-cruise-operations https://shipboardai.com/blog/ai-crew-management-cruise-operations Wed, 25 Mar 2026 00:00:00 GMT Managing 2000+ crew across nationalities, certifications, and compliance requirements, without reliable connectivity. Why on-vessel AI is becoming essential for cruise line HR operations. hello@shipboardai.com (James Calder) resilience https://shipboardai.com/blog/predictive-maintenance-ai-maritime-vessels https://shipboardai.com/blog/predictive-maintenance-ai-maritime-vessels Wed, 18 Mar 2026 00:00:00 GMT When your vessel's systems show early signs of trouble, you need answers immediately, not after the data uploads to shore. Here's why on-vessel AI is changing how maritime maintenance works. hello@shipboardai.com (James Calder) resilience https://shipboardai.com/blog/computer-vision-safety-monitoring-vessels https://shipboardai.com/blog/computer-vision-safety-monitoring-vessels Wed, 11 Mar 2026 00:00:00 GMT Man overboard detection, restricted area monitoring, and bridge watchkeeping, running on hardware aboard the ship, responding in under a second, no cloud required. hello@shipboardai.com (James Calder) edge-ai https://shipboardai.com/blog/why-cloud-ai-fails-at-sea https://shipboardai.com/blog/why-cloud-ai-fails-at-sea Wed, 04 Mar 2026 00:00:00 GMT Your satellite link can't support cloud-based AI on a vessel. Here's why, and what actually works when you're 200 miles from shore. hello@shipboardai.com (James Calder) resilience