According to a recent LinkedIn post from FriendliAI, the company is emphasizing a busy March roadmap that includes a new model deployment, endpoint feature enhancements, security milestones, and industry event participation. The post highlights that NVIDIA’s Nemotron 3 Super, described as an open model geared toward agentic tasks, reasoning, and production-scale workloads, is now available through a dedicated endpoint on FriendliAI’s platform.
Claim 55% Off TipRanks
- Unlock hedge fund-level data and powerful investing tools for smarter, sharper decisions
- Discover top-performing stock ideas and upgrade to a portfolio of market leaders with Smart Investor Picks
The post also points to new endpoint capabilities, including support for draft-model speculative decoding, hosted KV cache, and an Anthropic-compatible Messages API, suggesting a focus on performance and interoperability for AI inference customers. These technical additions may strengthen FriendliAI’s value proposition to enterprises seeking cost-efficient, low-latency inference infrastructure, potentially improving customer stickiness and pricing power.
FriendliAI’s LinkedIn content further notes that the company has achieved SOC 2 Type II and HIPAA compliance, positioning its platform for regulated enterprise and healthcare workloads that require higher security and auditability. For investors, these compliance milestones could expand the addressable market into sectors with stricter governance requirements, which often support longer contract durations and higher switching costs.
The post mentions that FriendliAI plans to participate in the HumanX 2026 event in early April, framing it as an opportunity to engage with customers and discuss the future of production AI inference. Active presence at specialized industry conferences may help the company build partnerships, accelerate enterprise sales pipelines, and increase visibility within the competitive AI infrastructure landscape.