Key Highlights

  • MindWalk Holdings deployed OpenFold3 using AMD Instinct MI325X accelerators via Vultr Kubernetes Engine.
  • The deployment achieved an estimated 5-fold reduction in antibody-antigen inference latency.
  • Setup times for the system were streamlined down to minutes.
  • The milestone highlights the scalability of the company's ReefIQ platform during its corporate rollout.
  • Shares closed up 8% at $1.47 following the announcement.

Article

MindWalk Holdings (NASDAQ: HYFT) shares rose 8% to close at $1.47 after the company disclosed a technical milestone in its computational biology infrastructure: the successful operational deployment of OpenFold3 running on AMD Instinct MI325X accelerators, delivered through Vultr Kubernetes Engine. The headline result is an estimated 5-fold reduction in antibody-antigen inference latency compared to the company's previous setup.

For context, antibody-antigen inference refers to computational modeling of how antibodies — the immune system's targeting molecules — interact with antigens, the molecular structures they bind to. This kind of modeling sits at the center of modern computational drug discovery, particularly for Biologics and antibody-based therapeutics, where predicting binding behavior accurately and quickly can meaningfully accelerate the early stages of drug candidate identification.

Cutting that inference time by roughly five-fold isn't just a technical vanity metric. In practical terms, faster inference means researchers can screen more candidate molecules, iterate on designs more quickly, and get from computational prediction to experimental validation in less time. For a company selling infrastructure or platform capabilities into the drug discovery space, that kind of throughput improvement is a tangible product differentiator that can be marketed directly to potential biotech and pharma customers.

The other notable detail is the reduction in setup time, which the company says has been streamlined down to minutes. Deployment friction is a real barrier in enterprise AI infrastructure sales — a platform that's powerful but slow or complex to stand up can lose out to competitors offering a smoother onboarding experience, even if the underlying performance is comparable. Minimizing that friction directly supports easier customer integration, a point management explicitly highlighted.

This milestone is being framed as evidence of the scalability and adaptable design of ReefIQ, the company's broader platform, during what's described as an expanded corporate rollout. Running on AMD's newer MI325X accelerators rather than relying solely on more commonly used alternatives also signals a degree of hardware flexibility, which can matter for cost management and for customers who have already standardized on AMD infrastructure in their own operations.

From an investor standpoint, the combination of faster processing, lower setup friction, and reduced ongoing hardware costs is being positioned as a package that could ease customer Acquisition and improve unit Economics as the ReefIQ rollout continues. Watch for follow-up disclosures on customer adoption, additional hardware partnerships, or expanded platform capabilities as the next indicators of whether this technical milestone translates into commercial traction.