Scuttlebutt

Scuttlebutt

Good AI is all about data governance — Hitachi Vantara's AI infrastructure leads

5 takeaways: why a CT scan's "just an image" metadata hides an abdominal scan of a 50-year-old man, and how trains that watch their own tracks became an AI data factory's proof point

Scuttlebutt's avatar
Scuttlebutt
Nov 19, 2026
∙ Paid

SiliconANGLE theCUBE · Nov 18, 2026 · 20m · 14m saved

With Arun Prasad Raman, VP of Hybrid Cloud and AI Infrastructure at Hitachi Vantara · Jeff Lundberg, Principal Product and Solutions Manager at Hitachi Vantara

Good AI is all about data governance — Hitachi Vantara's AI infrastructure leads

A CT scan's own file metadata says almost nothing — "just an image." Arun Prasad Raman says the file itself still encodes that it's an abdominal scan of roughly a 50-year-old man, and that only modern AI can pull that out.

"That's where the importance of unstructured data has grown in recent times, where all these data can be extracted and be used in the AI world." — Arun Prasad Raman

Arun and Jeff Lundberg, both of Hitachi Vantara, told SiliconANGLE theCUBE's John Furrier and Gemma Allen at SC26 why most AI pilots stall before reaching production, and it isn't the model or the GPUs.

The full segment is covered here so you can skip it. 20 minutes of audio, 6 minutes of reading.

Here are the 5 takeaways that matter.


This Scuttlebutt Summary is available in full to premium subscribers. If you're not one already, please consider subscribing to keep reading.


Key Takeaways

  • Files, logs and images carry intelligence that was always there, just hard to extract before AI

    • A CT scan's metadata alone can't say it's an abdominal scan of a 50-year-old man

  • The real pilot-to-production gap is data governance, not GPUs

    • Discover what's AI-ready, classify what's sensitive, define who can access it

  • Hitachi Vantara builds intelligence into the storage layer itself

    • PII detection flags SSNs and birthdates at the file level inside object storage

  • The "Hitachi IQ stack" powers sister business units, including trains that monitor their own tracks

    • Hitachi Rail uses real-time video and telemetry for predictive maintenance

  • Hitachi Vantara is shifting toward bespoke industry solutions over AI-factory price competition

    • AI's buyers inside large enterprises often aren't the traditional storage buyers

1. Intelligence in the Files

Arun argues unstructured data — files, logs, images — has always carried intelligence that was simply hard to extract.

A CT scan is "just an image" at the metadata level, but AI can pull out that it's an abdominal scan of roughly a 50-year-old man

"That's where the importance of unstructured data has grown in recent times, where all these data can be extracted and be used in the AI world." — Arun Prasad Raman

Jeff says what's changed is that AI adoption made that previously-latent value worth unlocking, while exposing how fragmented, access-controlled and compliance-constrained most enterprise data environments actually are.

"The biggest roadblock I think today is... the data... data fragmentation spread across different systems, on-prem, software as a service, out in the cloud." — Jeff Lundberg

That fragmentation was tolerable before AI. It's now a hard blocker.

2. Governance, Not GPUs

User's avatar

Continue reading this post for free, courtesy of Scuttlebutt.

Or purchase a paid subscription.
© 2026 The Golden Age LLC · Privacy ∙ Terms ∙ Collection notice
Start your SubstackGet the app
Substack is the home for writing, podcasting, and media founders.