Today’s headline news for Canadian IT solution providers:
- HPE chief technology officer for cloud and AI Fidelma Russo used her Discover general session to introduce “tokenomics” – the argument that agentic AI economics are fundamentally infrastructure economics. She told the Las Vegas audience that continuous AI agents can cost $13,000 per agent per month in the public cloud, and revealed that HPE’s own MindStone AI support platform achieved a 30x cost reduction by moving from the public cloud to HPE Private Cloud AI on-prem – a saving of roughly $100,000 per month.
- Vultr announced it is buying HPE and NVIDIA Blackwell Ultra rack-scale systems – the GB300 NVL72 – with 800GbE Spectrum-X networking to build out next-generation global AI data centres. Vultr CEO J.J. Kardwell called out “decentralized, latency-sensitive workloads” as a driver. The announcement contained no channel component.
- HPE unveiled Morpheus 9, the latest version of its GreenLake virtualization platform, with a built-in MCP server for agent-driven operations. HPE claims up to 90 percent cost reduction versus traditional virtualization, and says more than 2,000 customers and one million cores are already on VM Essentials. A platform migration program offers the first year of Morpheus and VM Essentials at no cost. Zerto’s recovery tools are positioned as an “undo” button for when autonomous AI agents make unintended infrastructure changes.
Podcast: Play in new window | Download
Subscribe: Apple Podcasts | Spotify | Amazon Music | Android | iHeartRadio | Youtube Music | RSS
Read Full Transcript
This epsisode of The Buzz is brought to you by HPE Discover 2026. Check out our full coverage of the event on ChannelBuzz.ca — you’ll find out HPE Discover 2026 News Hub in the menu bar at the top of the page.
Welcome to The Buzz from ChannelBuzz.ca, I’m Robert Dutt, today is Thursday, June 18th, and here’s what’s happening in the channel today.
Today, day three of HPE Discover 2026 in Las Vegas, and the story is the economics of the agentic enterprise. Let’s get to it.
HPE’s chief technology officer for cloud and AI, Fidelma Russo, took the main stage yesterday morning with a message that will resonate with anyone who has watched a client’s cloud AI bill spiral: continuous agentic AI is wildly expensive in the public cloud. Russo cited a figure of $13,000 per month, per agent, for continuous reasoning operations in the public cloud. That is not a pilot. That is production infrastructure.
Her answer is HPE’s take on “tokenomics” – the idea that AI economics are fundamentally infrastructure economics. It comes down to utilization, efficiency, and scale. And HPE has proof. Russo revealed that HPE built its own AI support platform, MindStone, and moved it from the public cloud to HPE Private Cloud AI on-prem. The result: a 30-fold cost reduction, saving roughly $100,000 per month.
That is the argument for why production AI is coming to the data centre. Not because it is fashionable, but because the math stops working anywhere else.
The alternative hyperscaler announced it is buying HPE and NVIDIA Blackwell Ultra rack-scale systems – specifically the GB300 NVL72 – along with 800-gigabit Ethernet Spectrum-X networking, to build out its next generation of global AI data centres. This is a procurement deal, not a partnership, but it is serious hardware at serious scale. Vultr CEO J.J. Kardwell framed it around “decentralized, latency-sensitive workloads across Vultr’s extensive global network.”
Now clearly, this isn’t a channel story unto itself at this moment. This is pure enterprise infrastructure. But it does signal that someone is actually buying the big AI factory gear HPE has been talking about all week.
The GreenLake platform now has a built-in MCP server for agent-driven operations, and HPE says Morpheus 9 delivers up to 90 percent cost reduction compared to traditional virtualization. There are more than 2,000 customers and a million cores already running on VM Essentials. To ease the migration pain, HPE is offering the first year of Morpheus and VM Essentials at no cost through a platform migration program.
There is a caveat: Zerto’s instant recovery and migration support is Morpheus-only for now. No KVM, no Kubernetes natively. But Zerto gets an interesting new job in this agentic world. Russo positioned it as the undo button for when autonomous AI agents make unintended changes to infrastructure – roll back to a known good state instantly.
I’ll be back tomorrow with a reporter’s notebook from the channel leadership breakfast panel at Discover, as we wrap up our coverage of the event this week.
That’s how we’re seeing the headlines from HPE Discover. I’m Robert Dutt for ChannelBuzz.ca, thanks for listening. Have a great day.
Be the first to comment