Mike Leone

VP & Principal Analyst at Moor Insights & Strategy covering the intersection of data, analytics, and AI. Love bleeding edge tech, Boston sports and family time.

[Mike Leone @LeoneDataAI]

I've joined @MoorInsStrat as VP & Principal Analyst covering data, analytics, governance, and storage — and where AI intersects them all.

Grateful to learn the playbook from @PatrickMoorhead and the MI&S team.

Sat down with Erin Zehr for my first piece on why I made the move: https://t.co/nglrQmTlSy

[Mike Leone @LeoneDataAI]

Off to the @couchbase #Catalyst2026 keynote. Room’s buzzing.

[Mike Leone @LeoneDataAI]

https://t.co/eYjmebrneI

[Mike Leone @LeoneDataAI]

I give @ServiceNow a lot of credit for Flow. Plenty of incumbents would've kept selling just the big platform. ServiceNow built a chat-first service desk a 200-person company can buy on a credit card. More in my latest Field Note, link below.

[Mike Leone @LeoneDataAI]

@Microsoft Got into it with @journoanirban for @CIOonline.
https://t.co/gPpHg3dK2f

[Mike Leone @LeoneDataAI]

Liked seeing @Microsoft back Ossie for two reasons. One, a lot of companies keep their business definitions in Power BI. And two, now, whoever wrote those definitions gets to take them wherever they go if they want.

[Mike Leone @LeoneDataAI]

Loved the joke on stage about agents building beautiful Power BI dashboards, then everyone just looking for the export to Excel. Too real. Some things AI can't fix. 😂 #FabConEurope #SQLConEurope

[Mike Leone @LeoneDataAI]

Fabric IQ can now build an ontology from your Power BI semantic models. With 20+ million semantic models in Power BI today, that's a big head start for a lot of teams. #FabConEurope #SQLConEurope

[Mike Leone @LeoneDataAI]

As a perf guy, obviously results may vary. But based on what they showed on stage, it's worth a double click.

[Mike Leone @LeoneDataAI]

We can keep talking agents, but GPU acceleration in Fabric's warehouse is still a sleeper story for me. #FabConEurope #SQLConEurope

[Mike Leone @LeoneDataAI]

Biggest cheer of the keynote so far went to billing. On-demand across Fabric, no upfront capacity needed. #FabConEurope #SQLConEurope

[Mike Leone @LeoneDataAI]

OneLake's getting a policy engine. Set the rules once, right where the data lives. #FabConEurope #SQLConEurope

[Mike Leone @LeoneDataAI]

Agents that leave artifacts behind, plus a Monitor hub that shows what they ran. That's how Fabric gets data teams to trust the agents. #FabConEurope #SQLConEurope

[Mike Leone @LeoneDataAI]

Most companies I talk to treat sovereignty as an insurance policy meaning companies pay for the option to meet whatever residency or control rules come their way, without rebuilding anything if those rules change. SQL Server on Azure Local going GA gives SQL shops that option without a rebuild. #FabConEurope #SQLConEurope

[Mike Leone @LeoneDataAI]

IQ is where this gets won or lost. Copilot and Agent 365 are only as good as the context it hands them, and that's where Fabric earns its spot in the story.

[Mike Leone @LeoneDataAI]

The AI pitch from @Microsoft at #FabConEurope #SQLConEurope comes down to people working in Copilot, agents governed in Agent 365, and Microsoft IQ giving both the same context on your business.

[Mike Leone @LeoneDataAI]

Ready for the @Microsoft #FabConEurope #SQLConEurope keynote!

[Mike Leone @LeoneDataAI]

Barcelona for @Microsoft #FabConEurope #SQLConEurope. Analyst summit kicks off today. Let's go!

[MoorInsightsStrategy @MoorInsStrat]

🔐 New research from VP & Principal Analyst Michael Leone (@LeoneDataAI) examines how confidential AI brings models to regulated data while protecting records and model weights.

The brief explores @VAST_Data Enclave and four requirements buyers should use to evaluate confidential enclaves.

https://t.co/qawkCigb8j

@PatrickMoorhead
@VAST_Data

[Mike Leone @LeoneDataAI]

@VAST_Data https://t.co/OqqkPT1Wzn

[Mike Leone @LeoneDataAI]

Today, @VAST_Data launched DataEnclave, bringing confidential AI to regulated companies that want top-tier models on data that can't leave the building. I wrote a research brief on why the model has to come to the data. Link below.