VMblog

delivers daily enterprise tech news, expert analysis, and interviews on #virtualization, #cloud, #AI, #cybersecurity, & modern datacenter technology

[VMblog @vmblog]

A kill switch can stop an #AIagent, but it can't undo what the agent already did. 🛑

@GraviteeIO CEO Rory Blundell explains why enterprises need layered controls, Guardian Agents, and humans in the lead as agents scale.

Read the @VMblog Q&A: https://t.co/Ri8E137P0w

#AI #AIAgents #AgenticAI #EnterpriseAI

[VMblog @vmblog]

Three systems. Three answers to “How many active customers do we have?” 41,200. 47,900. 44,600.

All from the same warehouse. All “working as designed.”

Building got cheap. Being wrong got expensive.

Meet the blue tick for data.

https://t.co/k92OHjKpNK

[VMblog @vmblog]

2027 won't reward the teams that adopted #AI fastest. It'll reward the teams that can prove what their machines did.

Don't miss these #predictions from @SumoLogic CISO Jeremy Powell: agentic SOCs, the CISO role splitting, continuous hunting, deepfake-proof verification, and the AI supply chain crisis.

https://t.co/23SMCQHKfd

#CyberSecurity #CISO #SOC #2027Predictions

[VMblog @vmblog]

#Observability covers metrics, traces and logs. But carbon? Largely missing.

SCI can standardize the calculation. #OpenTelemetry can standardize the signal.

Together, carbon could become #telemetry engineers observe alongside the signals they already use.

What we make observable, we can make actionable.

Check out this latest @VMblog article by Diana Todea of @VictoriaMetrics.

https://t.co/b9uHp8Rlkb

#GreenSoftware #KubeCon

[VMblog @vmblog]

Broadcom quietly pulled VMware's #VDDK, and migration plans took a hit.

@Platform9Sys's Sirish Raghuram explains why proxy- and storage-assisted methods may be a better path for CSPs: a 250 GB disk cloned in 4 seconds, and a full migration in about 15 minutes.

Read the @VMblog Q&A: https://t.co/RSAINNxrtR

#VMware #CloudProviders

[VMblog @vmblog]

Today, @KeeperSecurity announced Phishing Protection, a new feature in the Keeper Browser Extension that warns users before they enter credentials on fraudulent login pages. Built on a zero-knowledge local AI engine, the feature detects look-alike domains, spoofed websites and brand impersonation directly in the browser, helping users avoid handing over credentials to a cybercriminal.

https://t.co/iinLBAuERF

[VMblog @vmblog]

#Kubernetes is becoming the common substrate for #AI and modern apps. At #KubeCon, see how @Broadcom helps enterprises bring AI workloads to production.

https://t.co/YL10ejgeSw

#K8s #Broadcom

[VMblog @vmblog]

Cloud-first isn't the only playbook anymore.

Security demands, licensing shifts, and #AI compute needs are pushing IT leaders to ask a better question: what does this workload need, and can your team actually support it?

Autumn Salama of @DataBank_Ltd explores why workload-first is the smarter infrastructure strategy 👇

https://t.co/lQ5oZMSF6C

#Infrastructure #CloudStrategy #CyberSecurity

[VMblog @vmblog]

Misaligned #GPUs, NICs, and CPUs can cut NCCL bandwidth to 41% of baseline.

#Kubernetes has come a long way on topology awareness: Topology Manager, DRA device pairing, PodGroup placement. But three gaps remain.

Joongi Kim (@lablupinc CTO) says alignment should be a cluster-wide default, not something every workload must spell out.

https://t.co/Ox4JrvT9Bt

#AIInfrastructure #GPU #KubeCon #K8s

[VMblog @vmblog]

'I don’t code, and I've built ~80% of the software I use, tools I used to pay subscriptions for. If I can do that, why can’t a company with an IT budget?'

Isaac Correa of Hellomatik shares his #2027predictions: companies start canceling SaaS tools they've paid for since the 2010s and replace them with their own.

https://t.co/K0QpOXK21f

#predictions

[VMblog @vmblog]

#Phison is making a major move beyond NAND controllers. With $3.4B in first-half 2026 revenue, Pascari enterprise SSDs and aiDAPTIV are putting the company deeper into #AI infrastructure. @VMblog takes a closer look at the strategy and technology learned while at the @ITPressTour #70 in Palo Alto earlier this month.

https://t.co/B4DeAaKMWa

#ITPT #storage

[VMblog @vmblog]

.@ControlUp, a company reinventing IT operations with its AI platform, released findings from an analysis of more than one million real-world help desk tickets, revealing that sign-in problems account for up to 53% of tickets at individual enterprises analyzed. Password resets, account lockouts, disabled accounts and MFA were the main causes.

Learn more: https://t.co/unSzPd8nsp

[VMblog @vmblog]

#AIagents can write code in hours. But ~57% of software changes need a database update, and a schema approval can take up to 2 weeks.

GemDB's answer: the code is the schema. Change the data model and the database change is already done. No migration script, no downtime.

Learn more in this @VMblog Q&A with Sandeep Parikh, CPO at #GemDB.

https://t.co/QfpY7pmQ03

#AI #Databases #DeveloperTools

[VMblog @vmblog]

Some #cyberthreats never stay dead. 🧟

The Phobia DGA family fell from ~62,000 device matches to just 600–700 a day, then surged back to ~25,000 in May.

A quiet threat isn't a gone threat. DNS can show the early warning signs. Here's what to watch for in this @VMblog article by Yaëlle Harel of @efficientip:

https://t.co/Z1EYMCdXEB

#DNS #CyberSecurity

[VMblog @vmblog]

The #AIsecurity incident that does the most damage in 2027 may be one your own company set up.

@Dtex_ai's Marshall Heilman's #predictions: an agent follows authorized instructions, takes actions nobody would have approved, and a critical business operation goes down. No attacker needed.

Who owns that risk? The CEO and the board.

https://t.co/xRMPvS3hMn

#AIAgents #CyberSecurity #InsiderRisk #2027Predictions

[VMblog @vmblog]

Every #Kubernetes backup holds a hidden #datalake: app data, manifests, RBAC configs, all versioned. But querying it today means a 30-90 minute restore for one SQL query.

Expose backups as Apache Iceberg tables and you get time-travel SQL for compliance, security forensics, and #AIagents. No restore needed.

The building blocks are here. 🧱

Don't miss this @VMblog article by Michael Haigh of @Veeam #Kasten.

https://t.co/ZME67d0dHT

#KubeCon

[VMblog @vmblog]

In 2027, the hardest part of #AI won't be the model. It'll be everything around it.

Data validation, provenance, testing whole workflows, and knowing which decisions the model shouldn't make.

Mandar Chaudhari of Land IQ on building AI that works when the data is messy 👇

https://t.co/G9xWckLJTa

#predictions #2027predictions #VMblog

[VMblog @vmblog]

.@GetNerdio, the leading automated end-user computing (#EUC) platform for Microsoft Cloud technologies, released its 2026 State of the #MSP Business Report, based on a survey of hundreds of managed service provider decision-makers responsible for managing #Microsoft365 and #Azure across their customer bases.

Find out more: https://t.co/Aaqq4oMQwi

[VergeOS @VergeUCI]

Most VMware exits stall on testing, not migration.

VergeOS takes snapshots in under a second and gives each workload its own isolated Virtual Data Center for rigorous, simultaneous testing across multiple users and AI test harnesses for 100% post-migration confidence.