Owner of VMware, Broadcom, announced that its VMware Cloud Foundation platform is now AI native at the VMware Explore conference a few weeks ago.
It was the latest move by the company to keep up to speed with the rest of the technology industry’s wide and rapid adoption of large language models, yet came as the company battles bad press about licensing policy changes that have dogged it since it acquired virtualisation giant VMware in November 2023.
The ending of the platform’s free tier, reports of aggressive sales tactics to keep subscribers on board, and several court cases focused on existing agreements, including extant perpetual licences, have led many users to rethink what is often the basis of their IT stack. Nutanix, SUSE, and IBM have been among the beneficiaries from those leaving the VMware stable.
But the nature of VMware deployments means that they’re often complex, and extricating workloads out from heavily-virtualised environments running on the platform can come with high migration costs and not insignificant risks to an organisation’s QoS metrics. Better to stay and pay the devil you know than go out on a limb and migrate to an alternative.
By the same token, engineering AI into VMware’s offerings is fraught with danger and the potential for identical fallout. Re-architecturing the VMware platform to bake AI in at the core would mean it would be end-users’ stuttering workloads paying the price for any breaking changes. And the nature of software is that the deeper breaking changes are made, the greater the potential negative ramifications.
Broadcom’s initial aims are to make it simpler for its users to deploy AI models and agents inside their existing environments. VMware Private AI Services is to ship with VCF 9 subscriptions next year, and will comprise of all the elements required to build and run AI on-premise, or at least outside hyperscale facilities. It will include a model store (it’s expected that many organisations will turn to – at least in testing phases – open-source, smaller models), indexing services, vector databases, an agentic AI builder, and a ready-made API gateway to allow optimised machine-to-machine communications between separate AI models that need to work together.
Conference attendees were told AI’s presence in the enterprise was only going to grow, and so it only made sense that AI should be a feature of every VMware-based infrastructure. As it stands, what Broadcom is offering is a nod in the AI direction, but nothing unique nor new. The company also announced improvements to the VMware Tanzu Platform which include simpler publishing of MCP servers, and a new data lakehouse, Tanzu Data Intelligence.
Presumably low-hanging fruit for VMware’s own developers was Intelligent Assist for VCF, a chatbot with access to the VMware knowledgebase. The AI-powered ‘bot will be able to lengthen the time between a user raising an issue or question, and them getting to speak to a human who can help.
The excitement around widespread adoption of containers led many to declare that the end was nigh for ‘traditional’ virtualisation, much in the same way that the explosion of cloud services was to spell the end for on-premise databases, and thus see off Oracle. The reality was, and remains, that legacy infrastructure compels enterprise users to consolidate on the platforms they have invested in, despite rapacious licence fees and high costs.
VMware may be sprinkling the deals between it and its customers with a little AI fairy-dust, but it knows that its long-term income is guaranteed by the presence of legacy infrastructure at the core of the enterprise.
Source: https://www.artificialintelligence-news.com/