VMware's AI strategy: A focus on India and platform agnosticism explained

At VMware Explore 2024, Chris Wolf, global head of AI and advanced services in Broadcom's VMware Cloud Foundation Division, engaged in a candid Q&A session with media and analysts, shedding light on the company's AI strategy, recent workforce reductions, and their commitment to the Indian market.

Addressing Layoffs: Business Needs, Not AI

Wolf addressed concerns about recent layoffs at VMware, which some had attributed to the integration with Broadcom and the rise of AI. Wolf clarified, "Our reorganisation had nothing to do with AI." He told Business Today that the changes stemmed from the need to create a unified cloud platform, stating that “a cloud has to be a unified stack that gives you all of your infrastructure in a unified fashion.” He elaborated that before the Broadcom acquisition, separate teams handled product management for various parts of the cloud offering, leading to "a lot of overlap.”

Wolf reassured that the changes were purely for "business purposes," stating, “Hopefully what you’ve seen here today is there hasn’t been an impact on innovation. We’re innovating well, we're more focused than we've ever been, and I think you're going to see that in our products going forward.”

VMware's AI Strategy: Platform Agnosticism and Customer Choice

Wolf highlighted VMware's AI strategy, which centres around platform agnosticism and offering customers a wide range of choices.

"We, like a lot of companies, have been doing work and research work in artificial intelligence for more than 10 years," Wolf explained. He noted that their current AI strategy took shape around two years ago, based on the hypothesis that "the space was moving too quickly and most enterprises would not want to bet on a single platform."

He believes history has proven them right, stating, "Most customers, they don't want to consume the entirety of the AI services from one provider, period." This realisation shaped their approach.

"The first part is giving you an AI platform that has all of the tools you need and automation you need to be very successful," Wolf explained. “And then second, making sure the platform has the APIs and supports the services that customers require to build upon. And then supporting a very large ecosystem that promotes choice, both in open source and in the commercial space.”

Differentiation and Collaboration

Wolf distinguished VMware's AI strategy from competitors like Red Hat, highlighting their focus on the platform rather than a full-stack solution.

"Where we are different is we lead with the platform,” Wolf said. "And because of that, the entirety of the AI ISV community doesn't see us as a competitor, and it's very easy for us to work together with customers." He illustrated this by comparing VMware's approach to that of a company like Cohere, saying, “It's hard for me to go into a customer account with AWS because AWS sells competitive services to them. And we don't.”

India: A Key Player in VMware's AI Vision

Wolf acknowledged the significant role India plays in VMware's AI development. "A significant portion of the features you saw recently are developed by our engineering teams in India," he told Business Today. He praised the team's expertise and dedication, stating, "They are actively driving our AI agenda, doing it quite well."

Beyond development, Wolf highlighted their on-the-ground efforts in India, stating, "They're also partnering with our field teams to help with our first customers in India to see value from the platform."

Addressing the concern about a lack of AI expertise in the industry, Wolf shared how VMware is bridging the gap: "The way we've been helping to enable customers is our engineers have hands-on with these customers to make them successful with their deployment and first application."

He confirmed strong traction for their AI solutions in India, revealing, "For a product that launched in May, we're in the process of onboarding our second customer. So, we're seeing real strong traction."

Challenges and Opportunities in Private AI Adoption

Wolf identified the main challenge in private AI adoption as "identifying the right use case.” He explained, "We had a lot of hype really in the previous year. And now it's down to I want to be successful in AI, but I want to make sure I'm doing something that I can measure that has business value.”

He acknowledged the lack of expertise in the market, stating, "If you look into the system integrator space, there's a lot of systems integrators that are still just building their AI practices. And it's, so we're having to be selective in terms of who we work with to ensure we spend time vetting the integrator."

He suggested customer service as a good starting point for private AI adoption, "Because every industry vertical, they have some type of contact centre or customer service use case. It has their private data. It's an area where they can see some value there."

Security and Reliability

Wolf underlined the importance of security and reliability in private AI deployments, highlighting VMware's approach of leveraging existing enterprise controls.

"On the surface, what we wanted to do was bring the AI model and compute to where customer data was," Wolf said. "So they can they can use a lot of their existing security controls to protect the data."

He discussed the challenges of container governance and data security in AI, stating, “It is very easy to create new backdoors and bypass access management with AI if you're not careful.” He stressed their efforts to provide guidance and documentation to address these concerns.

VMware's Focus on Innovation

Despite the challenges, Wolf expressed confidence in VMware's AI trajectory and their commitment to delivering value to customers. He mentioned their work on confidential VMs and containers, stating, "That matters for the AI use case because through confidential computing, when you're running an AI model in shared memory, we will be able to provide guarantees of model confidentiality."

He also highlighted their work on resource scheduling, claiming, "There is nobody that has the sophistication of orchestration and being able to intelligently place workloads across GPUs, across clusters like what we have. There's nothing even close."