- Enterprises now realize AI is a "big-dollars game," said VAST VP John Mao
- But they're struggling to focus and find ROI
- AI is forcing enterprises to think about infrastructure issues they had not considered for decades
Enterprises are becoming comfortable with the nine-figure checks they need to write to build infrastructure that's ready for AI, said John Mao, VP of Global Business Development at VAST Data.
"There are a lot of people still trying to figure out how to justify the absurd amount of spend that they're seeing happening around them," Mao said, in an interview with a half-dozen journalists and analysts at the VAST Forward 2026 customer conference in Salt Lake City last week. "It's not a low-investment game — it's a big-dollars game, and that's very uncomfortable for a lot of executives who have to make that investment."
But reluctance to spend is fading, and VAST is now seeing enterprises spend $100 million on AI infrastructure, Mao said. These enterprises are moving from small proofs-of-concept (POCs) running on a few GPUs to production. "That gap for most enterprises is wider than they realize," Mao said.
Even as they spend, enterprises struggle to focus amid thousands of project ideas, and question where ROI is coming from.
AI is forcing enterprises to think about issues they have not had to consider for decade, such as liquid cooling and data center planning, as well as managing large clusters of GPUs. These organizations don't have the skills, Mao said. VAST simplifies some of those problems related to managing data across hybrid clouds.
At its conference last week, VAST introduced PolicyEngine and TuningEngine to help organizations deploy AI that's "secure, trusted and self-learning," the company said.
And VAST CEO Renen Hallak challenged telcos to regain leadership lost to AWS and Netflix.
Connecting personas for enterprise AI
Partners are a critical part of VAST's enterprise strategy, bridging the gap between business stakeholders and execution and implementation teams, Mao said. Partners can also bridge the gaps between different specialties in the enterprise — infrastructure, storage, applications and line of business — who have different concerns and may not even talk with each other.
Phil Manez, VAST VP for go-to-market execution, agreed. "We could be selling a product to one organization and selling to three different personas," he said. In other words, individual roles within an enterprise often have different concerns that might even conflict with one another, and VAST and its partners have to address all those concerns. "You've got to understand the culture of the enterprise," Manez added.
For example, a data analytics specialist at an enterprise might be concerned about Kafka, data warehouses and data lakes. An AI specialist would be concerned about scaling vector databases. Storage specialists worry about efficiency, resiliency and operational simplicity. And data analytics specialists often work around storage specialists, rather than collaborating with them. "They don't have to talk to storage, and they don't want to," Manez said.
Ironically, the renewed enterprise interest comes as supply chain constraints make it difficult for buyers, including enterprises, to acquire servers and components. "I feel like all I talk about now is the supply chain," Manez said.
Supply chain constraints play to VAST's strength, providing a common platform that can run traditional enterprise workloads, data analytics and AI together, rather than requiring separate, siloed platforms for each. "Those silos create a lot of stranded capacity and inefficiencies," Manez said.
VAST opportunity in the back office
VAST sees major opportunities for AI in automating back office operations, Mao said. "It's not the sexy stuff you hear about in the news, but it's the most logical place," he said.
VAST's TwelveLabs partnership is a great example of that kind of opportunity, Mao said. TwelveLabs provides video analytics — the company's service provides natural language descriptions of events on a video. In announcing the partnership with TwelveLabs last week, VAST showed this slide, where TwelveLabs is essentially providing an AI-generated play-by-play of a soccer match.
Nearly every enterprise today has security video, whose applications go beyond security when coupled with AI, Mao said. For example, a retailer could use that video to analyze store traffic — how many customers walk down every aisle in a store, how long they linger at the aisle cap and look at the ads there. Banks can use video analytics to identify money laundering.
TwelveLabs now runs on the public cloud — VAST's partnership with TwelveLabs, will enable the service to run on premises, making TwelveLabs more suitable for regulated industries and others where public cloud is not an option, Mao said. TwelveLabs started life in media and entertainment, where video is public; VAST will enable the company to move into other enterprise verticals.
VAST sees its mission as simplifying complexity, both for born-in-the-cloud startups like TwelveLabs, and for enterprises wading into the deep waters of AI. "I firmly believe we have an opportunity to make AI more democratized across the enterprise market, because we are trying to productize and codify a lot of the complexity out of the equation," Mao said.