Spencer Honeyman has moved to President at Vi Labs, the enterprise AI platform for healthcare, wellness, and life sciences that completed a $145 million financing at a $1.64 billion valuation in May - the same month Vi launched a suite of AI agents targeting hospital operators, health plans, and fitness enterprises.
Honeyman's progression through the company - from SVP of Strategic Partnerships to Chief Commercial Officer to CRO to now President - tracks the arc of a company that has stopped being a platform-in-development and started behaving like a growth-stage software business. The appointment itself isn't the story. What it signals is.
The market structure shift no one is tracking publicly
Enterprise AI has surged from $1.7 billion in 2023 to roughly $37 billion today, capturing 6% of the global SaaS market and growing faster than any software category, according to Menlo Ventures. Healthcare AI specifically was valued at $36.7 billion in 2025 and is projected to reach $505 billion by 2033.
But the transition happening inside these numbers is more important than the numbers themselves. The market is shifting from infrastructure AI - the models, the chips, the foundational layers - to application AI: the software that actually sits inside a hospital's workflow, a health plan's claims pipeline, or a wellness chain's member engagement system. The companies that win this transition don't need to own the model. They need to own the vertical.
This is what separates Vi from the training-focused names dominating public AI coverage. Vi doesn't build foundation models. It builds the execution layer - the AI agents that read a health plan's data and automatically trigger interventions, or that optimize a gym chain's member retention in real time. The company claims 4X ROI for enterprise clients. Whether that holds across scale is the question, but the architecture bet is clear: vertical AI agents, not horizontal model APIs.
Why the promotion timing matters
Honeyman's move to President arrives as Vi transitions from fundraising milestones to revenue execution. The company has raised $169.8 million across 12 rounds, with Alpha Partners as a lead investor - a firm known for backing businesses that can show unit economics, not just moonshot promises. The $145 million transaction at $1.64 billion put Vi in unicorn territory, but unicorns in enterprise AI are becoming table stakes, not moats.
What changes with a President who came up through commercial operations is accountability for the go-to-market machine. Early-stage enterprise AI companies hire presidents from product or engineering. Vi is promoting from revenue - which signals the company believes its bottleneck is no longer building the right technology. It's closing deals at scale.
Put plainly: when the commercial leader becomes the operator, the thesis has shifted from 'will this work' to 'can this grow.'
The competitive field is getting crowded - and unforgiving
Vi isn't alone in this space. Companies like Olive AI raised hundreds of millions to build hospital automation AI before struggling to prove unit economics. PathAI, Tempus, Abridge, Qure.ai - the field is dense, and most of these names are still private. The graveyard of healthcare AI startups that could demonstrate a pilot but never scaled to production revenue is long.
Vi's differentiation is its data architecture. The company operates on household-level data spanning health plans, providers, and wellness providers. That breadth - the ability to see a member across payers and providers - is what allows its agents to actually execute decisions rather than generate recommendations. That's the architectural advantage. The question is whether it's durable.
Here's where the risk lives: the enterprise AI market is growing fast enough that you don't need to monopolize it to build something large. But it's also growing fast enough that incumbents - UnitedHealth, Optum, Elevance, even the big tech clouds - can replicate the agent layer once the economics become clear. The window for a pure-play private company to own a vertical before the incumbents fold it into their platform is narrowing.
Where does this leave capital?
Vi is private. You can't buy it today. But the signal it's sending matters for how I think about capital allocation in the public AI trade.
The transition from infrastructure AI to application AI means the next wave of market-cap creation won't come from the companies selling shovels. It'll come from the companies using them - vertically, in specific industries, with data advantages competitors can't easily replicate. The infrastructure names have front-loaded their returns. The application layer hasn't.
That means I'm looking for public companies that sit between the infrastructure and the vertical outcome. Cloud platforms that embed healthcare-specific AI agents. Software vendors that are migrating from point solutions to AI-native workflows. The names that are doing today what Vi is doing privately.
The debate isn't whether enterprise AI in healthcare is a real market. It clearly is. The debate is whether the return profile of holding infrastructure names - where much of the upside is already priced in - is better than finding the application-layer stories that are still finding their footing. In my opinion, for investors who can tolerate the execution risk of younger business models, the answer is shifting toward the application side.
Honeyman's promotion to President at Vi is a small data point from a private company. But it tells me the people building the vertical AI layer believe the bottleneck has moved from technology to go-to-market scale. That's exactly the signal I look for when deciding whether to rotate capital from a crowded infrastructure trade into the next phase of the AI cycle.
What would change this view? If the enterprise AI market growth decelerates sharply - Menlo's $37 billion figure needs to hold - or if incumbents replicate the agent layer faster than vertical specialists can scale revenue. If either happens, the timing for rotating into application plays resets, and the infrastructure names regain their relative edge.