Own the P&L (Or Get Shut Down When the CEO Changes)

Many corporate innovation units don't make it past three years. They launch with a mandate that sounds durable — pilots, capability building, strategic optionality — until the executive who signed off on it leaves, and the next one reads that same mandate completely differently.

LG Nova was built specifically to not depend on that read: ventures deliberately outside LG's core hardware business, a ring-fenced budget, external investors with their own fund, and a 12-month incubation process that requires commercial evidence at every gate. Four ventures have launched so far, spanning rural healthcare to data center energy to mental health to AI marketing, each with its own P&L.

Below, we look at LG Nova’s full model: how the unit was designed, how the incubation funnel works, what the four ventures are actually doing in market, and why AI compression is raising the bar for every corporate innovation team still justifying itself on soft metrics.

Innovation units operate on a cycle that has more to do with leadership tenure than with the quality of the work. A CEO stands up a unit, gives it a mandate, and for two or three years the unit runs. Then the CEO departs or the board's priorities shift, and the new leadership asks a question no one can answer cleanly: what exactly is the unit worth?

Units justify themselves through indirect value: technology scouting, cultural signaling, ecosystem access. While all of those contributions are real, they are also subject to interpretation. A technology scouting function that one leader considers essential infrastructure looks like overhead to the next. The PR value of having an innovation lab matters to one CEO; the CEO who follows next wants to know why the lab hasn't produced revenue.

Sokwoo experienced this dynamic firsthand when he joined LG. The company had spent 70 years building DNA optimized for hardware manufacturing. Previous attempts to build new kinds of businesses inside the existing structure kept running into the same wall: the thought process, the talent, the distribution channels, everything was calibrated for factories and appliances.

But when Sokwoo looked deeper at what had survived leadership transitions at LG, one thing stayed with him.

Sokwoo doesn't dismiss indirect benefits. He points out that those benefits will be "analyzed and interpreted very differently depending on who you are." Capability building looks strategic to one CEO and discretionary to the next. Revenue doesn't have that problem. The P&L argument is simple: indirect value won't keep the unit alive when leadership turns over, because the new leader will read it differently. Revenue, on the contrary, reads the same to everyone.

When Sokwoo assessed what LG had to work with for building non-hardware businesses, the answer was: very little. The company's expertise was in hardware manufacturing. Its distribution channels were optimized for physical products. Its talent base had been shaped by seven decades of building appliances and electronics. None of that infrastructure could support the kind of asset-light, software-driven businesses LG wanted to create.

So he designed LG Nova to operate at a deliberate distance from the core business. The unit sits in Silicon Valley, separated from LG's Korean headquarters by geography and by mandate. The ventures LG Nova builds sit under LG's broader ABC strategy (Artificial Intelligence, BioTech, CleanTech) and have nothing to do with LG's existing product lines. The budget is completely ring-fenced from the business units.

That distance creates a specific dynamic inside the organization. Business units have little interest in what LG Nova builds, because the ventures sit outside their domain. An appliance division has no stake in a mental health platform, and a TV division has no reason to follow data center energy software. As a result, LG Nova doesn't receive much internal support, but there’s a catch: the business units don't interfere either. No one tries to redirect a healthcare venture to serve the TV division's roadmap or pulls engineering talent back toward the factory floor.

The ring-fenced budget carries its own pressure. LG Nova's funding doesn't come from business unit allocations, which insulates the unit from internal budget politics. The funding also comes with a timeline: these ventures need to reach billion-dollar scale within a defined window. Sokwoo owns P&L responsibility directly, with expectations that are more patient than a quarterly earnings cycle and tied to concrete commercial milestones. "I'll give you money, and with probably a bit less stringent timeline," is how the mandate was framed. The endpoint however was clear: the ventures need to make money.

The design also shaped the kind of talent LG Nova needed. Sokwoo describes two distinct profiles inside the unit. One group operates more like a traditional innovation team: generating ideas, running experiments, doing quick validation work. The second group is different entirely. Their job is building businesses: sales, operations, legal structure, customer acquisition. The second group's work doesn't look like innovation at all — it looks more like running a company. Both functions sit inside LG Nova, and keeping them together means the unit doesn't hand off a validated concept to someone else and lose control of the outcome.

LG has 100,000 employees, and the number who share his perspective is, by his account, very small. That isolation is part of why peer networks with innovation leaders at other companies matter as much as any internal relationship. The P&L eventually provides its own constituency, because revenue is hard to argue with. However, until the ventures produce that revenue at scale, the innovation leader holds the position largely alone.

LG Nova runs a six-stage incubation process that moves from exploration through to Series A funding and launch. Each stage has a defined timeline and budget:

The full cycle runs about 12 months from concept to launch, with commercial evidence required at every gate. That cadence produces a funnel: roughly 25 ideas enter per year, about 10 reach proof of concept, 5 become commercial MVPs, and 2 to 3 spin out as standalone ventures.

The process kills weak concepts early and concentrates resources on the ones that clear each gate with real commercial evidence. Stage 5 (revenue readiness) arrives at roughly the six-month mark, preserving enough runway for ventures to generate meaningful revenue before NovaWave Capital (LG Nova's partner fund) makes its investment decisions.

LG Nova's external orientation goes beyond typical corporate partnerships. Sokwoo's assessment when he started was that LG had virtually no internal assets useful for the kind of businesses the unit needed to build. The expertise, the distribution channels, the talent base: all of that was optimized for hardware. Building new, asset-light ventures required capabilities LG simply didn't possess.

So he built an ecosystem from scratch: startups, large corporate partners, potential customers, state governments, and venture capitalists. Each relationship was structured around mutual value. The ecosystem works because everyone in the network has a clear reason to stay. LG Nova's Mission for the Future program for example, which sources startup partners, received over 4,000 submissions from global startups in its first three years, with more than 100 selected to explore joint business opportunities. The ventures that eventually launch are built through collaboration with these external founders and domain specialists who bring capabilities LG doesn't have internally.

The funding side of the ecosystem runs through NovaWave Capital, a venture fund created to support LG Nova's portfolio companies. LG Electronics is the anchor limited partner. Outside investors, including AZ Venture Capital and others, have come in alongside them. NovaWave is managed by Aurion Capital, a global investment group, and focuses on scaling high-growth AI companies in energy, health, and business sectors. The fund led PADO's $6 million seed round in March 2026 and has expanded into a cross-regional network with hubs in Silicon Valley, West Virginia, and Arizona, where it launched the WaveX AI venture studio in partnership with the Arizona Commerce Authority at CES 2026.

The fund structure brings external capital into the operation, which reduces LG's financial exposure on any single venture. It also introduces investors who care about returns, and those investors hold the ventures to market standards. A venture that might coast inside a corporation on leadership enthusiasm or strategic alignment with a future roadmap has to show real commercial traction when outside money is involved.

Sokwoo draws a sharp line between this model and open innovation as a general practice. "Open innovation without a goal is just open innovation exercise," he says. Companies bring in startups, run accelerators, host demo days. The activity can look productive, but unless every external relationship is tied to a specific business outcome, the ecosystem produces motion without progress. In LG Nova's case, the goal is the same one that keeps the unit alive: building revenue-generating ventures with their own P&L that LG owns.

Primefocus Health, the first venture out of LG Nova, launched in 2024 under CEO Darren Sabo, who previously led commercialization of health ventures inside the unit.

Rural patients managing chronic conditions like diabetes and hypertension need ongoing clinical support, and the nearest provider may be a long drive away. Primefocus builds a modular platform that lets healthcare providers extend care from the hospital to the home: remote monitoring, AI-driven insights, personalized care plans, and an app-based patient portal where providers can track progress, communicate, and intervene between visits.

The platform's first major deployment is with Marshall Health Network, a not-for-profit academic health system serving 1.4 million residents across 40 counties in West Virginia, southern Ohio, and eastern Kentucky. The initial focus was obesity: a 250-patient pilot through Marshall's Obesity Clinic provided ongoing clinical support to rural patients without requiring frequent travel to clinical facilities. Early results showed improved patient engagement and better management of chronic conditions. At CES 2026, Primefocus and Marshall announced an expansion into breast cancer prehabilitation and rehabilitation support through Marshall's Edwards Cancer Institute.

The venture has also begun work on substance use disorder and healthy aging programs. Primefocus is the only LG Nova venture with what could be described as a natural adjacency to LG's consumer brand: the idea that LG is already trusted inside people's homes, and healthcare is an extension of that trust. Every other venture operates in a market where LG has no existing presence at all.

PADO launched in 2024 under CEO Wannie Park, a veteran of the energy technology sector with previous roles at companies that were acquired by Shell and Mysa.

The problem PADO addresses is the gap between how much computing power a data center could run and how much it actually runs. Mid-market colocation facilities (typically under 100 megawatts) promise customers near-perfect uptime, and to protect against outages, they tend to keep GPU utilization at around 30% to 40% of capacity. That conservative approach means significant computing power sits unused.

PADO's platform uses AI and machine learning to orchestrate the relationship between compute workloads, cooling systems, power infrastructure, and distributed energy resources. The company's core metric is Compute Per Megawatt: how much processing a facility can extract from each megawatt of electricity. PADO's thesis is that its orchestration can raise GPU utilization from the 30-40% range to 55-60%, and its tools can reveal 30-60% of unused capacity in legacy facilities where conservative planning has left throughput on the table. The potential result is 20-40% throughput gains within existing physical footprints, which means more revenue for the data center operator without building new facilities.

In March 2026, PADO closed a $6 million seed round led by NovaWave Capital. The company has signed partnerships with MARA, a publicly traded energy technology company, to develop power load balancing as a service, and with VESSL, an MLOps platform, to build what the two companies call the industry's first energy-oriented MLOps workflow: a system that ties machine learning workloads to energy economics and renewable availability. PADO is also participating in EPRI's DCFlex initiative on engineering grid-friendly data centers.

The data center power management market sits at roughly $22 billion (2024 estimate) and is projected to reach $40 billion by 2032. PADO's positioning targets mid-market operators underserved by hardware-heavy solutions, where software-based optimization can unlock stranded capacity without major capital expenditure.

ReliefAI Health, launched October 2025 under CEO Atul Singh (25 years in digital healthcare, including Teva, AbbVie, and Walgreens), tackles a specific number: 44% of therapy clients drop out within the first four weeks. Its app fills the gap between sessions — mood tracking, voice journals, CBT goal-setting — and feeds clinicians insights they can bill through CMS's Remote Therapeutic Monitoring codes, turning a clinical tool into a direct revenue stream. Its sharpest differentiator: unlike most digital mental health tools, which build their own separate provider networks, ReliefAI integrates into the clinic the patient already attends.

OnVibe, LG Nova's newest venture, launched in February 2026 under CEO Sahar Arshad and unveiled at CES 2026. It's an AI-powered marketing partner for small businesses that can't afford an agency — daily content recommendations based on a brand's performance and category trends, rather than generic content generation. It's also the earliest-stage of the four: five months old, still building its initial user base, with the commercial evidence the other three ventures already have still ahead of it.

All four ventures are AI-first, all launched within the last two years, and none existed inside LG's product lines before LG Nova built them. What's changed since the first one launched is how fast the market expects them to show revenue.

Two years ago, LG Nova's product development cycle followed a familiar corporate timeline. Conceive an idea, spend roughly a year building the product, assemble a sales team, go to market. Investors and internal stakeholders expected a two-to-three-year horizon before seeing returns. Today, that timeline has compressed dramatically. LG Nova's product development cycle for software-based ventures has gone from six months to roughly four weeks.

The compression is already standard in the startup world. Investors increasingly expect founders to have revenue before funding, because building a software product costs so little now that investors don't see a reason to fund anything pre-revenue. A product that took six months to build two years ago can now be assembled in hours. Distribution has shifted alongside production: some companies generate tens of millions in revenue through community-driven growth without a single salesperson.

The gap between AI adoption and AI results is real. According to McKinsey and BCG, 88% of companies are using AI in at least one function, and only 5% are generating value at scale. 60% report little to no material impact. Meanwhile, AI-native startups like Cursor ($500 million in annual recurring revenue with roughly 20 people in about three years), Midjourney ($200 million ARR with roughly 10 people in two years), and ElevenLabs ($100 million ARR with roughly 50 people in two years) are producing revenue-per-employee ratios 6 to 17 times higher than traditional SaaS companies. "You don't need money to build products anymore," Sokwoo says. "You build it and find somebody who's going to buy it. Then, people invest."

For corporate innovation units, the compression raises the bar on what a unit must demonstrate and how quickly the demonstration has to happen. The old cadence (a year of building, a year of piloting, a year of scaling) is already out of step with what external markets consider normal. An innovation unit that takes 18 months to launch a product is competing against startups that launch in weeks.

This compression makes the P&L argument more urgent. When product cycles were long, an innovation unit could justify years of investment before showing returns. Capability building and strategic positioning were reasonable interim metrics, because everyone understood that building took time.

That justification erodes when the market knows products can be built in weeks. The CEO who might have given the unit three years of runway now asks the question after one: where's the revenue?

LG Nova is five years in. Primefocus Health is working with a health network serving 1.4 million people across rural Appalachia. PADO has closed a $6 million seed round. ReliefAI is deployed in behavioral health clinics with a reimbursable revenue model. OnVibe launched commercially five months ago. The ventures are still early, but the P&L is building.

Every innovation leader eventually faces the same moment: the CEO changes, the strategy shifts, and someone new asks what the unit is worth. The leaders who lose that moment are the ones still answering with pilots, cultural impact, or strategic optionality. The leaders who survive it are the ones who can point to revenue.

That has always been true - what's changed is how quickly the moment arrives. When product cycles ran in years, an innovation unit had time to build credibility before the question came. When they run in weeks, the question comes much sooner.

That's the lesson from LG Nova's first five years. The ring-fenced budget, the external fund, the ventures outside the core, the 12-month incubation process: all of it exists to produce one thing, a P&L that speaks for itself when the leadership changes.

The innovation leaders who are already thinking about their own version of this are looking at their portfolios and asking a simple question: which revenue line can I build now, while the current CEO is still in the chair and the window is still open, that would make this unit too expensive to shut down?