Millie Summary:
- Venture firms have become increasingly effective at creating portfolio leverage through talent, networks, customers, and operating expertise. Technology, however, is still largely managed company by company.
- Similar technology challenges often show up across a portfolio, particularly around AI, specialized engineering needs, and growing technology costs.
- The opportunity will look different for every fund, but the goal is to use portfolio scale where it can make technology investment more efficient without taking those decisions away from individual companies.
______________________________________________________________________________
Written by Nicole DerMarderosian, Growth Development Analyst
Venture firms have spent years developing ways to make a portfolio more valuable than the sum of its investments. Executive networks reduce hiring friction, commercial relationships create access to customers and partners, and operating expertise gives management teams resources that would be inefficient for every company to recreate independently.
Technology has developed differently. Most technology decisions appropriately remain with the company, particularly those tied to the product and its competitive differentiation. Yet many of the costs around those decisions recur across a portfolio. Companies often face similar technology challenges independently, whether they are evaluating AI or trying to access expertise they don’t yet have internally. Those same needs can surface elsewhere across the portfolio.
Some of these costs show up directly in operating expenses, from infrastructure costs to additional technical capacity. Others are absorbed through time and opportunity cost. A company can spend months waiting for a specialized hire while an important initiative remains on the roadmap. Product engineers can be redirected toward infrastructure or modernization work because the necessary expertise is unavailable elsewhere. An AI experiment can consume months of engineering capacity before the team determines that the original use case simply isn’t practical in production.
For a venture-backed company operating against finite runway and specific growth milestones, these are economic costs even when they never appear as separate line items. A quarter lost waiting for technical capacity can affect a product milestone or customer commitment, while engineering time spent solving an infrastructure problem is engineering time unavailable to the product roadmap. Technology decisions made to solve an immediate problem can become costly as the company grows, putting pressure on margins later.
At the company level, many of these costs arise as needs change and are difficult to avoid. At the portfolio level, repeated exposure can make them increasingly predictable, which changes the economics of how they can be addressed.
We call this the Portfolio Experimentation Tax: the cost of companies repeatedly solving technology challenges on their own when some of that cost could be reduced at the portfolio level. A single company may have limited leverage over its technology spend. Across a portfolio, the economics and visibility into that spend can look very different. Experience can compound in a similar way. When the same needs continue to surface, companies should not always have to start from the same place.
Learning Across the Portfolio
AI provides a useful example because companies are experimenting quickly while the path from a promising use case to a valuable production system remains uneven.
Technology leaders expect a median of 50 percent of AI pilots to convert to production. Some experiments should fail because that is part of determining where the technology creates value. For a venture portfolio, the more consequential question is how much time and engineering capacity each company needs to spend reaching that conclusion.
The economics become meaningful well before model or infrastructure costs do. An AI initiative can consume months of time across the business while competing with other priorities. Understanding whether the use case has the data, economics, and path to production it needs can help management decide earlier where deeper investment makes sense.
This is where portfolio experience starts to matter. Each AI opportunity still needs to be evaluated against the company’s own product, data, and customers, but companies do not have to approach those decisions from scratch. Giving them access to technical guidance and pressure testing early can help them benefit from what has already been learned across similar initiatives, rather than discovering the same avoidable problems independently. The goal is to identify likely constraints and understand what needs to be true before significant engineering time is committed.
The economic benefit is better allocation of capital and engineering time. A weak initiative can be identified earlier, while a strong one can move forward with a clearer understanding of what production will require. The solution remains company-specific, but the learning curve does not have to reset with every investment.
The Economics of Engineering Capacity
Engineering capacity presents a different problem. Venture-backed companies need to own the technical capabilities that create lasting differentiation, but the need for specialized expertise often develops faster than the organization can build it internally.
That gap remains significant. 47 percent of organizations report being understaffed in AI and ML engineering, with similar gaps across cybersecurity, FinOps, platform engineering, and cloud computing. For a growing company, the challenge is not necessarily deciding whether that expertise is valuable. It is deciding when there is enough sustained demand to justify adding it permanently.
In the meantime, waiting has a cost. A company may need specialized engineering expertise to complete a modernization, deliver an important product initiative, or meet a customer commitment. If that expertise is unavailable, the work can move back a quarter while the company hires, or the existing engineering team can absorb it at the expense of something else on the roadmap. Neither cost necessarily appears in the hiring budget, but both affect how quickly the business can execute.
At the company level, these needs can be difficult to predict. A business may need deep cloud expertise for six months and much less of it once a modernization is complete. Another may need additional engineering capacity around a major product release. Hiring permanently for every peak in demand can add fixed cost that remains long after the immediate need has changed.
Across a portfolio, however, those individual peaks can start to look more like recurring demand. One company may need specialized expertise today and another six months from now. No single company needs to maintain every capability internally, but the portfolio can create more consistent access to those capabilities when they are needed.
This creates an opportunity to make specialized engineering capacity available across the portfolio without centralizing technology decisions. Companies can still decide what they build, how they build it, and which capabilities should ultimately remain in-house. Portfolio scale simply gives them another option when an important initiative cannot afford to wait for the organization to catch up.
The economic case is not that flexible capacity should replace permanent engineering teams. It gives companies another option between delaying an important initiative and permanently hiring for expertise they may only need at certain points in their growth. Across a portfolio, that gap can be addressed more efficiently.
Technology Economics at Portfolio Scale
Cloud and AI spend make the portfolio opportunity particularly visible because the costs can be measured directly. The 2026 State of FinOps found that 98 percent of respondents now manage AI spend, up from 63 percent in 2025. As more products depend on cloud infrastructure, data, and AI, the technology required to deliver them can have a growing impact on margins.
At the company level, venture firms can give founders access to tools such as FinOps reviews and AI assessments that help them understand those costs earlier. A FinOps review can identify unnecessary cloud spend and where infrastructure costs are likely to grow, while an AI assessment can help determine whether the economics of a use case still make sense as usage scales. For growing software companies, this gives founders a chance to control costs before they become harder to change or start putting pressure on margins.
Several companies may also be paying the same cloud provider, using similar data platforms, or buying overlapping software without any reason to look at that spend together. Looking across the portfolio can reveal where those costs overlap and where there may be opportunities to consolidate or negotiate better terms.
A single portfolio company may have limited negotiating power with a major technology vendor. Several companies buying similar services represent a different opportunity. Where technology spend overlaps, the size of the portfolio can create leverage that an individual company may not have on its own.
The companies still decide what technology they need and how they build. The fund can help founders better manage technology costs inside their companies while also using portfolio scale where it creates an economic advantage.
Where Portfolio Leverage Makes Sense
There is no standard technology strategy that should be applied across venture portfolios. The opportunity depends on the companies, their stage of growth, and where technology is creating the most pressure.
For one fund, the need may be helping founders make better decisions around AI before committing significant engineering time. For another, it may be giving companies access to specialized engineering expertise when an important initiative cannot wait for a new hire. Growing software companies may need help understanding cloud and AI costs as usage increases, while earlier-stage founders may benefit from senior technology guidance before they are ready to build out that leadership internally.
The point is to give companies access to the right support when they need it, while recognizing when the same needs are showing up elsewhere in the portfolio.
This is increasingly how we think about our work at MILL5. We support growing companies with AI assessments, FinOps, fractional CTO leadership, and software, cloud, data, and AI engineering. For venture firms, that creates another way to support founders as they grow, while building on the experience gained across the portfolio.
Venture capitalists already create this kind of advantage through talent, networks, customers, and operating expertise. Technology can work the same way, while companies remain independent in the technology that differentiates them.
To get started with a complimentary strategy session, contact Nicole DerMarderosian and the MILL5 team at nicoled@mill5.com.
