Most venture firms describe their focus in a sentence. Ours takes 44 of them. Every sector we invest in is a vertical, with its own team, two funds (Fund I, then Fund II) and written boundaries with its neighbors. This note explains how we got there, and the single question that runs through all of it.
Why verticals
Venture capital works best when the people deciding know the industry cold. A team that knows a vertical’s buyers, its engineers and its failure modes should find the right company earlier, and help it further, than a generalist can. Specialization also makes us easier to work with: a founder can see in a minute whether we invest in what they build.
Place matters too. Each vertical is run from the office whose industry fits it best: our headquarters verticals in Dallas, defense and government in Washington, D.C., fintech and health in Atlanta, energy and space in Houston, housing and robotics in Salt Lake City, and quantum and biotech in San Francisco.
How we chose them
We began with twelve hardware verticals, the hardware under AI and industry, chosen by comparing every hardware sector on four measures from public-market data: valuation multiples, volatility, revenue growth and capital intensity (source: Aswath Damodaran, NYU Stern datasets, January 2026).
The twelve are Grid & Power Equipment Tech, AgTech, Sensing & Measurement Tech and Telecom & Networking Tech in Dallas; Advanced Materials, Cooling & Thermal, Energy Production, Critical Minerals, Water and Flow Control Tech in Houston; and Power Semiconductor Tech and Semiconductor Equipment Tech in San Francisco.
We then added 32 more verticals across AI and software, life sciences, defense and government, energy and climate, real estate technology and consumer markets, and gave each its own team and two funds. The method is described in detail to prospective investors in our confidential materials. You can see all 44 verticals on our site.
Boundaries, in writing
Each company belongs to one vertical, based on what it makes. Our boundaries are written down, including what does not fit. A few examples from the hardware verticals:
- A solid-state transformer belongs in Grid & Power Equipment Tech. Software that optimizes the grid belongs in Energy Tech.
- A liquid cooling system belongs in Cooling & Thermal Tech. The power conversion that feeds the rack belongs in Grid & Power Equipment Tech, and the GaN or SiC devices inside it belong in Power Semiconductor Tech.
- An industrial pump or valve belongs in Flow Control Tech, unless it is part of a water treatment system (Water Tech) or downhole and wellhead equipment (Energy Production Tech).
When a company could fit two verticals, our Chief Compliance Officer decides which vertical it belongs to and records why. That keeps two of our funds from competing for the same first investment.
The one test
Inside every vertical, we apply the same seven written tests: vertical fit, technical proof, a measurable gain, customer signal, a capital path, a team and strategic buyers. Underneath all seven sits one question: will someone buy it?
A named customer matters more than a large market.
Three of the tests ask that question directly:
- Customer signal. A pilot, letter of intent or design partnership with a named customer. A market-size slide does not tell us whether one buyer will sign; a named customer does.
- A measurable gain over what customers use today. Buyers rarely switch for “better.” They switch for a number they can defend inside their own organization.
- Strategic buyers. At least three plausible acquirers we can name at the start. Large industrial and semiconductor companies buy technology they need. Market examples include Eaton’s acquisitions of Boyd Thermal and Fibrebond, and Vertiv’s acquisition of UtilityInnovation Group. These are market examples only; none is a Dark Ice Ventures investment.
The question also shapes the capital test. Hardware failures are expensive. Graphcore sold for about $400 million after raising about $700 million; Northvolt went bankrupt after raising more than $15 billion. In our hardware verticals, we invest only where the path to first revenue needs less than about $50 million of total funding.
Patient, where it counts
Not every important company is close to revenue. A vertical fund may carry platform programs inside its lane, such as Frontier, for technology that needs new physics or materials and may be more than about seven years from revenue, and Dreamers, $150,000 at pre-seed for first-time founders. The question still applies; it simply has a longer horizon.
What sits outside
Control buyouts, and buying, building or running real estate and real assets, belong to Dark Ice Capital and its eight platform companies. If you build there, we can introduce you to the right Dark Ice team, with your consent.
If you are building inside one of our verticals, tell us what you are building and what it replaces. We answer on fit within five business days.
Dark Ice Ventures · Oct 7, 2026
This note is for information only and is not an offer to sell or a solicitation of an offer to buy any security or fund interest.
