Most of what gets written about VeerNet is architecture: the encoder-decoder that lifts curves off scanned raster well logs, the loss functions, the metrics on held-out depths. That is the technical story, and it has its own whitepaper. This is a different story, about the same work seen from the side rather than the front. It is the account of how a paid research engagement for a Texas onshore operator became the substrate under a product company, and of the specific order in which that conversion had to happen. We think the order is the interesting part, because it is the part that is easy to get wrong and expensive to get wrong, and because nothing about a good model guarantees you clear it.
The claim we want to make is narrow and testable. A research consulting engagement converts into a fundable product company only after it crosses three gates, in sequence: a working pipeline, external demand proof, and an investable structure. Each gate exists to retire a particular risk, and each one is what makes the next one credible. Skip a gate, or take them out of order, and you are not early, you are stalled, because the thing the next gate depends on is not there yet. That staged logic is not ours; it is the customer-development and staged-financing view of venture formation [1] [3]. What we can add is a worked instance of it with real numbers attached.
Gate one: the delivery engagement was the pipeline
The first thing a research spin-out needs is not a pitch. It is a thing that works, and someone else paying for the work that makes it. We had both, because the pipeline was built as a delivery engagement rather than as speculative internal R&D. The engagement came in two shapes: an accelerated track of 16 weeks with 6 people at 180,000 EUR, and a standard track of 32 weeks with 4 people at 100,000 EUR. Either way, the client was paying to have their raster logs digitised, and the by-product of doing that at contract quality was the model, the training data, and the serving path that a product would later stand on.
This matters more than it sounds. Delivery revenue meant the pipeline was pressure-tested against a real operator's scans and a real deadline, not a benchmark we chose to flatter ourselves. It meant the R&D was de-risked by someone else's budget before a single equity dollar was at stake. And it meant that when we later described the technology to an investor, we were describing something in production for a paying customer, not a research demo. The gate that a science-heavy venture most often fails is exactly this one: producing good results is not the same as having a pipeline someone will pay to run, and the two are routinely confused [4]. Building the pipeline as delivery collapsed that confusion into a single funded activity.
Gate two: three signatures we did not write ourselves
A working pipeline tells you the thing can be built. It does not tell you anyone outside the room wants to buy it. That is a separate risk, and internal conviction does not retire it, because we are the worst possible judges of whether our own product is wanted. The evidence has to come from outside, and the cheapest credible form of outside evidence is a signed intent to buy before the product is finished [1].
We got three. Three Letters of Intent, from parties who were not us and not the original client, saying they would take the product when it existed. Three is not a market, and we were careful never to treat it as one. But three is the difference between "we believe operators need this" and "here are named counterparties who have said so in writing," and that difference is the entire content of gate two. It converts a demand hypothesis into a demand signal. It also does something quieter and more useful: it forces you to write down what, exactly, you are asking someone to commit to, which surfaces the gap between the delivery work you have done and the product you are claiming to sell. The LOIs were the first time our internal belief had to survive contact with someone else's signature.
Gate three: a cap table an investor could price
The third gate is the one research teams forget, because it has nothing to do with research. A working pipeline and outside demand still do not make a company you can invest in. Someone has to be able to price the thing, which means the structure has to be legible: who owns what, on what terms, for how long. In our case the load-bearing fact was that the product company issued 50 percent of its equity to the research firm on a fully-diluted basis, over a three-year term. That single line is what turned "a good model with some LOIs" into a cap table, and a cap table is what a seed investor is actually buying.
Only with that structure in place did the financing question become answerable. The ask was a 3,000,000 USD seed round against 18 months of runway. Notice that the number only means anything once the first two gates are behind it: the runway buys time to convert LOIs into contracts on a pipeline that already works, and the equity split tells the investor what their money buys a share of. Staged financing works precisely this way, releasing capital against risk that has already been retired rather than promised [3]. The seed ask is not the start of the story. It is what the three gates unlock.
What we would keep, and what we would run differently
The honest version is that we got the sequence roughly right and the pacing roughly wrong. Building the pipeline as delivery was the best decision in the whole arc, and we would do it again without hesitation, because it funded the risky part with someone else's budget and produced a production reference in the same motion. Treating three signatures as a signal rather than a market was also right, and it kept us from over-reading early interest. Where we were slower than we should have been was gate three: we treated the structure as paperwork to be done after the interesting work, when in fact it is the gate that makes the interesting work fundable, and doing it late cost us weeks we did not need to spend. The lesson we carry forward is that each gate is a test that retires a named risk [2], and structure is a risk like any other, not an afterthought to be cleaned up once the model is good.
Limitations
This is one spin-out, in one domain, backed by one engagement, and it should be read as a worked example rather than a template. The three gates and their order are the pattern we lived and the pattern the venture-formation literature describes, but the specific figures are ours and do not transfer: another team's pipeline might not be fundable as delivery work, three LOIs might be too few or beside the point in a different market, and a 50 percent equity split is a fact about our particular relationship between a research firm and its spin-out, not a recommended ratio. The 3,000,000 USD ask and 18-month runway are the numbers we put forward, not evidence about what a round should be. We also describe the sequence as we ran it, which means survivorship colours the account: we cannot show you the counterfactual where we took the gates out of order and it worked anyway, only argue from the structure why we think it would not have. And clearing three gates gets you to a fundable position; it does not get you to a funded, growing company, which depends on execution this note does not cover.
References
[1] Blank, S., and Dorf, B. The Startup Owner's Manual: The Step-by-Step Guide for Building a Great Company. K&S Ranch (2012). Outside commitment, not internal conviction, is the evidence a product is wanted, and a signed intent is a stronger signal than enthusiasm. https://www.wiley.com/en-us/The+Startup+Owner%27s+Manual%3A+The+Step+By+Step+Guide+for+Building+a+Great+Company-p-9781119690689
[2] Ries, E. The Lean Startup. Crown Business (2011). Treating each step of a new venture as a test that reduces a specific uncertainty before the next dollar is spent. https://theleanstartup.com/book
[3] Gompers, P., and Lerner, J. The Venture Capital Cycle, 2nd ed. MIT Press (2004). Investors release capital against milestones that retire risk, and structure and terms are part of what makes a round investable. https://mitpress.mit.edu/9780262572385/the-venture-capital-cycle/
[4] Pisano, G. P. Science Business: The Promise, the Reality, and the Future of Biotech. Harvard Business School Press (2006). Why science-heavy ventures are hard to fund and monetise, and why the path from research to a sellable product needs deliberate structure rather than good results alone. https://www.hbs.edu/faculty/Pages/item.aspx?num=21497
