Written by Will Jones
In June, my co-founder Jayant and I received a signed term sheet for $11.5 million. Pre-revenue, at a valuation north of $100 million, from a group of European investors who'd spent months courting us. Then we didn't take it.
I want to tell that story properly, because it isn't really a story about money, or caution, or some principle about dilution. It's a story about what we need from the people we build with, and why that was worth walking away from eight figures to protect.
Some context first, because it matters for how much weight to put on what follows.
I didn't start out in consulting.
I trained as an engineer, spent two years designing hardware at STMicroelectronics, then co-founded an automotive-components factory and ran it for four years to a profitable exit.
In 2010 I went back to school, an MBA at Michigan's Ross School of Business, and came out the other side at McKinsey, five years across the US and Middle East offices. In 2017 I left to start EZ, which is now the execution layer for more than 200 client organizations across the UAE, India and Saudi Arabia, profitably.
Jayant joined EZ in 2019, well before AI was the story everyone in tech was telling.
In 2025 I asked him to co-found Caspr with me, and he moved to it full-time.
I define what the analysis has to do and what good work looks like; he leads the research that builds the models that do it.
I'm telling you all of this because the decision I'm about to describe only makes sense against that backdrop: we had worked together for six years before we ever wrote a line of Caspr's code.
We started raising in January.
By spring we were in serious conversations with two or three investors, and by June we had a signed term sheet from the one that looked strongest on paper, on terms we'd negotiated hard for and mostly gotten.
They were European, and part of the pitch was access, introductions to partners and connections that would help us move fast in the European market once we were ready to expand beyond the US.
That access was the part that actually mattered to us.
Jayant and I can build the product and run go-to-market ourselves, we've spent the last four months proving exactly that.
What we can't manufacture on our own is a network we don't have yet.
So when we evaluated that term sheet, we weren't only pricing the company.
We were pricing the relationship.
The relationship didn't hold up.
We started the process of closing the round, and that is when the delays began on their side.
The introductions they'd promised to the partners in their firm who could make those European connections real didn't happen at anything close to the speed they'd described, and the people we needed access to stayed one step removed.
For a company trying to move as fast as we were, that lag mattered a great deal.
So we said no.
Not because $11.5 million isn't real money, it is, and walking away from final terms at that valuation was not a decision either of us took lightly.
But by the time we were deciding, Jayant and I had already realized something else: what we were building was bigger than the round we were raising for it, and we needed a partner who could move at our speed.
This wasn't the right fit.
We didn't sit around waiting for the next offer, either.
Instead, I self-funded these last few months, and we found grant funding too, including a grant from the Indian government.
That wasn't a fallback plan.
It was a bet.
We believed, and still believe, that what we're building is going to be worth substantially more than $100 million, and that the right way to prove it was to keep building on our own terms rather than take a discount on our own work.
Those four months turned out to be the real turning point.
Jayant, about ten years younger than me, and the kind of person we've joked at EZ has Mathematics as a first language, led the technical rebuild while I drove positioning and go-to-market, and together we went fully AI-first in how we operate.
I alone have sent more than 150,000 messages to AI tools in the last ninety days, working days that started early and ran past midnight often enough that my wife has noticed.
We repositioned entirely around what we'd actually built, which we now call Analytical AI.
What that means in practice is simple to say and hard to build.
When two credible sources give different numbers for the same thing, an analyst goes underneath both: what was counted, over what period, for whom, and why it was published.
Then they decide which number answers the question in front of them.
In one report on Saudi Arabia's industrial valves market, the published estimates sat far apart until Caspr put them on the same scope, and the answer was which band to use for the question being asked.
That is what Caspr does when sources disagree on a figure that matters.
The report is finished when it arrives, and the evidence is there if you want it.
The test I hold us to is whether you could cite a Caspr report as a source, the way you cite traditional research.
Our source base went from roughly a million documents to more than 25 million curated public and proprietary sources.
The product got dramatically better, because we spent the time building it instead of spending it in fundraising meetings.
What I took from it, and what I'd pass on to anyone mid-raise, is that the number is the easy part to evaluate.
The harder question is how a partner shows up once the ink is dry. Money was never actually the thing we were short on.
Caspr is live now. Anyone can go use it, ask it a real question, and see what those months produced.
What we're looking for hasn't changed since June: partners who move at the speed this category requires, and who see what we see in what we're building.
It is greenfield AI research, models built from the ground up to do analysis, with a great deal of research behind them, and we think it will create real value for the world.
We want to build it with people who are willing to be judged, as we are, on follow-through rather than on the size of the check.
That's a higher bar than a valuation. It's the only one that matters to us now.
