Technical progress matters only when it connects to a workflow someone needs to operate.
During the spring of 2026, Capicú completed more than 130 customer-discovery interviews while participating in I-Corps Puerto Rico Cohort 21 and BioLeap Gen. 3. In June, that work continued in New York during Tech Week through conversations with operators, scientists, founders, and builders across the biotechnology ecosystem.
The purpose was concrete: test where biological data becomes difficult to act on, understand what teams require from AI near the instrument, and identify the deployment constraints that cannot be solved by a model alone.
More than an interview count
I-Corps Puerto Rico applies the National Science Foundation's customer-discovery methodology to early-stage technical ventures. Cohort 21 ran through April and May with a demanding interview cadence and direct pressure on the assumptions behind each participating company.
Capicú completed more than 130 interviews across the broader discovery effort.
The number is useful because it required repetition across roles and settings. The value came from tracing where answers converged, where requirements conflicted, and where a technically attractive idea did not fit the way scientific work is actually performed.
Three constraints repeatedly shaped the work:
- Biological evidence is distributed across instruments, files, methods, and people.
- A useful result has to arrive while the sample or process can still be acted upon.
- New intelligence has to fit existing infrastructure without creating another disconnected system for scientific teams to maintain.
Those constraints now sit at the center of how Capicú develops Foundation Models and deploys AI into biological instrumentation.
Taking the questions to New York
New York Tech Week ran from June 1 through June 7. Capicú used the week to continue discussions around laboratory workflows, model qualification, and the role of instrument-level inference.
Conversations around SOSV and the Advanced Science Research Center at the CUNY Graduate Center brought those questions into a wider group of biotechnology operators and technical builders. The objective was not event attendance. It was to pressure-test where CEMI Nodes belong on a laboratory bench and what a quality-control scientist needs from a model before relying on its output.
Non-dilutive support for disciplined development
That discovery and product-development work was supported by $60,000 in non-dilutive funding:
- $5,000 through VentureWell E-Team Pioneer;
- an additional $20,000 through VentureWell E-Team Propel;
- $20,000 through BioLeap Gen. 3; and
- a $15,000 Apostando en Mayagüez seed-capital award from Perspectivas Globales, a program of OverComing Adversities.
At the BioLeap Gen. 3 Showcase, Capicú also received the Innovation Protection Prize from Ydea Law Group: five hours of legal advisory support valued at $1,500. That support is in-kind and separate from the $60,000 funding total.
The programs brought different forms of accountability to the same work: interview evidence, technical milestones, operating plans, and a clearer path from a scientific problem to a deployable system.
Building from Puerto Rico
Capicú is building from Mayagüez, within reach of one of the most concentrated biopharmaceutical manufacturing regions in the world. The combination of engineering talent, operating facilities, research institutions, and an increasingly connected startup ecosystem creates a practical place to build infrastructure for biological AI.
The spring was a period of sustained discovery, not a collection of isolated programs. More than 130 interviews, work across Puerto Rico and New York, and non-dilutive support gave the team a stronger evidence base for deciding what to build and how it has to operate.
The work continues with the same standard: assumptions should be tested before they become infrastructure.