Founded by Francesco Pappone, Giulio Starace, and Emanuele Rodolà, Paradigma wants to use its intelligenza artificiale to assist or automate the work of scientific research. The company was selected among ten startups for the Next Frontier AI Challenge program dedicated to developing frontier AI technology. SPRIND, the German federal agency for innovation, promotes this initiative and provides a total budget of €125 million, distributed across three competitive phases.

The selection does not immediately assign the full potential funding amount. Each of the ten startups will initially receive €3 million each, while subsequent sums depend on surpassing two additional evaluations. The maximum value of €26.5 million indicated for each project is cumulative and concerns only those that reach the final phase.

For Paradigma, the result achieved is entering the initial group with the first tranche predicted by the program. According to the Corriere della Sera, the Next Frontier AI Challenge seeks to accelerate Europe's development of advanced systems of artificial intelligence and reduce the delay compared to the United States. The structure of successive selections concentrates resources on projects remaining in competition.

The program combines financial support and performance verification, but the source does not specify the calendar of evaluations nor the criteria used to select the companies destined to proceed. Paradigma is an Italian startup born in Rome just a few months ago. Its project aims at building systems capable of assisting or automating the work of scientific research.

The ambition described by the source pertains to the entire process: from the initial formulation of hypotheses until experimental validation. Artificial intelligence would be employed as an operative tool throughout multiple stages of research, rather than being limited to generating texts or responses. The company was founded by Francesco Pappone, already a doctorate holder in Machine Learning, Giulio Starace, former OpenAI researcher, and Emanuele Rodolà, professor ordinary of Informatics at the University Sapienza of Rome.

The founders' paths connect academic research, automatic learning, and experience in developing artificial intelligence technology. The source does not provide further details on the societarian composition, current staff, or capital raised before accessing the competition. This is the first tranche reserved for all selected startups and constitutes the economic base to tackle subsequent phases.

However, this total budget also includes resources allocated to startups that will surpass the selections. Not all participants receive the same final amount. After the first phase, the number of projects remaining in competition decreases from ten to six.

Each of these six startups can initially obtain another €8 million. However, the possibility of entering the second tranche remains subordinate to selection: the source does not specify whether these funds are already assigned to the Italian startup. In the last phase, three realities among the six still competing will be identified.

To each could be recognized an additional sum of €15.5 million. Summing up the three tranches — 3 million, 8 million, and 15.5 million — the support for a project capable of completing the entire process would reach €26.5 million. The mentioned value represents therefore a maximum limit, not the funding already received by every finalist.

For Paradigma, participation offers resources to measure a particularly complex project: transferring intelligenza artificiale within a sequence of activities comprising hypothesis formulation and experimental validation. The source does not indicate which scientific disciplines will be addressed first, what technical results are already available or what degree of autonomy has been reached by systems. No declarations from the founders nor independent evaluations on proposed technologies are reported.

The next stage will be reducing finalists from ten to six, followed by the conclusive choice of three projects. Remaining details about times, evaluation parameters, and requested results for accessing new tranches remain unknown.