Multiverse Computing
Multiverse Computing is an AI model provider leader in AI model compression. Its compressed AI models use techniques derived from quantum computing, enabling the smooth implementation of large AI language models and other systems with more than half the cost and energy consumption without sacrificing performance. With a global presence, its headquarters are located in Donostia-San Sebastián, Spain, with offices in Canada, the United Kingdom, France, Germany, Italy, and the United States. Multinational companies such as Iberdrola, Moody's Analytics, the Bank of Canada, and Telefonica have used Multiverse Computing's products. In 2025, they were named one of the “100 Most Promising Companies in AI” by CB Insights, and Digital Europe awarded them the “Future Unicorn 2024 Award.”www.multiversecomputing.com
31
MENTIONS
7
EPISODES
6
PODCASTS
Search complete. 31 mentions across 7 episodes found for "Multiverse Computing".
Sep 7, 2026
Sebastian HassingerHOST
Is quantum-inspired actually quantum, or is it just a marketing label for classical algorithms that happen to have quantum ancestry? It's a fair question, and I think the most forthright and frankly satisfying answer I've encountered comes from today's guest, who certainly has the credibility to answer it.
Sebastian HassingerHOST
Román Orús is the co-founder and chief scientific officer of Multiverse Computing, which is currently the largest quantum software company in the European Union.
Sebastian HassingerHOST
He's also an Ikerbasque Research Professor at the Donostia International Physics Center in San Sebastián, Spain.
Sebastian HassingerHOST
He completed the first PhD thesis on quantum algorithms in Spain back in two thousand and six.
Sebastian HassingerHOST
And he's the recipient of the twenty twenty-four Physics Innovation and Technology Award from the Royal Spanish Society of Physics.
Sebastian HassingerHOST
So serious academic credentials, serious commercial track record.
Sebastian HassingerHOST
Just weeks before we recorded this conversation, Multiverse announced a five hundred and seventy million dollar Series C at a one point seven billion dollar pre-money valuation.
Sebastian HassingerHOST
The company started as a quantum software shop and has evolved into one of Europe's most aggressive players in efficient AI, compressing large language models using tensor network methods, releasing open source models, and partnering with NVIDIA on a reasoning model that outperforms much larger competitors on standard benchmarks.
speaker_5HOST
Yes, which allows enterprise clients to run Anthropix models on their own internal systems, completely avoiding the privacy leaks and regulatory nightmare of sending proprietary data to the cloud.
speaker_4HOST
We also saw Multiverse Computing, a highly specialized Spanish AI startup, launch Quasar 438b.
speaker_5HOST
Which is massive.
speaker_4HOST
This model came in with an intelligence index score of 43, officially topping all European rivals and proving that massive parameter counts are no longer exclusive to Silicon Valley.
LeahHOST
correct? We did.
LeahHOST
Finally, looking at the European landscape, a Spanish AI startup named Multiverse Computing has just launched Quasar 438b.
LeahHOST
This is a significant milestone because, as their very first model, it scored a 43 on the Artificial Analysis Intelligence Index.
AminHOST
A 43 for a debut model is impressive.
speaker_2HOST
Give us an example.
speaker_5HOST
Well, on one hand, you have the Spanish startup Multiverse Computing launching Quasar 438B.
speaker_5HOST
It scored a 43 on the index, which comfortably tops their EU rivals.
speaker_2HOST
OK, so there is innovation happening.
speaker_2HOST
Give us an example.
speaker_4HOST
Well, on one hand, you have the Spanish startup Multiverse Computing launching Quasar 438B.
speaker_4HOST
It scored a 43 on the index, which comfortably tops their EU rivals.
speaker_2HOST
OK, so there is innovation happening.
![[YOU] on AI · Ahora](https://particle.news/cdn-cgi/image/format=auto,width=128/https://cdn.particle.pro/url/media/4e91d5b0-2117-58a6-bbc5-99f50507c704/4ba59d61cb1532c3192cbd071151c34c20f282a4cc8dbcff1e4711bdb0859bab)
Edo SegalHOST
Daniel nos trae la perspectiva desde los laboratorios de frontera en Europa.
DanielCORRESPONDENT
La empresa española Multiverse Computing lanzó Quasar 438b el miércoles, un modelo de razonamiento con más de cuatrocientos mil millones de parámetros que la compañía reporta como el sistema de inteligencia artificial europeo con mayor puntaje, con cuarenta y tres puntos en el Artificial Analysis Intelligence Index y sesenta y nueve en Terminal Bench de programación agéntica.
DanielCORRESPONDENT
Ya está disponible a través de la API CompactAI de la empresa, un benchmark directo poco común para un laboratorio continental a esta escala.
Edo SegalHOST
Amber, ¿qué está pasando en el frente regulatorio?
AmberCORRESPONDENT
El gobierno de Estados Unidos presentó esta semana un escrito apoyando a OpenAI en su disputa de derechos de autor con The New York Times, argumentando que el entrenamiento de modelos de inteligencia artificial con material protegido por derechos de autor podría calificar como uso legítimo bajo ciertas condiciones.