Gradium Ignites Voice AI Revolution with $70M Seed Funding
January 24, 2026, 3:38 pm
Gradium, an AI startup, secured $70M in seed funding. It launches Audio Language Models (ALMs). ALMs revolutionize voice AI. They deliver natural, ultra-low latency interactions. This technology outperforms existing systems. Gradium's mission is ubiquitous, high-quality voice interfaces. They target gaming, customer care, and healthcare. The company leads European seed rounds in 2025. This signals a new era for conversational AI.
Gradium has secured a significant $70 million seed funding round. This AI startup launched today. It aims to revolutionize voice artificial intelligence. The company focuses on Audio Language Models, known as ALMs. This investment establishes Gradium as a leader in the generative audio space. It represents the largest European tech seed round of 2025.
Audio Language Models are specialized AI systems. They process, understand, and generate natural language. They use combined audio-text data. Natural language acts as a "supervision signal." This approach allows ALMs to perform voice tasks with superior effectiveness. They surpass general-purpose Large Language Models (LLMs) in this domain. ALMs are the "audio-native counterpart" to LLMs. They support more natural, expressive voice interactions. They offer dramatically lower latency. Conversations with AI feel more realistic.
ALMs are trained on specific datasets. These datasets pair audio with descriptive text. This process teaches the complex relationships between sound and language. The natural language supervision technique replaces traditional labeling methods. Natural speech guides the ALMs. It teaches them to understand and generate specific words.
Existing voice AI systems are often subpar. They prove brittle and costly. They often fail to deliver truly natural interactions. Gradium directly addresses these critical limitations. ALMs outperform LLMs across all voice AI tasks. This includes accurate speech recognition. It extends to realistic audio generation. Audio classification also benefits significantly. Gradium's mission is ambitious. It wants voice to become the primary human-machine interface.
The company seeks to transform AI assistants and agents. Conversational interactions must feel more natural. They must also be realistic. Gradium eliminates a long-standing trade-off in voice technology. It combines ultra-realistic expressivity with accurate transcription. It adds ultra-low-latency interactions. All these capabilities come at a price point suitable for widespread adoption. High-quality voice can finally achieve ubiquity.
Gradium’s team boasts exceptional talent. Researchers and engineers come from prestigious institutions. Google DeepMind alumni are key contributors. Meta's FAIR research team members are on board. Experts from Jane Street Capital also joined. CEO Neil Zeghidour assembled this formidable group. He co-founded the company in September 2025. Other founders include Olivier Teboul, Laurent Mazaré, and Alexandre Défossez. The company possesses one of the industry's highest concentrations of generative audio expertise.
Gradium maintains a crucial ongoing collaboration. It works closely with Kyutai. Kyutai is a nonprofit AI research lab. This partnership ensures Gradium’s access to frontier research. It keeps the company at the forefront of ALM innovations. The founders initially developed the core ALM concepts during their time at Kyutai.
Gradium's platform is now fully live. Its initial models are immediately accessible. They support a growing range of languages. English, French, German, Spanish, and Portuguese are currently included. More languages are under active development. The company offers flexible plans. These cater to diverse needs. They range from small developer teams to large enterprise-scale deployments.
Early adopters are already utilizing Gradium's production-ready systems. They are generating revenue. Targeted industries include gaming. Customer care operations benefit significantly. Language learning platforms are integrating the technology. Healthcare applications are emerging. AI agents see dramatic improvements in their capabilities.
Gradium's $70 million seed round stands out prominently. It represents the largest seed funding secured by any European tech company in 2025. European tech companies collectively raised over €2.4 billion in seed funding that year. This amounted to 3.3 percent of the total €72 billion invested across the sector. Seed rounds constituted 14.2 percent of all tech deals. This highlights robust early-stage investment activity in the European ecosystem.
Software and artificial intelligence sectors led industry funding. They recorded the highest deal count. They also commanded the largest round sizes. Investor interest in scalable, AI-driven businesses remains exceptionally strong. Gradium exemplifies this powerful trend. Its significant achievement underscores the growing prominence of European innovation. It positions Paris, France, as a global hub for advanced AI development.
Gradium’s launch marks a pivotal moment for voice technology. Voice AI is progressing beyond basic commands. It now aims for truly human-like dialogue. This transformation impacts how humans interact with all forms of technology. Smarter, more capable AI assistants will emerge. More intuitive applications will follow rapidly. Gradium is actively building the foundational infrastructure for this future. Its Audio Language Models promise to unlock voice AI's true, expansive potential. The path to ubiquitous, natural human-machine voice interaction is becoming clearer.
Gradium has secured a significant $70 million seed funding round. This AI startup launched today. It aims to revolutionize voice artificial intelligence. The company focuses on Audio Language Models, known as ALMs. This investment establishes Gradium as a leader in the generative audio space. It represents the largest European tech seed round of 2025.
Audio Language Models are specialized AI systems. They process, understand, and generate natural language. They use combined audio-text data. Natural language acts as a "supervision signal." This approach allows ALMs to perform voice tasks with superior effectiveness. They surpass general-purpose Large Language Models (LLMs) in this domain. ALMs are the "audio-native counterpart" to LLMs. They support more natural, expressive voice interactions. They offer dramatically lower latency. Conversations with AI feel more realistic.
ALMs are trained on specific datasets. These datasets pair audio with descriptive text. This process teaches the complex relationships between sound and language. The natural language supervision technique replaces traditional labeling methods. Natural speech guides the ALMs. It teaches them to understand and generate specific words.
Existing voice AI systems are often subpar. They prove brittle and costly. They often fail to deliver truly natural interactions. Gradium directly addresses these critical limitations. ALMs outperform LLMs across all voice AI tasks. This includes accurate speech recognition. It extends to realistic audio generation. Audio classification also benefits significantly. Gradium's mission is ambitious. It wants voice to become the primary human-machine interface.
The company seeks to transform AI assistants and agents. Conversational interactions must feel more natural. They must also be realistic. Gradium eliminates a long-standing trade-off in voice technology. It combines ultra-realistic expressivity with accurate transcription. It adds ultra-low-latency interactions. All these capabilities come at a price point suitable for widespread adoption. High-quality voice can finally achieve ubiquity.
Gradium’s team boasts exceptional talent. Researchers and engineers come from prestigious institutions. Google DeepMind alumni are key contributors. Meta's FAIR research team members are on board. Experts from Jane Street Capital also joined. CEO Neil Zeghidour assembled this formidable group. He co-founded the company in September 2025. Other founders include Olivier Teboul, Laurent Mazaré, and Alexandre Défossez. The company possesses one of the industry's highest concentrations of generative audio expertise.
Gradium maintains a crucial ongoing collaboration. It works closely with Kyutai. Kyutai is a nonprofit AI research lab. This partnership ensures Gradium’s access to frontier research. It keeps the company at the forefront of ALM innovations. The founders initially developed the core ALM concepts during their time at Kyutai.
Gradium's platform is now fully live. Its initial models are immediately accessible. They support a growing range of languages. English, French, German, Spanish, and Portuguese are currently included. More languages are under active development. The company offers flexible plans. These cater to diverse needs. They range from small developer teams to large enterprise-scale deployments.
Early adopters are already utilizing Gradium's production-ready systems. They are generating revenue. Targeted industries include gaming. Customer care operations benefit significantly. Language learning platforms are integrating the technology. Healthcare applications are emerging. AI agents see dramatic improvements in their capabilities.
Gradium's $70 million seed round stands out prominently. It represents the largest seed funding secured by any European tech company in 2025. European tech companies collectively raised over €2.4 billion in seed funding that year. This amounted to 3.3 percent of the total €72 billion invested across the sector. Seed rounds constituted 14.2 percent of all tech deals. This highlights robust early-stage investment activity in the European ecosystem.
Software and artificial intelligence sectors led industry funding. They recorded the highest deal count. They also commanded the largest round sizes. Investor interest in scalable, AI-driven businesses remains exceptionally strong. Gradium exemplifies this powerful trend. Its significant achievement underscores the growing prominence of European innovation. It positions Paris, France, as a global hub for advanced AI development.
Gradium’s launch marks a pivotal moment for voice technology. Voice AI is progressing beyond basic commands. It now aims for truly human-like dialogue. This transformation impacts how humans interact with all forms of technology. Smarter, more capable AI assistants will emerge. More intuitive applications will follow rapidly. Gradium is actively building the foundational infrastructure for this future. Its Audio Language Models promise to unlock voice AI's true, expansive potential. The path to ubiquitous, natural human-machine voice interaction is becoming clearer.