Free Intensive Nov 21–22Free 2-day hands-on Vibecoding Intensive on Nov 21–22 at 16:00 MSK (Sat–Sun) · Live Cohort 3 starts Nov 23Join Free →Cohort 3 (Nov 23)
InfrastructureInsight · On RadarTNW→ original

ElevenLabs and Google Cloud bet on NVIDIA Blackwell GPUs

ElevenLabs has signed a multiyear agreement with Google Cloud to expand their infrastructure partnership. The company will get access to NVIDIA Blackwell-class GPUs to scale its products, voice synthesis and conversational AI agents, for large corporate clients. The deal highlights growing demand for compute in voice AI and intensifying competition among cloud providers for leading AI startups.

AI-processed from TNW; edited by Hamidun News
ElevenLabs and Google Cloud bet on NVIDIA Blackwell GPUs
Source: TNW. Collage: Hamidun News.
Listen to article

The race for voice AI is entering a new phase, and the key players are redrawing their alliances. On February 26, London-based ElevenLabs, one of the most prominent voice AI startups, announced an expanded strategic partnership with Google Cloud. Under the renewed multiyear agreement, ElevenLabs gets access to Google's expanded cloud infrastructure and, most importantly, to NVIDIA's latest Blackwell-class GPUs. This is not just another corporate deal: it signals where the voice technology industry is heading.

To understand why the partnership matters, it helps to look at ElevenLabs' trajectory. The company was founded in 2022 by former Google and Palantir employees and within a few years became one of the market leaders in voice synthesis. Its technology can generate realistic speech in dozens of languages, clone voices and build conversational AI agents capable of natural dialogue. Its customers include media companies, publishers, game developers and large corporations. But scaling for the enterprise segment became the bottleneck: real-time voice processing demands enormous compute, and standard cloud capacity is no longer enough.

This is where NVIDIA Blackwell GPUs come in. The next-generation architecture succeeds Hopper. Blackwell chips, including the flagship B200, promise a multiple-fold performance gain in inference, which is exactly what real-time speech generation needs. Training a model can be done once, over a span of weeks, but inference is continuous: it runs every time a user asks for text to be voiced or talks to an AI agent. The faster and more efficient the inference, the lower the latency, the higher the quality and the lower the cost per request. For a company handling millions of voice requests every day, this is critical.

Choosing Google Cloud as an infrastructure partner is not accidental either. The three cloud giants, AWS, Azure and Google Cloud, are fighting hard for AI startups, and each such contract becomes a reputational asset. Over the past year and a half, Google Cloud has aggressively built up its AI portfolio: its own TPU chips, Gemini integration and specialized AI instances. Landing ElevenLabs strengthens Google's position in generative AI and shows that the platform can handle demanding voice synthesis workloads at a level that satisfies one of the market leaders.

For ElevenLabs itself, the partnership opens the door to enterprise customers on a completely different scale. Large organizations considering voice AI agents for customer service, internal communications or product interfaces set strict requirements for reliability, response speed and the system's ability to handle peak loads. Access to Google Cloud infrastructure with Blackwell accelerators lets ElevenLabs guarantee a level of service that was previously available only to giants such as Amazon or Microsoft with their own voice platforms.

The broader context deserves attention too. The voice AI market is growing explosively: by various estimates, its size will exceed $30 billion by 2027. Competition is heating up: OpenAI is developing ChatGPT's voice capabilities, Amazon is improving Alexa, and dozens of startups offer specialized solutions for dubbing, podcasts and contact centers. In these conditions, access to cutting-edge hardware is not just a competitive advantage but a condition of survival. Companies that are first to master Blackwell for inference will be able to offer faster, better and cheaper service, and capture the lion's share of a growing market.

There is one more aspect that cannot be ignored. AI startups' deepening dependence on cloud giants creates an asymmetry of power. Google Cloud gets not only a customer but also a deep understanding of how advanced voice models are built, what loads they generate and where the technology is heading. For ElevenLabs, this means balancing the benefits of the partnership against the risks of lock-in to a single provider. Still, the wording "expanded partnership" suggests the company had worked with Google Cloud before, so the choice was deliberate and time-tested.

Ultimately, the ElevenLabs and Google Cloud announcement is less news about a single deal than an indicator of the voice AI market's maturity. The technology has moved from experimentation to industrial-scale deployment, and the deciding factor is no longer just model quality but the ability to run it on infrastructure that can withstand the load of a real business. Whoever controls compute controls the future of voice intelligence.

Story
ZK
Zhemal Khamidun — CPO at Alpina Digital, product manager of AlpinaGPT. Building AI products since 2019. Hamidun News is a pet project that grew into a production aggregator with 6,000+ readers.
✦ Unified AI Workspace · Nomad

Test Frontier LLMs & Autonomous Agents in One Workspace

Run Claude Opus 5.5, GPT-6, Gemini 3.1, and DeepSeek side-by-side in Nomad AI (nomadnet.ai) — unified model routing, deep research, and multimodal workflows under one subscription.

Try Nomad AI →Launch Nomad Agent
What do you think?
Loading comments…