Cerebras Programs and Perplexity AI are becoming a member of forces to problem the dominance of typical search engines like google and yahoo, asserting a partnership that guarantees to ship near-instantaneous AI-powered search outcomes at speeds beforehand thought unimaginable.
The collaboration, introduced in an unique VentureBeat report, facilities on Perplexity’s new Sonar mannequin, which runs on Cerebras’s specialised AI chips at 1,200 tokens per second — making it one of many quickest AI search programs accessible. Constructed on Meta’s Llama 3.3 70B basis, Sonar represents a major guess that customers will embrace AI-first search experiences in the event that they’re quick sufficient.
“Our partnership with Cerebras has been instrumental in bringing Sonar to life,” Denis Yarats, Perplexity’s CTO, mentioned in an announcement. “Cerebras’s cutting-edge AI inference infrastructure has enabled us to achieve unprecedented speeds and efficiency.”
AI search simply bought quicker — and large tech ought to concentrate
The timing is notable, coming simply days after Cerebras made headlines with its DeepSeek implementation, which demonstrated speeds 57 instances quicker than conventional GPU-based options. The corporate seems to be leveraging this momentum to ascertain itself because the go-to supplier for high-speed AI inference.
In keeping with Perplexity’s inner testing, Sonar outperforms each GPT-4o mini and Claude 3.5 Haiku “by a substantial margin” in person satisfaction metrics, whereas matching or exceeding dearer fashions like Claude 3.5 Sonnet. The corporate’s evaluations present Sonar attaining factuality scores of 85.1 out of 100, in comparison with 83.9 for GPT-4o and 75.8 for Claude 3.5 Sonnet.
Specialised {hardware}: The brand new battleground for AI firms
The partnership displays a rising development of AI firms in search of aggressive benefits by specialised {hardware}. Cerebras CEO Andrew Feldman just lately argued that such technological advances develop reasonably than contract the market. “Every time compute has been made less expensive, they [public market investors] have systematically assumed that made the market smaller,” Feldman instructed ZDNET in a current interview. “And in every single instance, over 50 years, it’s made the market bigger.”
Trade analysts recommend this alliance might strain conventional search suppliers and different AI firms to rethink their {hardware} methods. The power to ship near-instant outcomes might show notably compelling for enterprise prospects, the place velocity and accuracy straight impression productiveness.
Market impression: Can specialised chips reshape enterprise search?
Nevertheless, questions stay concerning the scalability and cost-effectiveness of specialised AI chips in comparison with conventional GPU-based options. Whereas Cerebras has demonstrated spectacular velocity benefits, the corporate faces the problem of convincing prospects that the efficiency advantages justify potential premium pricing.
The partnership additionally highlights the more and more aggressive panorama in AI search, the place firms are racing to distinguish themselves by velocity and accuracy reasonably than simply uncooked mannequin dimension. For Perplexity, which has been gaining consideration as an AI-native different to conventional search engines like google and yahoo, the Cerebras partnership might assist set up it as a severe contender within the enterprise search market.
Perplexity plans to make Sonar accessible to Professional customers initially, with broader availability coming quickly. The businesses didn’t disclose the monetary phrases of their partnership.
Day by day insights on enterprise use instances with VB Day by day
If you wish to impress your boss, VB Day by day has you coated. We provide the inside scoop on what firms are doing with generative AI, from regulatory shifts to sensible deployments, so you may share insights for max ROI.
An error occured.
