By Key Points
- Inflection AI shifts from Nvidia to Intel Gaudi 3 accelerators.
- The new platform offers 2x better price performance for enterprises.
- Intel Gaudi 3 provides affordable, high-performance AI training.
- Inflection 3.0 targets enterprise AI apps with Intel’s cloud service.
Inflection AI has made a strategic shift, choosing Intel Gaudi 3 accelerators over Nvidia GPUs to power its latest enterprise platform, Inflection 3.0.
This move allows the company to offer AI model fine-tuning services with significant improvements in both performance and cost efficiency.
With Intel Gaudi 3 at the core, Inflection AI can now provide enterprises with up to 2x price performance improvements compared to traditional AI accelerators like Nvidia’s H100.
The switch to Intel Gaudi 3 is a game-changer for Inflection AI. The startup, which originally focused on developing a conversational assistant called Pi, has pivoted toward building AI models for businesses using proprietary datasets.
With Inflection 3.0, these enterprises can fine-tune their AI models for specific applications, resulting in highly customized and efficient AI solutions.
Inflection AI Enterprise offering ditches Nvidia GPUs for Intel’s Gaudi 3 https://t.co/Ao3bLyT9qe
— Ratan Jyoti (@reach2ratan) October 7, 2024
Intel Gaudi 3’s Price-Performance Advantage
Intel Gaudi 3 brings serious firepower to the table with its cutting-edge specs. It features 128 GB of HBM2e memory, delivering a massive 3.7 Tbps of bandwidth.
The accelerator’s performance shines in dense floating-point tasks, offering 1,835 teraFLOPS for FP8 and BF16 operations.
This makes Intel Gaudi 3 not only faster but also more cost-effective for AI training and inference compared to Nvidia’s popular H100.
The affordability of Intel Gaudi 3 is one of the key reasons Inflection AI decided to make the switch.
As AI development costs rise, businesses are increasingly seeking more economical solutions. By using Intel Gaudi 3, Inflection AI ensures its customers can build and deploy AI models without overspending on hardware.
Intel is aggressively pricing Gaudi 3 to compete with Nvidia, offering an eight-accelerator system for just $125,000. This makes Intel Gaudi 3 approximately 33% cheaper than an equivalent Nvidia H100 setup.
The lower cost combined with high-performance capabilities gives businesses more flexibility to scale their AI infrastructure.
Flexibility in AI Model Deployment with Intel Gaudi 3
Another benefit of Inflection AI’s adoption of Intel Gaudi 3 is the platform’s flexibility.
Although Inflection 3.0 will run primarily on Intel’s Tiber AI Cloud, customers have the option to deploy their models on various hardware systems, both in the cloud and on-premises.
This is a crucial feature for businesses concerned about data security and infrastructure control.
Starting in Q1 2025, Inflection AI will also offer physical systems based on Intel Gaudi 3 for enterprises that prefer to manage their own data centers.
This on-premises option allows businesses to leverage the power of Intel Gaudi 3 while maintaining full control over their data and models.
While Intel has long been considered an underdog in the AI chip space, Intel Gaudi 3 is helping the company gain ground against established competitors like Nvidia and AMD.
Intel’s focus on affordability and high-performance AI solutions is resonating with enterprises that need cost-effective, scalable AI infrastructure.
Intel Gaudi 3’s Impact on the AI Market
Intel’s decision to price Gaudi 3 aggressively, combined with its technical advantages, positions it as a serious contender in the AI chip market. As more companies like Inflection AI turn to Intel Gaudi 3, Nvidia’s dominance could face significant challenges.
The promise of faster training times, more affordable AI development, and flexible deployment options make Intel Gaudi 3 a powerful solution for enterprises looking to integrate AI into their operations.
With the arrival of Nvidia’s Blackwell GPUs and AMD’s MI325X accelerators on the horizon, Intel has timed the release of Gaudi 3 well.
Its performance, pricing, and flexibility could make it a go-to option for businesses ready to invest in AI without the steep costs traditionally associated with such technology.
