Intel announced a major expansion of its artificial intelligence offerings, unveiling multiple new processors optimized for AI workloads across PCs, servers and the cloud in an attempt to catch up to rivals Nvidia and AMD.
New PC and Laptop Chips Bring AI Capabilities to Consumers
The centerpiece of Intel’s launches is the new 13th Gen Intel Core desktop processors, code-named “Meteor Lake,” which are the first x86 CPU to have a dedicated neural processing unit integrated on die. The neural processing unit provides accelerated performance for AI-enhanced creation apps and tasks.
These new Core processors, branded as Core i9-13980HX and Core i7-13620H, deliver up to 5.6 GHz Turbo frequency and feature Intel’s fastest integrated graphics. Major PC manufacturers like Dell, HP and Lenovo announced upcoming laptops taking advantage of the new AI-focused chips.
|Max Turbo Frequency
|Intel Arc A770M
|Intel Arc A730M
In benchmarks, the integrated GPU performed better than AMD’s latest Phoenix graphics in both gaming and content creation workloads. The addition of a dedicated AI engine also provided substantial performance improvements in AI-enhanced workflows compared to prior generations.
New Xeon Server Chips Target AI in Data Centers
On the data center side, Intel launched its 3rd Gen Intel Xeon CPU, code-named “Emerald Rapids,” featuring up to 60 cores per processor. The new Xeon chips are optimized for AI inferencing and training performance in the cloud.
Intel claimed up to a 9x increase in AI training performance compared to the prior generation Sapphire Rapids chips. This could help Intel be more competitive with AMD’s EPYC server CPUs and Nvidia’s AI accelerators in the lucrative data center market.
|Intel Xeon CPU Max
|Up to 56 Performance-cores, 8 Efficient-cores, 150W TDP, AI Optimization
|Intel Xeon CPU Platinum
|Up to 60 cores, 270W TDP, Support for 8-Channel DDR5 Memory and PCIe Gen 5
Discrete Intel GPU and Habana AI Accelerator Also Announced
Rounding out its portfolio, Intel announced the data center-focused Intel Data Center GPU Max Series, code named “Ponte Vecchio.” It combines multiple chips using Intel’s new multi-tile architecture to deliver high performance optimized for AI and high performance computing applications.
Intel also introduced its first AI training accelerator built using its Habana Labs technology. The Habana Gaudi3 is manufactured on TSMC’s 5nm process node and is claimed to deliver faster training performance than Nvidia’s flagship H100 GPU.
Outlook: More AI Innovation Expected from Intel with Help from Partners
With these slew of new product introductions across PCs and the data center, Intel is signaling its ambition to be a major player in AI chips going forward. While they have some catching up to do to industry leaders like Nvidia, the combination of Intel’s x86 CPUs, Habana AI accelerators, Arc GPUs and oneAPI software stack provides an alternative AI ecosystem.
Through its IPU Alliance which includes other chipmakers like Qualcomm and TSMC, Intel is trying to establish its Neural Processing Unit and oneAPI software as an industry standard. If more PC and data center hardware partners adopt Intel’s AI solutions, it could see rapid innovation that allows it to close the gap versus rivals.
To err is human, but AI does it too. Whilst factual data is used in the production of these articles, the content is written entirely by AI. Double check any facts you intend to rely on with another source.