Get your free personalized podcast brief
We scan new podcasts and send you the top 5 insights daily.
The AI hardware market is splitting into two distinct segments: training and inference. While NVIDIA dominates training, the larger, long-term opportunity lies in inference. This is creating a market for specialized, memory-optimized chips from companies like Cerebras and Grok designed for running models efficiently.
Related Insights
The AI inference process involves two distinct phases: "prefill" (reading the prompt, which is compute-bound) and "decode" (writing the response, which is memory-bound). NVIDIA GPUs excel at prefill, while companies like Grok optimize for decode. The Grok-NVIDIA deal signals a future of specialized, complementary hardware rather than one-size-fits-all chips.
While focus is on massive supercomputers for training next-gen models, the real supply chain constraint will be 'inference' chips—the GPUs needed to run models for billions of users. As adoption goes mainstream, demand for everyday AI use will far outstrip the supply of available hardware.
The AI hardware market is fragmenting. Google is now producing two distinct eighth-generation TPUs: one for training (8t) and one for inference (8i). This move away from one-size-fits-all GPUs shows that optimizing for specific AI workloads is the next competitive frontier.
Despite its high valuation post-IPO, AI chipmaker Cerebras's long-term strategy focuses on inference, not just training. The bet is that inference will become a much larger segment of the AI compute market. By developing chips specifically optimized for this task, Cerebras aims to take significant market share from NVIDIA.
While Nvidia dominates the AI training chip market, this only represents about 1% of the total compute workload. The other 99% is inference. Nvidia's risk is that competitors and customers' in-house chips will create cheaper, more efficient inference solutions, bifurcating the market and eroding its monopoly.
The primary bottleneck for AI inference is now memory (HBM), not compute. To circumvent this, industry giants Nvidia and AWS are making multi-billion dollar deals for systems from Groq and Cerebrus that use on-chip SRAM, which is faster and not subject to the same supply constraints.
The AI inference process is being broken apart, with different stages of the transformer architecture running on different specialized chips. For example, the compute-heavy "prefill" step and the memory-heavy "decode" step can be handled by separate hardware. This explains NVIDIA's strategic interest in Grok, which excels at the decode portion.
The era of dual-purpose AI chips is ending. The overwhelming demand for real-time processing from AI agents is forcing companies like Google and NVIDIA to create dedicated, inference-optimized hardware. This marks a fundamental and permanent split in the AI infrastructure market, separating training from inference.
The inference market is too large to remain monolithic. It will fragment into specialized platforms for different use cases like real-time video, long-running agents, or language models. This specialization will extend to hardware, with high-throughput, low-latency-need tasks (like agents) favoring cheaper AMD/Intel chips over NVIDIA's top GPUs.
Beyond the simple training-inference binary, Arm's CEO sees a third category: smaller, specialized models for reinforcement learning. These chips will handle both training and inference, acting like 'student teachers' taught by giant foundational models.
