AI chips and export controls
Who makes the chips, who can buy them and what they cost to run. Analysis of Nvidia, custom silicon and export controls from The AI Realist, and measured benchmarks of AWS Trainium and Inferentia, Intel CPUs and Apple silicon.
16 articles and 16 videos. Last updated August 9, 2026.
AI chips and export controls: articles
August 9, 2026
The Model Is the MachineAMD just bought a chip company whose product runs exactly one model. That is the point.
June 11, 2026
Nvidia Won the Cloud. Now It Wants the Laptop.Nvidia's new laptop chip trails Apple where it counts and it was built to win anyway.
May 27, 2026
Two Chips, One Decade, One WinnerTwo hyperscalers bet on custom AI silicon a decade ago. One built a chip that its own models run on.
May 25, 2026
Huawei Can't Buy EUV. It Says It Doesn't Need To.A new “scaling law” reframes the chip race from space to time. The export-control wall was built to deny the first. It has no answer for the second.
May 12, 2026
Below the SiliconTrump flies to Beijing on Wednesday. Here's the recipe his hosts control, element by element.
April 10, 2026
The Verification TaxThe GPU ate the AI narrative. The CPU kept doing the work.
April 3, 2026
Your Parents PaidNVIDIA built the world's most profitable hardware company by treating its consumer GPUs as a recruitment pipeline. Now many recruits are buying Macs.
April 3, 2026
What to Buy for Local LLMs (April 2026)A practitioner's guide to inference and training, why they're not the same machine, and why you may still need cloud
April 1, 2026
What China Shipped While the West Was Banning ChipsSix domains where China's deployment record contradicts the restriction narrative.
March 23, 2026
Musk's Chip GambitOn March 14, 2026, Elon Musk posted five words to X: “Terafab Project launches in 7 days.” The announcement described a $25 billion semiconductor fabrication facility targeting 2nm process technolo...
March 20, 2026
Every Country Needs Sovereign AI. Jensen Is Selling It.On March 16, Jensen Huang stood on the GTC stage in San Jose and showed the audience a map. Dozens of countries. More than 1 million Nvidia GPUs are deployed worldwide through the company's cloud p...
March 15, 2026
AWS Built Its Own AI Chip. Now It Needs Someone Else's.In March 2024, I explained to my YouTube audience how LLM inference actually works.[1] Two distinct phases, I said. Prefill, the phase where the model processes your entire prompt in parallel, fil...
March 4, 2026
South Korea Powers the World's AI Chips. Where are its LLMs?On January 15, 2026, South Korea's Ministry of Science and ICT announced the results of the first elimination round in what the Korean press had taken to calling the AI Hunger Games.[1] Five govern..
March 1, 2026
Chip and MortarOn Friday morning, Amazon announced a $50 billion investment in OpenAI, $15 billion upfront, another $35 billion conditional on milestones that reportedly include an IPO or the achievement of AGI ...
February 14, 2026
Jensen's COMECON: How Nvidia Built an Empire of Captive CloudsIn January 2026, Nvidia invested an additional $2 billion in CoreWeave, bringing its total stake to $5.3 billion and to 11.5% of the company. The press release used the language of partnership: "de...
August 20, 2025
A not-so-silent revolution is happening in AI inferenceThe combination of smaller yet more efficient language models, advances in open-source inference frameworks, and hardware acceleration is enabling an unprecedented pace of innovation for CPU infere...
AI chips and export controls: videos
May 31, 2026
Benchmarking TurboQuant with MLX on Apple SiliconKV-cache quantization in MLX: I benchmarked TurboQuant on Apple Silicon to find out how much memory does it really save, and where it breaks.
September 11, 2025
Running AFM-4.5B on Intel CPUs with OpenVINOUnlock the full potential of your Intel CPU server processor! In this demo.
July 14, 2025
SLM Inference on a Windows laptop 🤯 Intel Lunar Lake CPU/GPU/NPU + OpenVINOUnlock the full potential of your Intel Lunar Lake processor! In this demo.
November 20, 2024
Run performant and cost-effective GenAI Applications with AWS Graviton and Arcee AILive session on Twitch, 11/19/2024 We first discuss why AWS Graviton CPU instances are a great fit for AI inference.
March 20, 2024
Phi-2 on Intel Meteor Lake - Coding questionMarch 20, 2024
Phi-2 on Intel Meteor Lake - Physics questionJanuary 23, 2024
Deep Dive: Advanced distributed training with Hugging Face LLMs and AWS TrainiumFollowing up on the "Hugging Face on AWS accelerators" deep dive this video zooms in on distributed training with NeuronX Distributed, Optimum Neuron.
January 16, 2024
Deep Dive: Hugging Face models on AWS AI AcceleratorsExplore the technical intricacies of optimizing Hugging Face models on AWS accelerators in this detailed walkthrough.
May 17, 2023
Transformer training shootout, part 2: AWS Trainium vs. NVIDIA V100In this video, I compare the cost/performance of AWS Trainium with the NVIDIA V100 GPU.
April 14, 2023
Accelerate Transformer inference with AWS Inferentia 2In this video, I show you how to accelerate Transformer inference with AWS Inferentia 2, a custom chip designed by AWS.
February 24, 2023
Transformer training shootout: AWS Trainium vs. NVIDIA A10GIn this video, I compare the cost-performance of AWS Trainium, a new custom chip designed by AWS, with NVIDIA A10G GPUs.
February 7, 2023
Accelerate PyTorch Transformers with Intel Sapphire Rapids, part 2In this video, you will learn how to accelerate PyTorch inference with an Intel Sapphire Rapids server running on AWS.
January 2, 2023
Accelerate PyTorch Transformers with Intel Sapphire Rapids, part 1In this video, you will learn how to accelerate a PyTorch training job with a cluster of Intel Sapphire Rapids servers running on AWS.
November 17, 2022
Accelerate Transformer inference with AWS InferentiaIn this video, I show you how to accelerate Transformer inference with AWS Inferentia, a custom chip designed by AWS.
October 13, 2022
Accelerate Transformer training with AWS TrainiumIn this video, I show you how to accelerate Transformer training with AWS Trainium, a new custom chip designed by AWS.
December 16, 2019
Deep Learning hardware acceleration with AWS InferentiaIn this video, I show you how to get started with AWS Inferentia, and Amazon EC2 Inf1 instances.
About the author
Julien Simon is an AI Operating Partner at Fortino Capital. He was VP and Chief Evangelist at Arcee AI, Chief Evangelist at Hugging Face and Global Evangelist for AI and machine learning at AWS, and wrote “Learn Amazon SageMaker” (Packt). He has given 693+ talks in 37 countries and published 502+ technical articles. See his experience and Wikidata entry.
Method: figures come from filings, court documents, vendor documentation and measurements I ran, and each piece dates its numbers.
Go further
More topics
Frequently Asked Questions
- Where can I read independent analysis of AI chips, GPUs and export controls?
- Julien Simon covers ai chips and export controls in 16 articles and 16 videos on julien.org. The newest articles are "The Model Is the Machine" (August 9, 2026); "Nvidia Won the Cloud. Now It Wants the Laptop." (June 11, 2026); "Two Chips, One Decade, One Winner" (May 27, 2026). Who makes the chips, who can buy them and what they cost to run. Analysis of Nvidia, custom silicon and export controls from The AI Realist, and measured benchmarks of AWS Trainium and Inferentia, Intel CPUs and Apple silicon.
- Where are Julien Simon's pieces on ai chips and export controls published?
- Articles appear first in The AI Realist newsletter (airealist.ai) and are archived on julien.org. Videos are on the YouTube channel @juliensimonfr, each with a full transcript on julien.org.
- Which videos does Julien Simon have on ai chips and export controls?
- 16 videos on the YouTube channel @juliensimonfr, each with a transcript on julien.org. The newest are "Benchmarking TurboQuant with MLX on Apple Silicon" (May 31, 2026); "Running AFM-4.5B on Intel CPUs with OpenVINO" (September 11, 2025); "SLM Inference on a Windows laptop 🤯 Intel Lunar Lake CPU/GPU/NPU + OpenVINO" (July 14, 2025).