Mustafa Suleyman says Artificial Intelligence compute growth is still accelerating

Mustafa Suleyman argues that Artificial Intelligence development is being propelled by simultaneous advances in chips, memory, networking, and software efficiency rather than nearing a hard limit. He contends that rising compute capacity and falling deployment costs will push systems beyond chatbots toward more capable agents.

Mustafa Suleyman frames modern Artificial Intelligence progress as an exponential phenomenon that runs counter to human intuition about linear change. From the time he began work on Artificial Intelligence in 2010 to now, the amount of training data that goes into frontier Artificial Intelligence models has grown by a staggering 1 trillion times, from roughly 10¹⁴ flops for early systems to over 10²⁶ flops for today’s largest models. He presents that jump as the central force behind recent advances and rejects recurring claims that development is close to hitting a wall.

He argues that the acceleration is coming from several layers of the computing stack at once. Nvidia’s chips have delivered an over sevenfold increase in raw performance in just six years, from 312 teraflops in 2020 to 2,250 teraflops today. Microsoft’s Maia 200 chip, launched this January, delivers 30% better performance per dollar than any other hardware in the company’s fleet. HBM3 triples the bandwidth of its predecessor, while interconnect technologies such as NVLink and InfiniBand link hundreds of thousands of GPUs into warehouse-size supercomputers. Where training a language model took 167 minutes on eight GPUs in 2020, it now takes under four minutes on equivalent modern hardware. Moore’s Law would predict only about a 5x improvement over this period. He says the industry saw 50x instead, and notes a shift from two GPUs training AlexNet in 2012 to over 100,000 GPUs in today’s largest clusters.

Suleyman also points to software gains that are making models cheaper to train and serve. Research from Epoch AI suggests that the compute required to reach a fixed performance level halves approximately every eight months, much faster than the traditional 18-to-24-month doubling of Moore’s Law. The costs of serving some recent models have collapsed by a factor of up to 900 on an annualized basis. He says those trends indicate that Artificial Intelligence is becoming radically cheaper to deploy even as capabilities improve.

Looking ahead, he describes a continued surge rather than a slowdown. Leading labs are growing capacity at nearly 4x annually, and since 2020, the compute used to train frontier models has grown 5x every year. Global Artificial Intelligence-relevant compute is forecast to hit 100 million H100-equivalents by 2027, a tenfold increase in three years. He says that could amount to another 1,000x in effective compute by the end of 2028, and that by 2030 an additional 200 gigawatts of compute could come online every year. He links that scale to a transition from chatbots to nearly human-level agents capable of carrying out extended, semiautonomous work across industries, while acknowledging energy as the clearest constraint. A single refrigerator-size Artificial Intelligence rack consumes 120 kilowatts, equivalent to 100 homes, but he argues that falling solar and battery costs create a path for cleaner scaling.

58

Impact Score

China and the US are leading different Artificial Intelligence races

The US leads in large language models and advanced chips, while China has built a major advantage in robotics and humanoid manufacturing. That balance is shifting as Chinese developers narrow the gap in model performance and both countries push to combine software and machines.

Congress weighs Artificial Intelligence transparency rules

Bipartisan lawmakers are pushing a federal transparency standard for the largest Artificial Intelligence models as Congress works on a broader national framework. The proposal aims to increase public trust while avoiding stricter state-by-state requirements and heavier regulation.

Report finds California creative job losses are not driven by Artificial Intelligence

New research from Otis College of Art and Design finds California’s recent creative industry job losses stem from cost pressures and structural shifts, not direct worker displacement by generative Artificial Intelligence. The technology is changing workflows and expectations, but it is largely replacing tasks rather than entire jobs.

U.S. senators propose broader chip tool export ban for Chinese firms

A bipartisan proposal in the U.S. Senate would shift semiconductor equipment controls from specific fabs to targeted Chinese companies and their affiliates. The measure is aimed at cutting off access to advanced lithography and other wafer fabrication tools for firms such as Huawei, SMIC, YMTC, CXMT, and Hua Hong.

Trump executive order targets state Artificial Intelligence laws

Executive Order 14365 lays out a federal strategy to discourage, challenge, and potentially preempt state Artificial Intelligence laws viewed as burdensome. Employers are advised to keep complying with current state and local rules while preparing for regulatory uncertainty in 2026.

Contact Us

Got questions? Use the form to contact us.

Contact Form

Clicking next sends a verification code to your email. After verifying, you can enter your message.