Prediction Open AI assessment confidence: 95% certainty 3/5 debate potential 3/5

Morin predicts AI compute will be 95% inference within five years

Steeve Morin · Steeve Morin: Why Google Will Win the AI Arms Race & OpenAI Will Not | E1262 · 20VC with Harry Stebbings · Feb 24, 2025 · at 0:15

Steeve Morin, founder and CEO of ZML, discusses the future workload shift in AI infrastructure from model training to inference.

0:00 / 0:04exact quote · 4.8s

▶ Watch the full episode on YouTube → 720p mp4 · rendered on demand · StarZero watermark

“In five years, I would say 95% inference, five percent training.”

quote is from the automated transcript, cleaned for reading: filler sounds and stutters are removed, nothing is rephrased. names can be misheard (the analysis reads context, assessments check outside sources). how →

The claim, plainly

Steeve Morin predicts that within five years, 95% of AI compute workload will be for inference and 5% for training.

a quote alone can be ambiguous, so this restates it with the references and context of the conversation around it resolved. AI-written; the tape below the player is the source.

Assessment

The target horizon for this five-year prediction made on February 24, 2025, is February 24, 2030. As of the current date (July 23, 2026), the timeline has not elapsed. Current industry analyst reports from Gartner, McKinsey, Deloitte, and Iron Mountain project that inference will represent between 50% and 80%+ of total AI compute or infrastructure load by 2030, but finalized empirical aggregate data for 2030 is not yet available.

Gartner projects that over 65% of AI-optimized IaaS spending will support inference by 2029, and over 80% of data center workload accelerators will be dedicated to inference by 2030.
McKinsey predicts inference will exceed 50% of total AI compute workload by 2030.
Iron Mountain projects AI inference reaching roughly 80% of total AI critical IT load capacity by 2030.
ABI Research forecasts inference workloads will not overtake training workloads in power consumption and volume until 2033.

Checked against

[1] thetwentyminutevc.com [2] introl.com [3] weka.io [4] radical.vc [5] mckinsey.com [6] mirantis.com [7] deloitte.com [8] ironmountain.com

assessed as of 2026-07-23

This assessment is an AI reading of a public statement, made against the cited references with the confidence shown. It can be wrong. The quote, the tape and the references are all on this page; if something is off, use report an error below. full disclaimer →

⏳ prediction date: 2030-02-24; checked again automatically on the next resolution wave

Mentioned

AI compute

⚑ Spot an error? Report it

More from Steeve Morin

Opinion

Morin: Nvidia is far from the most efficient AI hardware platform

“But it's by far not the most efficient platform. And arguably, even in terms of software, it's not the best software platform.”

Steeve Morin Feb 24, 2025 ▶ 6:08 Steeve Morin: Why Google Will Win the AI Arms Race & OpenAI Will Not | E1262 · 20VC with Harry Stebbings

Prediction Not checkable as stated

Morin: The market bubble around Nvidia H100 GPUs will burst

“There's going to be a need for inference. Very hard to say whether it will be worth, you know, everybody's money to do it on H 100. That is a bubble that I think will blow some time.”

Steeve Morin Feb 24, 2025 ▶ 6:44 Steeve Morin: Why Google Will Win the AI Arms Race & OpenAI Will Not | E1262 · 20VC with Harry Stebbings

Insight

Morin: GPUs are a clever workaround, not natively built for AI

“GPUs are, you know, are a good trick for AI, but they're not built for AI.”

Steeve Morin Feb 24, 2025 ▶ 11:13 Steeve Morin: Why Google Will Win the AI Arms Race & OpenAI Will Not | E1262 · 20VC with Harry Stebbings

Prediction Open · timeframe Dec 2025

Morin: AI chip oversupply will lead to GPUs selling at 30% value

“I very much worry there will be an oversupply of these chips. The problem is, is that, you know, remember, the chips are the collateral. So, you know, somewhere, you know, in the US or whatever, there's going to be a data center with like a thousand GPUs that …”

Steeve Morin Feb 24, 2025 ▶ 24:31 Steeve Morin: Why Google Will Win the AI Arms Race & OpenAI Will Not | E1262 · 20VC with Harry Stebbings

Assertion Contradicted

Morin: GPUs cannot deliver latent space AI reasoning at scale

“Fundamentally, GPUs cannot deliver, deliver this, plain and simple at scale.”

Steeve Morin Feb 24, 2025 ▶ 31:24 Steeve Morin: Why Google Will Win the AI Arms Race & OpenAI Will Not | E1262 · 20VC with Harry Stebbings

Prediction Not checkable as stated

Morin: Nvidia may lose market dominance in AI inference and training

“I think that there's a shot that they don't.”

Steeve Morin Feb 24, 2025 ▶ 35:25 Steeve Morin: Why Google Will Win the AI Arms Race & OpenAI Will Not | E1262 · 20VC with Harry Stebbings

Morin predicts AI compute will be 95% inference within five years

More from Steeve Morin

Turn any episode into a week of clips.