Prediction Open AI assessment confidence: 95% certainty 4/5 debate potential 3/5

Feldman: AI will rely less on transformers in 3 to 5 years

Andrew Feldman · Andrew Feldman, Cerebras Co-Founder and CEO: The AI Chip Wars & The Plan to Break Nvidia's Dominance · 20VC with Harry Stebbings · Mar 24, 2025 · at 34:05

Andrew Feldman, co-founder and CEO of Cerebras, discusses future shifts in fundamental AI model architectures away from current transformer dominance.

0:00 / 0:03exact quote · 3.9s

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

“We won't be as dependent on transformers in three years or five years as we are now.”

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

The AI industry will be significantly less reliant on transformer architectures in three to five years than it is today.

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 claim specifies a timeframe of three to five years from March 24, 2025, setting the evaluation horizon between March 2028 and March 2030. As of July 2026, only 16 months have elapsed. While hybrid architectures (e.g., Transformer-SSM models like Nemotron-3 Nano Omni and Bamba) and non-transformer models (such as Liquid AI's LFMs) show early growth in specific deployment niches, top-tier frontier models remain predominantly Transformer-based. Final resolution cannot be determined until industry reports and deployment surveys can be assessed during the target 2028–2030 period.

The target evaluation window for the 3-to-5-year prediction is March 2028 to March 2030.
As of Q1 2025 baseline, Transformer architectures accounted for over 95% of top-tier foundation models.
By mid-2026, hybrid architectures (such as Mamba-Transformer hybrids) and non-transformer Liquid LFMs have gained market share in edge and high-throughput contexts, but cloud-hosted frontier LLMs on primary leaderboards remain anchored in Transformer/MoE backbones.

Checked against

[1] youtube.com [2] youtube.com [3] algorithmine.com [4] medium.com [5] mindstudio.ai [6] askaibrain.com [7] waylandz.com [8] amazon.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: 2030-03-24; checked again automatically on the next resolution wave

Mentioned

Transformer

⚑ Spot an error? Report it

More from Andrew Feldman

Assertion Partly supported

Nvidia Conceals Massive Chip Field Failure Rates to Keep Customers Waiting

“You don't talk about the field failure rate of your products, which are massive. Rather you sort of keep talking about the future in an effort to convince people to wait, to make a good decision rather than go with technology that's, that that's better and pre…”

Andrew Feldman Oct 6, 2025 ▶ 12:47 Cerebras CEO, Andrew Feldman on Why Raise $1BN and Delay the IPO & Why NVIDIA’s Worried About Growth · 20VC with Harry Stebbings

Assertion Supported

Elon Musk and Tesla Failed at Building Wafer-Scale Chips for Dojo

“And after we did it Elon tried Dojo and they failed.”

Andrew Feldman Oct 6, 2025 ▶ 23:29 Cerebras CEO, Andrew Feldman on Why Raise $1BN and Delay the IPO & Why NVIDIA’s Worried About Growth · 20VC with Harry Stebbings

Assertion Supported

Feldman: Nvidia Has Zero CUDA Software Lock-In for AI Inference

“In inference, it's not real at all. There's no CUDA lock-in in inference. None. Well, you can move from OpenAI on an NVIDIA GPU to Cerebrus to Fireworks Service on something else to Together to perplexity with 10 keystrokes. I mean, anybody who actually uses A…”

Andrew Feldman Mar 24, 2025 ▶ 42:10 Andrew Feldman, Cerebras Co-Founder and CEO: The AI Chip Wars & The Plan to Break Nvidia's Dominance · 20VC with Harry Stebbings

Opinion

Feldman states the Trump administration is unwaveringly better for business

“Unwaveringly better for business. You know, there are things I agree with, there are things I disagree with in this administration, but unwaveringly better for business.”

Andrew Feldman May 26, 2026 ▶ 57:10 Cerebras CEO on the Future of Data Centres, Token Costs & Memory | Should US Companies Sell to China · 20VC with Harry Stebbings

Assertion Not checkable as stated

No Company Ever Went Bankrupt by Overpaying Extraordinary Talent

“No company ever went bankrupt by paying extraordinary people too much.”

Andrew Feldman Oct 6, 2025 ▶ 43:00 Cerebras CEO, Andrew Feldman on Why Raise $1BN and Delay the IPO & Why NVIDIA’s Worried About Growth · 20VC with Harry Stebbings

Assertion Partly supported

Feldman: Cerebras chips are faster than Nvidia at both training and inference

“No, we're faster on both, but the software challenges In training are real.”

Andrew Feldman Oct 6, 2025 ▶ 24:13 Cerebras CEO, Andrew Feldman on Why Raise $1BN and Delay the IPO & Why NVIDIA’s Worried About Growth · 20VC with Harry Stebbings

Feldman: AI will rely less on transformers in 3 to 5 years

More from Andrew Feldman

Turn any episode into a week of clips.