What podcasts say about Neil Movva
Every statement, with the speaker, the exact quote and the moment it was said.
What Neil Movva has said on podcasts
45 statements · 32 positive · 9 negative · 1 mixed · 3 neutral
- on SAIL Research LLM software stackPositiveAug 25, 2026· Invest Like the Best with Patrick O'Shaughnessy
SAIL built its LLM stack to maximize GPU efficiency
“we tried to build the entire LLM software stack around peak GPU efficiency”
Open the episode · Neil Movva - Making AI 10x Cheaper - [Invest Like the Best, EP.488]Listen at 15:16
SAIL aims to produce more tokens per chip than any other company.
“We wanted to squeeze out more tokens from the same chip than anyone else in the world.”
Open the episode · Neil Movva - Making AI 10x Cheaper - [Invest Like the Best, EP.488]Listen at 15:21
SAIL aims to reach the hardware’s maximum achievable performance
“We're chasing 100% speed of light.”
Open the episode · Neil Movva - Making AI 10x Cheaper - [Invest Like the Best, EP.488]Listen at 18:21
AI usage will shift from chatbots toward proactive or background agents.
“I believe that's the most profound change we're going to see in the next year. We're going to move away from chatbots to more proactive or background agents.”
Open the episode · Neil Movva - Making AI 10x Cheaper - [Invest Like the Best, EP.488]Listen at 20:25
AI usage will shift from chatbots toward proactive background agents
“We're going to move away from chatbots to more proactive or background agents.”
Open the episode · Neil Movva - Making AI 10x Cheaper - [Invest Like the Best, EP.488]Listen at 20:28
- on Background-agent AI infrastructurePositiveAug 25, 2026· Invest Like the Best with Patrick O'Shaughnessy
Background-agent workloads favor throughput-oriented AI infrastructure
“In that world, it makes a lot more sense to build a stack around throughput.”
Open the episode · Neil Movva - Making AI 10x Cheaper - [Invest Like the Best, EP.488]Listen at 20:31
Some non-NVIDIA chips offer better compute per dollar
“There are other chips that definitely rank higher than Nvidia on flops per dollar”
Open the episode · Neil Movva - Making AI 10x Cheaper - [Invest Like the Best, EP.488]Listen at 24:18
NVIDIA is effectively required for tensor parallelism
“Nvidia is basically mandatory for that.”
Open the episode · Neil Movva - Making AI 10x Cheaper - [Invest Like the Best, EP.488]Listen at 24:31
- on Cerebras wafer-scale systemsPositiveAug 25, 2026· Invest Like the Best with Patrick O'Shaughnessy
Cerebras-style wafer-scale systems can reach roughly 1,000 tokens per second
“So there you go. You have a path to a thousand tokens per second.”
Open the episode · Neil Movva - Making AI 10x Cheaper - [Invest Like the Best, EP.488]Listen at 28:56
Cerebras-style accelerators will be used in hybrid systems
“What happens to them is some hybrid.”
Open the episode · Neil Movva - Making AI 10x Cheaper - [Invest Like the Best, EP.488]Listen at 29:02
- on Million-token context windowsNegativeAug 25, 2026· Invest Like the Best with Patrick O'Shaughnessy
Using the full million-token context window is not generally beneficial.
“I don't think it's actually great to hit the full length.”
Open the episode · Neil Movva - Making AI 10x Cheaper - [Invest Like the Best, EP.488]Listen at 31:10
KV-cache requirements will constrain high-speed weight-storage architectures.
“However, the KV cache is going to be a big thorn in your side.”
Open the episode · Neil Movva - Making AI 10x Cheaper - [Invest Like the Best, EP.488]Listen at 31:24
KV-cache requirements will constrain fast inference systems
“the KV cache is going to be a big thorn in your side.”
Open the episode · Neil Movva - Making AI 10x Cheaper - [Invest Like the Best, EP.488]Listen at 31:24
- on GPU-Cerebras hybrid inferencePositiveAug 25, 2026· Invest Like the Best with Patrick O'Shaughnessy
Hybrid systems may place attention on GPUs and MLPs on Cerebras chips
“You would like to put the attention possibly on the GPU and the MLP on the Cerebris chip.”
Open the episode · Neil Movva - Making AI 10x Cheaper - [Invest Like the Best, EP.488]Listen at 32:50
- on Large GPU matrix multiplicationNeutralAug 25, 2026· Invest Like the Best with Patrick O'Shaughnessy
Large matrix multiplication on GPUs reaches roughly 70–80% peak utilization and is power-limited.
“That operation runs at 70, 80% of peak utilization, and it's limited not by software, but by power.”
Open the episode · Neil Movva - Making AI 10x Cheaper - [Invest Like the Best, EP.488]Listen at 42:32
- on Large GPU matrix multiplicationPositiveAug 25, 2026· Invest Like the Best with Patrick O'Shaughnessy
Large matrix multiplications reach roughly 70–80% GPU utilization
“That operation runs at 70, 80% of peak utilization”
Open the episode · Neil Movva - Making AI 10x Cheaper - [Invest Like the Best, EP.488]Listen at 42:32
Rack-scale computing will drive future AI efficiency and speed
“my belief is that that shape of compute is the future of efficiency and speed.”
Open the episode · Neil Movva - Making AI 10x Cheaper - [Invest Like the Best, EP.488]Listen at 43:27
- on SAIL heterogeneous chip servingPositiveAug 25, 2026· Invest Like the Best with Patrick O'Shaughnessy
SAIL can make any chip viable when priced appropriately
“There's no bad chips. There's only bad pricing. I will make any chip work at the right price.”
Open the episode · Neil Movva - Making AI 10x Cheaper - [Invest Like the Best, EP.488]Listen at 45:14
- on Advanced Micro Devices, Inc. (AMD)PositiveAug 25, 2026· Invest Like the Best with Patrick O'Shaughnessy
Meta and OpenAI have purchased substantial quantities of AMD chips.
“I think AMD is actually somewhat popular amongst some large buyers. I think publicly, Meta and OpenAI have bought a ton of AMD chips.”
Open the episode · Neil Movva - Making AI 10x Cheaper - [Invest Like the Best, EP.488]Listen at 45:56
- on Advanced Micro Devices, Inc. (AMD)PositiveAug 25, 2026· Invest Like the Best with Patrick O'Shaughnessy
AMD AI chips have gained adoption among large buyers
“AMD is actually somewhat popular amongst some large buyers.”
Open the episode · Neil Movva - Making AI 10x Cheaper - [Invest Like the Best, EP.488]Listen at 45:56
SAIL intends to integrate TPUs and Trainium into heterogeneous serving systems.
“Yes, we love TPUs. We're going to make TPUs work. Yes, we love Tranium. We're going to make Tranium work.”
Open the episode · Neil Movva - Making AI 10x Cheaper - [Invest Like the Best, EP.488]Listen at 47:20
SAIL intends to integrate TPUs and Trainium into heterogeneous serving
“We're going to make TPUs work. Yes, we love Tranium. We're going to make Tranium work.”
Open the episode · Neil Movva - Making AI 10x Cheaper - [Invest Like the Best, EP.488]Listen at 47:21
AI inference spending will continue increasing monotonically
“I do think inference spend monotonically increases.”
Open the episode · Neil Movva - Making AI 10x Cheaper - [Invest Like the Best, EP.488]Listen at 50:45
Training clusters can support inference workloads
“Training is a superset workload over inference.”
Open the episode · Neil Movva - Making AI 10x Cheaper - [Invest Like the Best, EP.488]Listen at 51:36
- on AI training and inference clustersMixedAug 25, 2026· Invest Like the Best with Patrick O'Shaughnessy
Training clusters can support inference, whereas inference clusters may not support training.
“Training is a superset workload over inference. You can make any training cluster work for inference, but maybe not vice versa.”
Open the episode · Neil Movva - Making AI 10x Cheaper - [Invest Like the Best, EP.488]Listen at 51:36
- on Distributed inference data centersPositiveAug 25, 2026· Invest Like the Best with Patrick O'Shaughnessy
Inference can run effectively in distributed one-megawatt data centers
“increasingly we're seeing a few new thinkers realize that inference is going to be suitable for these distributed one megawatt data centers.”
Open the episode · Neil Movva - Making AI 10x Cheaper - [Invest Like the Best, EP.488]Listen at 52:45
SAIL will use distributed small compute pools across the United States for inference.
“We're quite in agreement with that. And we are very happy to buy small pools of compute across the United States and use that as our inference fleet.”
Open the episode · Neil Movva - Making AI 10x Cheaper - [Invest Like the Best, EP.488]Listen at 52:52
SAIL treats machine performance as improvable rather than fixed.
“We don't accept the state of the machines we work on is fixed.”
Open the episode · Neil Movva - Making AI 10x Cheaper - [Invest Like the Best, EP.488]Listen at 1:05:08
Neil favors offloading KV cache to flash memory.
“I'm quite passionate about that idea.”
Open the episode · Neil Movva - Making AI 10x Cheaper - [Invest Like the Best, EP.488]Listen at 1:10:42
Model architectures could be redesigned to enable greater KV-cache flash offloading
“what would we have to change about the model architecture to make offloading KVCache to flash work at a much greater level.”
Open the episode · Neil Movva - Making AI 10x Cheaper - [Invest Like the Best, EP.488]Listen at 1:10:46
Statements are attributed to the speaker as said on the episode and reflect their view at the time, not PodLume's. They are not advice.