Dec 12, 2024 · 42m · no-priors
No Priors Ep. 93 | With Akash Systems' Felix Ejeckam and Ty Mitchell
0:00 / 0:00
▶ Watch on YouTube →
gold bands on the timeline = statements, start to end. Hover to read, click to jump. CC turns on captions
Akash Systems co-founders Felix Ejeckam and Ty Mitchell join host Sarah Guo to discuss how their lab-grown diamond cooling technology solves critical AI thermal bottlenecks, accelerates semiconductor performance, and strengthens domestic hardware manufacturing.
How this conversation actually went
Every chapter scored 0–10 on four independent dynamics. Hover any point for the reasoning behind the score. The hosts hold 15.7% of the talking time here. How this is scored →
The hosts as informed peer 4.9 Guest teaching 5.2 Guest disagreement 0.6 The hosts pushing back 0.9
speaking balance: gold is the hosts, purple is the guest (3 minute bins)
Sharpest disagreement ▶ 33:10 Rejecting earnings-only offshoring logic
Ty firmly pushes back against the corporate consensus of chasing margins abroad, arguing the US must manufacture domestically regardless of short-term quarterly earnings pressures.
Hardest push from the hosts ▶ 32:42 Questioning domestic lithography and fab feasibilitySarah directly questions the feasibility of near-term reshoring, asking whether it is genuinely credible that the US can stand up competitive fabs and lithography tools in a realistic window.
Biggest teaching moment ▶ 14:05 Laptop ice-pack thermal barrier explanationFelix explains the physical limits of external cooling by illustrating how placing an ice pack beneath a laptop fails to cool internal chips due to thermal interface barriers.
The host holds their own ▶ 38:10 Probing AI in inverse design and DFT simulationsSarah demonstrates deep technical familiarity with computational material science by drilling into density functional theory, inverse design, and chemical search spaces.
the scores for every segment, with the reasoning behind each
| Chapter | Topic | The hosts as informed peer | Guest teaching | Guest disagreement | The hosts pushing back | Why |
|---|---|---|---|---|---|---|
| Scientific Origins and Transitioning from Space to AI | 4 | 5 | 0 | 0 | Sarah prompts the founders to share their academic background and the genesis of Akash Systems. Felix and Ty detail their PhD research merging disparate materials and compare the extreme power density requirements of space radios against AI GPUs. | |
| Demonstrating Diamond Thermal Conductivity with an Ice Experiment | 3 | 6 | 0 | 0 | Felix conducts a live physical demonstration showing a diamond wafer rapidly melting an ice cube using ambient body heat. Sarah observes while Felix explains the economic and thermal impact of direct diamond integration. | |
| The AI Thermal Bottleneck and Hybrid Cooling Architectures | 5 | 6 | 1 | 2 | Ty explains how synthetic diamond integrates with existing cooling architectures like liquid cooling and fan-out packaging. Sarah raises reliability concerns regarding mechanical moving parts in data centers, and Felix explains why cooling must start at the material level using a laptop ice-pack analogy. | |
| Thermal Throttling, Server Architecture, and Accelerating Moore's Law | 6 | 6 | 1 | 1 | Sarah connects datacenter operational bottlenecks to thermal throttling and non-deterministic hardware behavior. Ty outlines the multi-layer thermal interface sandwich inside modern servers, while Felix claims diamond cooling can hyper-accelerate Moore's Law by eliminating thermal crosstalk. | |
| Diamond Synthesis and Semiconductor Integration at the Atomic Level | 3 | 6 | 0 | 0 | Sarah asks how growing diamonds works for industrial form factors. Ty outlines the chemical vapor deposition process from seed crystals, and Felix explains that their proprietary advantage lies in atomically bonding diamond to semiconductors. | |
| Deploying Sovereign AI Infrastructure with NextGen Data Center | 4 | 5 | 0 | 0 | Felix explains their deployment with NextGen Data Center in India, highlighting sovereign cloud requirements that keep computation and data within national borders. | |
| Domestic Semiconductor Manufacturing, CHIPS Act, and National Security | 6 | 4 | 1 | 1 | Sarah brings up Akash's CHIPS Act award and highlights the strategic importance of RF spectrum mastery in modern defense. Ty and Felix discuss domestic manufacturing and historical parallels with radar in WWII. | |
| Reshoring Hardware and Scaling with AI-Powered Manufacturing | 6 | 5 | 2 | 4 | Sarah challenges whether reshoring advanced fabs and lithography to the US is feasible in the near term. Ty rejects a purely short-term earnings focus in favor of national security resilience, and Felix argues AI-driven manufacturing will offset offshore labor advantages. | |
| Leveraging AI to Accelerate Material Science and Scientific Discovery | 7 | 4 | 0 | 0 | Sarah demonstrates domain knowledge by asking about using AI models for inverse design, chemical space search, and accelerating DFT simulations. Ty and Felix enthusiastically agree, discussing the potential to remove compute and imagination bottlenecks in materials science. |
Statements from this episode (14)
Mitchell: Materials innovators must build entire systems to convince customers
“When you have a materials level advancement, as the person who has that material level advancement, you really have to make the system to convince people that you have the solution.”
Ejeckam: Space cooling power densities are 10x higher than AI servers
“The power densities that we call in space are at the level of 10 to the three watts per square centimeters, so four to 5000 watts per square centimeter. The chips that we call in on grounds and AI in a typical server is a full order of magnitude less than that…”
Ejeckam: A 10-degree chip temperature drop is worth millions to data centers
“Initially, we're starting off with 10 degrees, which is already worth millions for any data center who has a small number of servers.”
Ejeckam: Akash expects 20 to 40 degree chip cooling within 12 months
“We're looking at further reductions, 2030, 40 degrees down the line over the next 12 months.”
Ejeckam: Akash cut semiconductor temperatures in space by 80 to 90 degrees
“In space, we already reduced temperatures there, 80 to 90 degrees.”
Mitchell: Nvidia's Blackwell rollout was delayed by heating issues
“When NVIDIA has introduced a couple of new chips first the Samsung HVM there were heating issues back then with Blackwell again, heating issues came up where that delay that rollout was delayed due to due to heating issues.”
Ejeckam: Akash ships diamond- and liquid-cooled Nvidia H200 servers
“The servers that we ship today, the H 200 from Nvidia, for example is both liquid cooled and diamond cooled.”
Ejeckam: AI data center demand will crash the grid without novel cooling
“We believe that a consciousness that if a material science and more specifically a physics or chemistry approach to solving the heat problem is not used today, that the needs of AI data centers around the world is based on projections today will crash the grid…”
Mitchell: Power, water, and heat will halt 2x-3x compute scaling cycles
“Where the drive to increased performance. You know, whatever, two to three times every two, three, four years. We're not going to be able to do that just because drawing power and water to the site and then getting the parts to run with this heat buildup, and …”
Ejeckam: Akash is entering market as a server maker, not chipmaker
“We are buying chips. We're not making GPUs. We're taking the hottest chips in the world and we're cooling them down so that you can, we can open the envelope of performance for the system architect. Okay. And so we fit in, we're coming into the world as a serv…”
Ejeckam: Diamond cooling will deliver six years of chip performance gains in two
“We think that with our diamond technology, that we will be able to hyper accelerate Moore's law so that You know, in, in, in two years, we will be achieving what previously folks had to wait six, seven years in the past to get to in terms of performance”
Mitchell: Growing synthetic diamonds uses standard semiconductor CVD methods
“It's really no different than growing other other semiconductor materials. If you're growing silicon or silicon carbide or gallium arsenide or indium phosphide, any of these electronic substrates you start with a seed crystal and then you use a typically a, yo…”
Mitchell: US will onshore fabs and lithography machines
“Yeah, it, the U.S. Will do it. The only question is whether the U.S. Will do it because it has to do it or because it wants to do it.”
Ejeckam: AI will 10x to 100x manufacturing, beating China without wage cuts
“I do think that AI will play a role in manufacturing and sort of 10 exiting or even a hundred exiting manufacturing so that we can actually outperform humans in other countries. I think that's the way it's going to look. Okay. So we will not have to reduce lab…”