Rob Wachen
Co-founder of Etched
aka Rob Laakin
“Our hardware is going to generally be able to get you an order of magnitude more concurrency at a given level of interactivity. That directly translates into tokens per watt, tokens per dollar, all the things people care about when they're actually serving these giant mixture of expert models at scale.”
“We were able to create a new mechanism of running at much lower voltages, a new type of power delivery that we call low voltage inference. And we think all AI chips in the future are going to be low voltage chips.”
“For our first gen product, we built it on a different supply chain than the Rubens. We're on 4 nanometer, Rubens on 3 nanometer, we're on a different HPM than Rubens and so forth. So it actually is not a zero sum thing... it's not a decision between a gigawatt of a GPU and a gigawatt of us, it's 2 gigawatts.”
“The decision explicitly not to build an arbitrary graph compiler, not to support arbitrary Pytorch, not to support arbitrary cuda, not to support arbitrary ONNX graphs. But instead we envisioned a world where there was going to be under 100 models that actually mattered.”
“I firmly believe we are on a global march of inference becoming majority of global GDP. It may take more than 10 years, but it's going to happen.”
Rob Wachen
One-line summary: Co-founder of Etched; tracked for low-voltage-inference, concurrency-per-megawatt, and additive-supply-chain claims — interested party on all Etched claims. (AssemblyAI diarization mislabels him 'Rob Laakin' in the 2026-06-30 source.)
What they're known for
Brief factual context — fill in.
Why they matter to stock-market
Why this person's claims are tracked here — fill in.
Said
Speaker-attributed claims extracted from diarized sources. Each bullet mirrors one entry in quotes: frontmatter — keep them in sync.
-
"Our hardware is going to generally be able to get you an order of magnitude more concurrency at a given level of interactivity. That directly translates into tokens per watt, tokens per dollar, all the things people care about when they're actually serving these giant mixture of expert models at scale." — 2026-06-30-podcast-invest-like-the-best-etched-building-ai-hardware-to-make-inference (2026-06-30)
-
"We were able to create a new mechanism of running at much lower voltages, a new type of power delivery that we call low voltage inference. And we think all AI chips in the future are going to be low voltage chips." — 2026-06-30-podcast-invest-like-the-best-etched-building-ai-hardware-to-make-inference (2026-06-30)
-
On tsmc-capacity-shortfall-and-pricing-power:
"For our first gen product, we built it on a different supply chain than the Rubens. We're on 4 nanometer, Rubens on 3 nanometer, we're on a different HPM than Rubens and so forth. So it actually is not a zero sum thing... it's not a decision between a gigawatt of a GPU and a gigawatt of us, it's 2 gigawatts." — 2026-06-30-podcast-invest-like-the-best-etched-building-ai-hardware-to-make-inference (2026-06-30)
-
On cuda-moat-erosion-at-inference:
"The decision explicitly not to build an arbitrary graph compiler, not to support arbitrary Pytorch, not to support arbitrary cuda, not to support arbitrary ONNX graphs. But instead we envisioned a world where there was going to be under 100 models that actually mattered." — 2026-06-30-podcast-invest-like-the-best-etched-building-ai-hardware-to-make-inference (2026-06-30)
-
On enterprise-token-budgeting:
"I firmly believe we are on a global march of inference becoming majority of global GDP. It may take more than 10 years, but it's going to happen." — 2026-06-30-podcast-invest-like-the-best-etched-building-ai-hardware-to-make-inference (2026-06-30)
Sources
Related
Cross-links — fill in.