Jim Seto opened his session at Mining Disrupt Miami 2024 by saying he would rather not be on stage at all. The CEO of ePIC Blockchain has spent about 30 years in semiconductors, most of it in mobile and PC technology, and said he normally leaves the talking to his team and the company's products. He then spent his slot on one idea: inflection points, how often people miss them, and where he saw them in Bitcoin mining hardware.
The session was recorded at Mining Disrupt Miami, June 24–26, 2024. Bitcoin traded at $61,272 on the first day of the event and at $81,502 on October 17, 2026.
Why a fast industry moved slowly
Seto said that coming from mobile and PC hardware, where technology, markets and customer requirements change quickly, he expected Bitcoin mining to grow just as fast. It did not, and he asked himself why.
But then I also realized that there was an actual, call it an inertial resistance to it.
Jim Seto · 2:01
He set the question against a slide of earlier shifts. Information storage moved from books to magnetic media to terabytes that fit on a fingertip, which spread knowledge. Communication moved from carrier pigeons to wired and then wireless networks. Stored value, in his telling, moved from coins to gold to a decentralized monetary system. The resistance to that last step, he said, came with political overlays and with the investment that infrastructure needs, so it was still a long road. His conclusion was that the industry is still early, which he called a great opportunity.
Inflection points people miss
Seto's next point was that inflection points often pass unnoticed until after the fact.
Here's an example where an inflection point actually happened, and people didn't realize it.
Jim Seto · 3:48
His example was a company that went to market with a plan to move books around the world. He guessed that 90% of the people who saw that plan did not think it would work, and that the 10% who understood the vision saw how it could change shopping, interaction and logistics. He left the outcome as a rhetorical question and said it had gone very well.
He then turned to phones. One maker, he said, believed a phone had to have a keyboard; another thought grayscale was enough and color was not needed. Customers wanted a full display with the keyboard under the glass. The makers that listened to customers moved ahead, he said, while others kept improving the mousetrap they already had.
Mining infrastructure has grown up
Seto showed early photos of mining setups, from home miners to the first industrial sites, and acknowledged that they suggest things had not gone well. He took the other view. The new facilities ePIC has seen, he said, pay attention to architecture and to the whole system, and at some industrial miners the team visited you could almost eat off of the floor. In his reading, the industry has recognized that world-class scale is needed, and that change is already under way.
ePIC's three lines of business
ePIC's plan, Seto said, came from two years of customer meetings.
Again, kind of meeting with the customers in the last two years, listening to what they want, providing innovation with the passion that we have for operational excellence.
Jim Seto · 7:53
The company started by designing SoC chips for blockchain, which he called its DNA. It does not design chips now, but clients ask it to assess the SoCs on the market and what to watch out for. Referring to an earlier speaker's point on diminishing returns, Seto said Moore's Law may have made chips close to a level playing field, and that the differentiation now sits in what is added on top: the system, the full solution and the integration layer above it.
The second line is fleet enhancement: control boards and firmware that improve miners ePIC did not design. The third is complete systems, with announcements planned between the talk and the end of that summer. He listed four competencies the company combines, firmware, system design, SoC design and an understanding of the technology, and said few hardware companies can put all four together and partner with customers upstream.
Who is entering mining hardware
Seto described the ecosystem as a dial: the supply chain on one half, from foundries and system builders to ASIC providers, and on the other half the customers buying product and the sites where hardware is deployed. He pointed to newer entrants as well. The financial sector, including very large stablecoin issuers, exchanges and remittance and e-commerce providers are all moving into the space, he said, and he read that as a sign the industry has matured.
From shoe boxes to custom miners
The usual product, Seto said, is the air-cooled shoe box, mainly from makers in China. When ePIC rolled out its own designs it worked with data centers on 4U prototypes and built some miners to be immersion ready from the start. The bigger change in this hardware refresh cycle, he said, is that infrastructure considerations now feed back into the requirements.
So we get requirements actually to do more custom miners, not the traditional shoe boxes, but more of a larger scale with unique features.
Jim Seto · 12:35
Customers also want a provider that can help extend the life of their infrastructure, which he said is what a mining site is: infrastructure you keep reusing. He tied this to efficiency, as hardware moves from 34 joules per terahash to 30, to 26 and lower.
The fact that you could actually reuse the infrastructure really starts to matter.
Jim Seto · 13:12
At scale, he added, networking, three-phase distribution and site integration all shape how the hardware is designed. Customers also asked for flexibility. At ePIC's booth, 382, the company showed control boards built on a common module, with the carrier underneath changed to fit each form factor, including Bitmain and MicroBT machines and ePIC's own custom systems.
The AI comparison
About 10 minutes in, Seto noted he had not yet said the word AI, which he said had taken a lot of the oxygen out of the room. He showed a chart of NVIDIA, Bitcoin and gold over the previous five years, and said a fourth curve would have beaten NVIDIA.
And that would be actually a system, firmware, customization, integration provider.
Jim Seto · 15:25
He named that company as Supermicro. His point was that value in a hardware boom can sit with integrators that people do not notice. Infrastructure providers might serve both markets, he said, but ePIC itself would stay focused on hardware for blockchain mining.
He closed by describing ePIC as a focused engineering company in a fast growth phase, increasingly asked to help with road mapping and with defining what the future data center looks like.