I only have so many hours in a day, but I will always set at least one of those aside to read these fascinating “hands on” stories about industrial policy design and implementation. A few things jumped out from this episode:
- the enormity of managing $39B with-a-B worth of grants. Damn.
- the importance of talent. The importance of talent. The importance of talent.
- very few people discuss the real questions about opportunity costs. Good to see that issue raised here in a way that makes you appreciate that $100B spent in one big and critical area, means $100B is not available to spend in other big and critical areas. Tough, gnarly choices (one that may eventually haunt us when it comes to data center over-investments).
- whatever mistakes the CHIPS team might have made, the fact that the people on the team were above reproach will be an enduring legacy. Especially when compared to the neoroyalism/nepotism driving many current investment and manufacturing decisions. It’s a third rail too many people want to avoid touching publicly. Yet it’s dominating decisions made today.
- Secretary Raimondo’s reputation will only grow over time.
- Browbeating industry works until it doesn’t. Then it fails badly. Maybe a cautionary tale here about the Pentagon and Anthropic.
Brilliant episode and a brutally honest recap of the $39B "KKR-style" capital deployment. This was the necessary first step to construct the balance sheet for System A's survival.
However, observing this from a physical/thermodynamic perspective, the real deep water hasn't even been tested. They indeed successfully secured the "shells" of TSMC and Intel, but Fabs are highly dissipative physical nodes. You can print fiat to smooth out the CapEx differential, but you cannot print the 10GW baseload power deficit TSMC will face in 2026, nor the upstream photoresist esters (purity), nor the downstream PCBs (where China has locked in 60% of global capacity).
As System C (Silicon Valley/AI) shifts irreversibly from bits to atoms, industrial policy will have to mutate from "filling out Excel subsidy models" into brute-force "physical Hardening" (e.g., forcibly stripping civil grid priority to feed AI, or militarizing critical mineral supply chains). This Hard Reboot has just begun.
I’ve written a series of deep dives on my Substack exploring this structural Hardening and how it reprices underlying hard assets. I would love to hear your thoughts on the "thermodynamic bottlenecks" of this industrial policy.
I only have so many hours in a day, but I will always set at least one of those aside to read these fascinating “hands on” stories about industrial policy design and implementation. A few things jumped out from this episode:
- the enormity of managing $39B with-a-B worth of grants. Damn.
- the importance of talent. The importance of talent. The importance of talent.
- very few people discuss the real questions about opportunity costs. Good to see that issue raised here in a way that makes you appreciate that $100B spent in one big and critical area, means $100B is not available to spend in other big and critical areas. Tough, gnarly choices (one that may eventually haunt us when it comes to data center over-investments).
- whatever mistakes the CHIPS team might have made, the fact that the people on the team were above reproach will be an enduring legacy. Especially when compared to the neoroyalism/nepotism driving many current investment and manufacturing decisions. It’s a third rail too many people want to avoid touching publicly. Yet it’s dominating decisions made today.
- Secretary Raimondo’s reputation will only grow over time.
- Browbeating industry works until it doesn’t. Then it fails badly. Maybe a cautionary tale here about the Pentagon and Anthropic.
Very well done
Brilliant episode and a brutally honest recap of the $39B "KKR-style" capital deployment. This was the necessary first step to construct the balance sheet for System A's survival.
However, observing this from a physical/thermodynamic perspective, the real deep water hasn't even been tested. They indeed successfully secured the "shells" of TSMC and Intel, but Fabs are highly dissipative physical nodes. You can print fiat to smooth out the CapEx differential, but you cannot print the 10GW baseload power deficit TSMC will face in 2026, nor the upstream photoresist esters (purity), nor the downstream PCBs (where China has locked in 60% of global capacity).
As System C (Silicon Valley/AI) shifts irreversibly from bits to atoms, industrial policy will have to mutate from "filling out Excel subsidy models" into brute-force "physical Hardening" (e.g., forcibly stripping civil grid priority to feed AI, or militarizing critical mineral supply chains). This Hard Reboot has just begun.
I’ve written a series of deep dives on my Substack exploring this structural Hardening and how it reprices underlying hard assets. I would love to hear your thoughts on the "thermodynamic bottlenecks" of this industrial policy.