Gutenberg Lab — Morning Paper

Why US Electric Utilities Can't Innovate — Five Barriers to Becoming Tech Companies

Participants

David Roberts (Volts podcast host) × Quinn Nakayama (Innovation lead, PG&E) × Hannah Green (Energy go-to-market lead, Microsoft; formerly at Pice)

Bottom line

US electric utilities face surging demand and soaring rates where innovation is the only math that works, yet they spend just 0.2% of revenue on R&D — still nowhere near the 1% target set 34 years ago.

3-Line Summary

• US electricity demand is rising for the first time since the 1970s (driven by data centers, EVs, electrification), requiring $250 billion in annual capital spending, while residential rates have jumped 40% since 2021 and political tolerance for further increases has evaporated
• Utilities have operated as "pipes and wires companies" for 100 years, requiring project management and civil engineering skills, but now must become technology companies needing CTOs, CPOs, and product managers — a complete organizational transformation
• Technologies arrive "60-70% baked" and require utility subject matter expertise and data to reach 100%, but utilities lack product development mindset and test-and-reject rather than commit to co-development

Three Key Points

1. Most smart meters are "flip phones": AMI 1.0 meters rolled out 22 years ago can only read meters and maybe voltage — they can't run sophisticated apps. AMI 2.0 meters (measuring voltage at 32 kHz resolution, app-capable) exist at only "a handful" of utilities nationwide

2. 70% of pilot programs never scale: Technologies arrive 60-70% complete and need utility SME and data to reach 100%, but utilities lack product development mindset. Current approach: test, find reasons it doesn't work, reject, "talk to you in four years"

3. Scaling requires demonstrating ROI within one financial year: Every functional area feels 30-50% underfunded, and VPs/directors won't accept additional budget haircuts unless technology pays off within a year. This is why pilots die — can't show fast enough financial return

Editorial Perspective

The utility innovation stagnation isn't about immature technology — it's an organizational structure problem. Organizations that functioned as infrastructure companies for 100 years are being asked to become tech companies overnight. This isn't just a skill shift; it requires simultaneous transformation of all five elements: strategy, structure, people, process, and technology. In demand-surge regions like California, political blowback against rate increases is creating pressure to defer capital spending, paradoxically generating motivation for innovation. But at the national level, many executives remain "drunk" on data center capital spending, suggesting this inflection point may still lie ahead.

Source: Volts "Why can't utilities innovate?" (August 21, 2026)

https://www.volts.wtf/p/why-cant-utilities-innovate

AI Disclosure: Produced with AI assistance; facts and analysis reviewed by our editorial team.

• --

Treasury Buybacks, Trump Speech, and SEC Rules Fire Triple Catalyst for Crypto Rally

Featuring

David Hoffman × Ryan Sean Adams (Bankless co-hosts)

Bottom line

Wednesday, August 19, 2026 marked a potential inflection point as three catalysts fired simultaneously — the US Treasury doubled bond buyback capacity, Trump endorsed crypto projects at the White House, and the SEC released a 402-page proposed rulebook — driving Bitcoin and Ethereum to their largest single-day gains since the "10/10" crash.

3-Line Summary

• Bitcoin hit $72,300 (up 14.5% on week), Ethereum $2,330 (up 23%), driven by liquidations including one account losing $30M shorting ETH
• US Treasury doubled bond buyback cap from $2B to $4B, pushing 30-year yields from 5.3% to 5.1% (rebounded to 5.24% Thursday morning). Gold and silver added $1.3 trillion in market cap
• SEC proposed three token issuance exemptions (Startup: $5M cap; Reg A: $20M-$75M/year; Rule 400 safe harbor) and standardized disclosures, delivering regulatory clarity without Congressional legislation

Three Key Points

1. Treasury QE Mechanics: Treasury Secretary Bessent's bond buybacks are not "pure QE" (no Fed balance sheet expansion) but swap 30-year bonds for T-bills, effectively increasing money supply velocity. Raoul Pal calls this the "Bessent Put" (5% yield ceiling). Markets reacted as if it were QE — gold/silver/Bitcoin rallied while AI stocks fell 2%

2. Trump's Hyperliquid Name-Drop: At a White House event attended by Coinbase CEO Brian Armstrong, Robinhood CEO Vlad Tenev, Kraken CEO Arjun Sethi, Ripple CEO Brad Garlinghouse, Chainlink founder Sergey Nazarov, the Winklevoss twins, and a16z's Chris Dixon, Trump explicitly said "Hyperliquid" on live TV. Hyperliquid gained 25% on the week

3. SEC's Three Exemptions: (1) Startup exemption: raise up to $5M, no accredited investor limits, no resale restrictions; (2) Reg A for tokens: Tier 1 = $20M/year (unaudited), Tier 2 = $75M/year (audited); (3) Rule 400 safe harbor: token deemed non-security if issuer ceased "essential managerial efforts" and files certification. David: "This does away with airdrops, points programs, yield farming — all the convoluted workarounds"

Editorial Perspective

Bessent's bond buybacks are not "pure QE" but achieved the same market effect by swapping long-dated bonds for T-bills, which markets treat as money-like instruments. Yet David's hesitation — "I don't know if a short squeeze is the confirmation I want" — captures the core uncertainty: is this the start of a sustained bull market, or the final liquidation before capitulation? The answer hinges on two variables over the next few weeks: (1) whether ETF inflows and spot volumes confirm durability beyond positioning unwind, and (2) whether the AI credit bubble (data center buildout on debt) blips like 2021's blockspace oversupply, pulling crypto down despite debasement tailwinds. The SEC delivered clarity without Congress, but as David notes, "No one wants new L1s anymore" — regulation arrived just as appetite for what it enables evaporated.

Source: Bankless "ROLLUP: Is the Bull Market Back? | Treasury QE | Trump Pumps Crypto | SEC Token Rules" (August 21, 2026)

http://podcast.banklesshq.com/

AI Disclosure: Produced with AI assistance; facts and analysis reviewed by our editorial team.

• --

Automation Creates Jobs, Not Unemployment — The Case Against Humanoid Hype

Featuring: David (host, Bankless) × Shahin Farshchi (General Partner, Lux Capital)

Bottom line: The humanoid robot hype is economically motivated, not practically grounded—the real automation revolution is happening with specialized robots optimized for specific industrial and commercial tasks.

3-Line Summary

Shahin Farshchi of Lux Capital challenges the humanoid robotics frenzy, arguing that the true automation revolution is unfolding in industrial and commercial settings with specialized robots. AI integration and hardware cost deflation are democratizing automation for small and mid-sized businesses that previously couldn't afford it. Crucially, automation creates higher-quality jobs rather than destroying employment—a historical pattern that contradicts popular fears.

3 Key Takeaways

1. Automation increases employment, not the reverse: Historically, economies with MORE automation have LESS unemployment. Farshchi's portfolio company Formic serves customers who automate because they "can't find workers"—and post-automation, they create more jobs that are higher quality with lower churn. "The enemy of jobs and automation is not the robots. It is the cheap labor."

2. Robotics is in its early-1990s internet equivalent: Just as 1980s computers were defined by hardware specs and the 1990s internet abstracted compute into a capability, robotics is now transitioning from being discussed as physical machines to being understood by their outputs. This abstraction signals mass proliferation has begun.

3. AI democratizes programming, not just intelligence: What required "many PhDs, millions of dollars, many years" can now be achieved by undergrads or high schoolers downloading open-source models (e.g., Physical Intelligence) and fine-tuning them. The bottleneck was never hardware—it was the engineering cost of integration.

Editorial Perspective

The humanoid hype reflects an economically convenient narrative—one universal solution for all problems—rather than engineering reality. History teaches us that specialized tools outperform generalized ones. The 1970s automotive welding arm and the Amazon warehouse shuttle share almost nothing except metal, plastic, and chips, yet both excel because they're optimized for their contexts. The real revolution isn't in form factor but in accessibility: AI-enabled programming and hardware cost deflation are opening doors for small and mid-sized businesses that couldn't afford automation before. This democratization, not the humanoid chassis, is what makes the current moment analogous to the early internet.

Source: Bankless "Bullish on Automation and Robotics, but not Humanoid Robots" (August 31, 2026)

http://podcast.banklesshq.com/

AI Disclosure: Produced with AI assistance; facts and analysis reviewed by our editorial team.

• --