Gutenberg Lab — Morning Paper

What comes after the data center backlash?

Featuring: Shayle Kann (Energy Impact Partners) × Brian Janis (Co-founder & Chief Commercial Officer, Cloverleaf Infrastructure)

Bottom line: The data center backlash is not a technical problem but a values fight over AI itself, and you cannot design your way out of it.

3-line summary

The data center industry faces a bifurcated crisis: solvable technical grievances (power rates, water use, infrastructure costs) versus an intractable values fight over AI and Big Tech. Even when developers address every legitimate community concern—funding grid upgrades, lowering rates, minimizing water—opposition persists because it is rooted in distrust of AI, not the facilities. The resolution of this tension will determine whether the U.S. can build AI infrastructure at scale or cedes leadership to other countries.

Key takeaways

1. Emotion vs. facts: Janis notes, "The industry has tried to attack this with facts and really miss the emotional resonance of, like, people just aren't happy about this stuff right now." Even when transparent, funding infrastructure, and lowering rates, communities respond "I just don't believe you" or "I don't like Meta." This is a values fight, not a technical one.

2. Micro good, macro bad: Data centers lower rates locally (utilities structure deals to guarantee this) but raise rates nationally (supply chain inflation, equipment shortages, higher gas demand). Both are true simultaneously, making the politics intractable: communities rationally want data centers nearby but also rationally believe they raise prices overall.

3. Elections create multi-year headwind: Midterms and the 2026 presidential election make data centers a wedge issue. Both parties label opponents as "pro data center," and it is working. Janis says, "That's a playbook that others are going to try to use." This political incentive structure persists through the early 2030s. "Everyone who thought, oh, I just got a constraint in 2027 or 2028, you have the same constraint in '29 and '30 and possibly into the early 2030s."

Editorial perspective

The data center backlash is the latest iteration of "technology anxiety" that recurs at every platform transition. Just as the rise of the internet sparked fears of monopoly and control, and the App Store economy triggered alarm over Apple's dominance, the AI economy is now generating emotional rejection of Big Tech. The difference this time is that physical infrastructure (power, water, land) is involved, so opposition movements directly intersect with local politics and become election issues. As Janis points out, this is a values fight that facts cannot resolve, and the headwind will persist until AI delivers visible results in medicine or materials science. The industry's failure to sell AI as transformative technology—rather than "chatbots and slop memes"—will constrain U.S. infrastructure buildout into the early 2030s.

Source: Catalyst, "What comes after the data center backlash?" (September 3, 2026)

https://www.latitudemedia.com/podcasts/catalyst

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

• --

Physical AI Funding Hits $50B in H1 2026 as VCs Abandon "Don't Do Hardware" Dogma

Guests

Jason Calacanis (host) × Caitlin Holloway (Founding Partner, 776) × Paige Dougherty (Founding Partner, Behind Genius Ventures)

Bottom line

The Physical AI funding surge is not a bubble but a structural shift driven by top engineers choosing rocket engines over ad tech—talent migration preceded capital, not the other way around.

3-Line Summary

Physical AI raised nearly $50B in H1 2026 (up 80% YoY, 450 deals), exceeding 2022-2024 combined. The decade-old VC mantras of "don't do hardware" and "don't sell to government" collapsed as launch costs fell 97% (from $10K/kg in 2014 to projected $273 by 2040) and the Pentagon deployed ChatGPT/Grok to 3M personnel. The defining shift: top engineers now choose "meaningful work" building rocket engines in Moses Lake, Washington over optimizing ad click-through rates—talent moved first, capital followed.

3 Key Points

1. Orbital data centers went from $150M to $2.3B in five months: 776-backed StarCloud achieved a 15x markup after putting the first H100 chip in orbit last year. NVIDIA is now designing space-specific chips based on StarCloud's flight data, validating the "infrastructure for the inevitable" thesis of investing upstream (suppliers to labs) rather than downstream (applications). Manhattan West led the $250M round with NVIDIA and Cisco participating.

2. Pentagon deployed ChatGPT and Grok to 3M personnel: ChatGPT MIL (paperwork/logistics) and Grok for Government (supply chain analysis) rolled out across military workforce. Claude excluded after Anthropic refused to modify usage restrictions for military applications. The government sales bottleneck shifted from "winning the contract" (multi-year cycles) to "executing with a 30-person company"—execution capability became the moat.

3. Anthropic reversed data policy under enterprise pressure: Facing consensus belief that frontier labs will use customer data to build competing products (Cursor → Claude Code, Figma → Claude Design), Anthropic implemented 30-day retention windows and customer cloud environments. Legal tech firm Litera and Launch portfolio company Advocates are building vertical LLMs to avoid dependency, echoing the "be your own bitch" platform risk lesson from the social media API era.

Editorial Perspective

This mirrors the App Store platform transition. In 2008, talent moved before capital—the best engineers chose iOS development before the venture dollars arrived. Today, the same pattern: top engineers are choosing rocket engines over ad optimization, and the $50B in capital is following that human migration. When launch costs drop 75% in a decade (from $10K/kg to projected $1,500 by 2030), orbital data centers shift from science fiction to economic rationality. The challenge for early-stage investors is maintaining ownership through billion-dollar fundraising paths. Verticalized applications (factory quality assurance AI, metal servicing software) offer faster commercialization and less dilution than general-purpose robotics—competing on speed to revenue rather than capital scale.

Source

Source: This Week in Startups "VC experts on why Physical AI funding is heating up | E2333" (September 2, 2026)

https://4a885955-6823-4b22-a3e1-a526c25516a5.libsyn.com/vc-experts-on-why-physical-ai-funding-is-heating-up-e2333

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

• --