Thursday, November 5, 2026

Voters rejected data centers 40 of 41 times, putting AI's compute build-out at risk

Data center measures appeared 41 times on ballots in 10 states, and voters sided with the AI build-out once. Democratic governors-elect in Texas, Georgia and Ohio campaigned on data center regulation. OpenAI leaders reportedly discussed resetting next year's capacity expectations, while Apple and Goldman Sachs are cutting AI spend internally. In agent commerce, vendors are building checkout and identity tooling, but Uber reports almost no one books rides through agents.

Sources read
20
Claims that stand
7
Quotes found in the source
19 of 19
Claims struck
0
Corrected or withdrawn
0

We quote it or we drop it. Every claim below carries the source’s own words, checked against the text we read.

Since 4 Nov

What we told you last time, ruled on from today’s evidence.

  • Gone quiet
    Nothing new today on these. A limit of what we read, not a verdict.

    Google makes Search access conditional on schools enabling AI · EU court adviser says Gemini's reuse of press content may breach copyright · Menlo: enterprise AI deployment spend hits $177B as agent traffic surges · Banks start packaging AI toolkits for small business customers · Solopreneurs use AI to grow before making a first hire · Freshworks crosses $1B ARR with 30% fewer employees · Privacy-focused AI home assistant Ollie shuts down under Big Tech pressure · African tech startups raised $684M in first nine months of 2026

Every correction we have made →

Each mark on a claim’s line is an outlet, placed by when it published: 11:15 to 06:55 UTC.

AI

  1. Voters reject data center build-out in 40 of 41 ballot measures

    In Tuesday's elections, data center bans, moratoriums and rezonings appeared 41 times in 10 states, and the AI build-out won once, in Harwood, N.D.; OpenAI leaders then discussed resetting 2027 data center capacity expectations. Local permitting is now a binding constraint on AI compute, which feeds directly into model pricing, capacity and the growth assumptions behind AI startup valuations.

    On Tuesday the artificial intelligence boom was on the ballot 41 times in 10 states. By Wednesday, with most results in, voters had sided with it just once.

    3 more receipts

    senior leaders at OpenAI were discussing whether the results meant the company would have to reset expectations regarding its data center capacity next year

    Billionaire investors Bill Ackman, Chamath Palihapitiya and Brad Gerstner warned that Democratic opposition to data centers could threaten the artificial intelligence boom.

    Wall Street rallied despite those concerns, with the Dow surging more than 1,500 points and Big Tech stocks adding more than $750 billion in market value

  2. New governors in Texas, Georgia, Ohio ran on data center regulation

    Democratic governors-elect Hinojosa in Texas, Bottoms in Georgia and Acton in Ohio made data center regulation central to their campaigns, and Texas holds about a quarter of planned U.S. capacity. State-level rules in Texas could slow the national compute build-out far more than scattered local moratoriums.

    Governors-elect Gina Hinojosa in Texas, Keisha Lance Bottoms in Georgia, and Amy Acton in Ohio—all Democrats—made data center regulation a key part of their campaigns.

    2 more receipts

    Texas is the big one, accounting for about a quarter of the U.S. data center capacity in the works, according to one recent analysis

    The AI buildout has boosted the stock market, but that growth benefits a smaller, wealthier slice of Americans.

  3. Apple cuts employee AI token budgets as enterprises rein in spend

    Bloomberg, via Reuters, reports that Apple reduced employee budgets for AI tokens to control costs, after Goldman Sachs restricted access to some AI coding tools. Large buyers are moving from unlimited AI usage to metered usage, which pressures AI vendors whose revenue depends on token consumption.

    has reduced employee budgets for AI usage as part of a broader effort to rein in costs

    1 more receipt

    Bloomberg recently reported that Goldman Sachs restricted access to some AI coding tools following a review of productivity and risk management

AI agents that find customers for small businesses and founders

  1. Agent-commerce tooling ships, but Uber reports almost no agent bookings

    Descope launched Envoy so consumer apps can accept personal AI agents as users, and $10B startup Instinct says users link credit cards quickly; Uber says only four people booked a ride via an AI agent in its busiest week last quarter. Founders being told to sell to AI agents should treat agent-driven demand as early and unproven, because real transaction volume is still close to zero.

    Descope Inc. today launched Envoy, a service that lets consumer websites and apps accept personal artificial

    3 more receipts

    Instinct founder Noah Shinn says rapid credit card linking shows trust, as AI agent rivals like Meta's Muse and OpenAI's Dots chase shopping.

    AI agents are increasingly handling purchasing decisions that humans used to make themselves.

    During its busiest week last quarter, Uber said only four people booked a ride using an AI agent.

  2. Small businesses urged to test agents before granting autonomy

    Guidance aimed at small business owners recommends testing AI agents on multi-step customer workflows, such as HVAC service requests, before giving them more independence; MarTech covers agents' growing role in customer experience. For small firms, the practical barrier to agents that handle customers is trust and testing, not raw capability.

    For small business owners, a routine service request can set off several tasks.

    1 more receipt

    Paul Meinshausen, co-founder of Aampe and head of agentic at MoEngage, joins Conversations with MarTech to discuss the evolving role of AI agents in

tech startups

  1. Women TechEU Allies offers €12M in equity-free grants to 160 startups

    The EU-backed Women TechEU Allies scheme gives about €75K each to 160 early-stage deep tech startups with women in top management roles, plus mentoring and a route to EIC support. Equity-free money of this size matters to deep tech founders when venture capital is concentrating in large AI rounds.

    The programme targets early-stage deep tech startups founded or co-founded by women who hold a top management position, such as CEO or CTO.

    2 more receipts

    the scheme offers equity-free money, mentoring and a bridge to further EIC support.

    Women TechEU Allies: 160 Startups, €12M Grants [2026] What €75,000 Buys a Software-Heavy Deep-Tech Startup

  2. Startups face a flood of AI-written job applications

    Sifted reports that startups want AI-literate hires but are being swamped by AI-generated CVs and cover letters, and are changing their hiring processes in response. Hiring signal is degrading for small teams, raising the cost of finding real talent just as headcounts get leaner.

    CVs and cover letters that reek of AI are piling into inboxes. Here’s how some startups are stemming the tide

Why it matters

The AI economy now faces constraints from two directions. Voters are restricting the physical supply of compute, and enterprise buyers like Apple and Goldman Sachs are rationing token spend. Tighter supply and more cautious demand together undercut the assumption that AI usage and capacity only grow, which underpins much current startup financing. The market's rally on gridlock suggests investors are not yet pricing in a slower compute build-out.

For founders betting on AI agents to bring in customers, the gap between infrastructure and usage is the key fact. Identity layers, payment linking and 'sell to agents' playbooks are arriving before consumers actually transact through agents, as Uber's four bookings show. The near-term opportunity is agents that small businesses test and supervise on their own workflows, not agent-driven consumer demand.

What to watch

  • Whether OpenAI or other hyperscalers formally cut or delay 2027 data center capacity targets after the election results.
  • What data center rules Texas governor-elect Hinojosa proposes, and the outcome of the three data center ballot measures still to come this year.
  • Whether agent platforms such as Instinct, Muse or Dots disclose actual transaction volumes that contradict Uber's near-zero figure.

Sources read

SourceOutletPublished, UTCQuoted
Voters rejected Trump’s ‘booming’ economy narrative and the AI that underpins it nbcnews.com05:151×
The Election Day Revolt That Threatens America’s AI Boom wsj.com02:151×
An AI-native experimental laboratory for autonomous biomolecular operations nature.com22:15read, not quoted
University to lead cybersecurity for national Fabric for AI-Driven Science project news.unl.edu06:32read, not quoted
2026 election results show widespread opposition to AI data centers scientificamerican.com23:153×
What Wall Street billionaires are saying about Democrats' midterm sweep cnbc.com23:152×
Apple cuts employee AI budgets in cost-control push, Bloomberg News reports reuters.com20:152×
Selling to AI Agents Requires a Different Playbook. Most Founders Have Not Written It Yet inc.com21:151×
Descope launches Envoy to let consumer apps accept personal AI agents as users siliconangle.com14:151×
Almost no one uses AI agents to book their Uber, company says techcrunch.com22:151×
How businesses can test AI agents before giving them more independence fox26medford.com19:151×
AI is coming for banks' business, but the industry can fight back americanbanker.com12:15read, not quoted
How AI agents help improve your customer experience martech.org18:151×
A $10 billion AI startup is finding out how quickly people will hand over their wallets businessinsider.com19:151×
Meet The Six Startups From RDCW’s Climate Tech Showcase & Social grepbeat.com06:55read, not quoted
New Startups Push to Innovate Strength Training Tech athletechnews.com21:15read, not quoted
Women TechEU Allies: 160 Startups, €12M Grants [2026] shattered.io15:151×
Women TechEU Allies: €12M, 160 Startups, €75K Each [2026] tech-insider.org15:152×
Tech companies tell job applicants: Be AI-pilled (but not too AI-pilled!) sifted.eu11:151×
Japan wants US tech companies to share undisclosed software flaws as risks mount asia.nikkei.com16:15read, not quoted

AI Builder’s Edge

OpenAI says Codex-driven parallel development overwhelmed its monorepo. It built Pando, a Git-compatible backend that cut median push latency by nearly 5x and now serves nearly a petabyte of Git data a day. Builder write-ups on Claude Code credit constraints, tests and small merges for fast results, and show it driving apps never built for AI. On the creator side, ChatGPT now generates editable animated GIFs alongside OpenAI's new Images 2.6 model.

Sources read
6
Claims that stand
4
Quotes found in the source
6 of 6
Claims struck
0

We quote it or we drop it. Every claim below carries the source’s own words, checked against the text we read.

Each mark on a claim’s line is an outlet, placed by when it published: 07:16 to 18:16 UTC.

  1. OpenAI built Pando because Codex broke its Git infrastructure

    OpenAI replaced its monorepo backend with Pando, a Git-compatible system built for concurrent pushes and high read throughput, after Codex-driven load pushed reliability below 90%. Agent-parallel development moves the bottleneck from writing code to source control and CI, and teams scaling coding agents will hit the same load pattern.

    In production, Pando cut median push latency by nearly 5x. It now delivers nearly a petabyte of Git data per day.

    2 more receipts

    One engineer now has several changes ready for CI at once, and each push can set several jobs fetching the same commit.

    We suffered slow fetches and pushes along with failed merges, with reliability at times falling below 90%.

  2. Claude Code site rebuild credits constraints, tests and small merges

    A developer rebuilt a website on Next.js 16 with Claude Code in two days and credits constraints, tests, small merges and scoping decisions, not code volume. It is a concrete, repeatable pattern: fence the agent with constraints and tests, and merge in small increments to keep agent output reviewable.

    I rebuilt my site on Next.js 16 with Claude Code in 2 days. The line count hides what mattered: constraints, tests, small merges, and what to build.

  3. Claude Code used to drive apps not built for AI

    XDA reports using Claude Code to control applications that lack AI integrations, describing it as a new workflow. It extends Claude Code from repo work to automating legacy or closed apps, so builders can script tools that have no AI integration.

    I used Claude Code to control apps that weren't built for AI, and it opens up a whole new workflow

  4. ChatGPT adds editable animated GIFs with Images 2.6

    ChatGPT can now generate animated GIFs with adjustable speed, frame-level editing and sharing, rolling out with OpenAI's Images 2.6 model. Frame-level control inside ChatGPT lowers the effort of producing short-form animated assets, which creators can use for social content.

    The update, which is rolling out alongside OpenAI’s new Images 2.6 model, allows you to adjust the GIF’s animation speed, edit individual frames, and share it with others.

Why it matters

OpenAI's Pando write-up puts a number on the strain coding agents create. When each engineer runs several changes in parallel, pushes, fetches and CI jobs multiply until source control fails. Teams adopting Codex or Claude Code at scale should expect infrastructure, not model quality, to become the constraint. OpenAI kept full Git compatibility, which signals that Git itself survives agent-scale work as long as the backend is rebuilt.

The practitioner reports point the same way. The speed gains come from process discipline, meaning constraints, tests and small merges, and from aiming agents at surfaces that were never designed for them. On the creator side, the gains are coming as features inside general assistants, such as GIF editing in ChatGPT.

What to watch

  • Whether OpenAI upstreams Pando-related changes to Git, as it says it will continue contributing to Git to support agent-scale work.
  • How quickly the Images 2.6 rollout and GIF editing reach all ChatGPT users, and whether frame-level editing extends to longer animation or video.
  • Further detail on Anthropic's Claude update addressing video game cloning, which CNET flagged as minor but which bears on what creators can generate.

Know someone watching this? Pass it on.

20 sources, at least 14 minutes of reading. This briefing took about 3. You’re done.

Counted from the 6 articles we could read in full, at 230 words a minute; 14 more sat behind a paywall or gave us only a headline.