> ## Content Index
> Fetch the complete content index at: https://www.digitalpolitics.co/llms.txt
> Use this file to discover other available public pages before exploring further.

# Why do politicians still use Facebook?
- URL: https://www.digitalpolitics.co/facebook-political-ads-midterms-brussels-effect/
- Published: 2026-10-05T11:00:29.000Z
- Updated: 2026-10-07T11:03:51.000Z
- Description: Ahead of the US midterm elections, I built a digital political ad daily tracker that showed how much campaigns still rely on Meta's platforms.
- Author: Mark Scott
- Tags: #edition, #archive, #democracy-information, #platforms-online-safety, #digital-power-geopolitics, Australia, Brussels, data centers, digital competition, digital markets act, digital regulation, digital services act, disinformation, Donald Trump, elections, European Commission, European Union, information environment, Instagram, Facebook, Mark Zuckerberg, meta, platform governance, platforms, Social Media, United States, US election, US states

**IT'S MONDAY, AND THIS IS DIGITAL POLITICS.** I'm [**Mark Scott**](https://www.linkedin.com/in/markscott82/?isSelfProfile=true&ref=digitalpolitics.co), and you find me in Vilnius for the annual EU DisinfoLab [conference](https://www.disinfo.eu/conference/?ref=digitalpolitics.co). I'll also be in Brussels on Oct 8 to moderate this [panel discussion](https://www.atlanticcouncil.org/event/transatlantic-digital-relations-sovereignty-security-and-competitiveness/?ref=digitalpolitics.co) on the transatlantic digital relationship (there are a few seats left if you're in town.)

**— The US midterm elections on Nov. 3 explain** why politicians worldwide still use Meta's Facebook and Instagram as campaign infrastructure to reach voters.

**— The Brussels Effect was supposed to export** the European Union's regulatory vision globally. But there's one thing to write EU-style rules. It's another to implement them.

**— Electricity usage associated with** artificial intelligence is expected to more than double, globally, over the next decade. 

**Let's get started.**

---

## HOW FACEBOOK CORNERED THE MARKET ON ONLINE POLITICAL ACTIVITY

**IT'S BEEN A WHILE SINCE I DELVED INTO** [political ads](https://www.digitalpolitics.co/newsletter057free/) on social media (it used to be my [bread](https://www.politico.com/news/2020/10/28/union-group-spending-facebook-swingstates-433457?ref=digitalpolitics.co) and [butter](https://www.politico.com/news/2020/03/06/stealth-political-ads-flourish-on-facebook-122539?ref=digitalpolitics.co).) But with less than a month to go before the midterm elections across the United States, I figured it was time to wade back in. 

**That's why I created the Digital Politics** [Facebook Ad Dashboard](https://www.digitalpolitics.co/midtermelections/) (US midterm election edition.) 

**It tracks, daily, how much money** is going into political, election and issue-based ads on Instagram and Facebook nationwide. It also offers granular details on "toss-up" races, as defined by the [Cook Political Report](https://www.cookpolitical.com/?ref=digitalpolitics.co), across the US Senate, House of Representatives and Governor elections. Want to know who's spending the most in a specific swing seat or which demographics a candidate is targeting? Click into the race and it's all there. 

**The data includes how California dominates overall ad spend** — at 10.2 percent of the total. How the top 10 US states, by ad spend, represent 54.3 percent of the overall money earmarked for November. And how **James Talarico**, the Democratic Party candidate for Texas' Senate seat, is currently the overall top ad buyer on Meta (just ahead of a Union-linked dark money operation.)

**Those "toss-up" election figures don't include** external groups like Super PACS, so it's not the full picture. That's by design. I wanted to focus solely on what the candidates, themselves, were spending via their official Facebook and Instagram accounts, mostly to avoid the messiness of figuring out which elections external groups were supporting or attacking.

**OK, that's the headline announcement.** If you want to chat through how I did this, please [reach out](mailto:digitalpolitics@protonmail.com). The team over at [WhoTargets.me](https://whotargets.me/en/?ref=digitalpolitics.co) also does a great job of tracking these social media ads (also beyond Meta.)

Thanks for reading Digital Politics. If you've been forwarded this newsletter (and like what you've read), please sign up [here](https://www.digitalpolitics.co/newsletter058/#/portal/). For those already subscribed, reach out on digitalpolitics@protonmail.com

**So why do this? Honestly, I wanted to see if I could do it.** But the less geeky question was to answer a simple question: In a world where Facebook is becoming less relevant by the day, do politicians still see much use in paying Meta for the privilege of targeting users with these party-political messages? With people polarized more than ever on social media — and artificial intelligence-based content flooding these platforms — does Facebook even matter anymore?

**On those questions, I have an emphatic answer:** yes. 

**Based on the dashboard I created,** the collective money spent across Meta's platforms from mid-June (when I started pulling the figures) currently totals just under $250 million. For other countries' elections, that would represent an eye-watering number. But projections from AdImpact suggest digital advertising — across Meta, Alphabet, Snap and X — [will collectively hit](https://adimpact.com/blogs/adimpact-reveals-2026-election-cycle-to-reach-record-11.6-billion-in-ad-spending?ref=digitalpolitics.co) $1.6 billion by Nov. 3.

**That, in itself, is bonkers money.** But then you realize digital will only represent 14 percent of the overall political advertising spend in this year's US midterm elections. Traditional TV broadcasting (at 48 percent of the estimated $11.6 billion total) and smart TV/streaming services (at 23 percent of the total) still outmuscle the tech bros of Silicon Valley. 

**Yet these top-line numbers only tell half the story.** The fact that campaigns spend money on social media is not new, especially in 2026\. The question of why they spend it on these platforms — compared to other routes — is what interests me.

**On that point, it's worth remembering social media** [is not really "social"](https://www.digitalpolitics.co/newsletter068/) anymore. In Meta [documents](https://storage.courtlistener.com/recap/gov.uscourts.dcd.224921/gov.uscourts.dcd.224921.627.0%5F2.pdf?ref=digitalpolitics.co) released in connection to a US Federal Trade Commission lawsuit last year, the company admitted just as much. "Only a fraction of time spent on Meta’s services – 7 percent on Instagram, 17 percent on Facebook – involves consuming content from online “friends” (“friend sharing”)," according to the document. "Today, more than half of the usage on Facebook and Instagram is unconnected content."

**That's troubling news for politicians.** As much as they can build up audiences on these platforms, if Meta's recommender systems are not filling followers' feeds with such content, then does it even count? In 2026, having a social media presence — and even buckets of followers — is not what moves the needle. It's the ability to target users with advertising, as well as specific "viral" moments driven by social media algorithms, that truly count for how candidates can reach would-be supporters.

**Well, that's my theory anyway:** Meta's shift to recommender systems designed to display unconnected content — in the name of engagement — tacitly forces politicians (and others) to pay for direct engagement these groups once got for free. That makes Meta (and other platforms) essential digital campaign infrastructure that also fattens these companies' profits. 

**To test it, I chose two specific midterm elections** — [Colorado's 8th Congressional District](https://en.wikipedia.org/wiki/Colorado%27s%5F8th%5Fcongressional%5Fdistrict?ref=digitalpolitics.co) (the Fighting Eighth!) and [Arizona's 6th Congressional district](https://en.wikipedia.org/wiki/Arizona%27s%5F6th%5Fcongressional%5Fdistrict?ref=digitalpolitics.co) (the Fighting Sixth!) — and pulled data for the public posts and Meta ads for each of the Democratic and Republican Party candidates to see how these politicians were cutting through the noise (and recommender systems). 

**One note: I make no judgement about the** candidates' political positions or the outcome of either race. These elections were chosen because they are both toss-ups. Each race also has a Republican incumbent to provide as close to an apples-to-apples comparison as possible.

**To keep this straightforward, I combined the engagement** (likes, comments, shares) from all four politicians' public posts between Sept 14-29, see the chart below. I then pulled the data (spend, number of ads, viewer impressions, demographics, and location) from each candidate's Meta ads over the same period, see below. 

**It's not a like-for-like comparison.** Engagements (how people engaged with public content) and viewer impressions (how often ads were displayed to Meta users) are not the same. But it's the closest I could get without each campaign handing me their internal data.

![](https://storage.ghost.io/c/df/80/df8046d2-bf27-4315-9b97-ced0ef736260/content/images/2026/10/public-post-performance-1.png)

![](https://storage.ghost.io/c/df/80/df8046d2-bf27-4315-9b97-ced0ef736260/content/images/2026/10/candidate-paid-advertising--1--1.png)

Source: Meta Ad Library

**What's interesting is that almost all the public posts**, from all candidates, performed the same. They all got less than 500 total engagements per post. The outlier was Manny Rutinel, the Democratic Party challenger in Colorado, after one of his posts ([of him working out at the gym](https://www.facebook.com/reel/1771136310864382/)) went viral. But when you strip out that recommender system-boosted content, Rutinel's per-post average engagement leveled out to where the others' figures were.

**Those are so-so figures for toss-up elections**, though not exactly a sign of overwhelming organic traction with voters online.

**That's where the ads came in.** Three politicians bought fewer than 10 ads, with overall impressions varying between 180,000 to 2.1 million, based on Meta's data. But, again, Rutinel — the only politician with a 'viral' moment — bought the most ads over the four candidates to reach the most number of people (4.2 million). 

**So why did the politician with the most reach** in his public posts also pay for the most paid-for messages? Part of that is likely down to Rutinel's campaign strategy, whose ads were delivered to users outside of Colorado. But another part is almost certainly down to what Facebook has now become.

**Yes, viral posts can allow politicians to break out** of Meta's algorithm-driven cage. There is no way, however, of predicting when that will happen, and for what reason (good or bad.) Ads, therefore, remain the more reliable way for campaigns to put a message in front of voters in the era of personalized, recommender-driven feeds. It explains why politicians spend money on Facebook and Instagram, even as organic engagement becomes harder to retain. 

---

## Chart of the Week

**Artificial intelligence is now all about** electricity**.** The International Energy Agency [expects](https://www.iea.org/data-and-statistics/charts/electricity-consumption-by-data-centres-2020-2035?ref=digitalpolitics.co) global data-center electricity consumption to rise from 485 terawatt-hours (TWh) in 2025, or about 46 million US homes, to 1,193 TWh by 2035.

**That does not mean AI alone explains every extra electron.** 

**But it does mean the next phase of AI policy** will not be decided only by model safety rules, copyright lawsuits or competition policy. It will also be framed around who can build this electricity generation infrastructure at a time when global warming is going from bad to worse.

![](https://storage.ghost.io/c/df/80/df8046d2-bf27-4315-9b97-ced0ef736260/content/images/2026/10/ai-electricity-story--1--1.png)

Source: [International Energy Agency](https://www.iea.org/data-and-statistics/charts/electricity-consumption-by-data-centres-2020-2035?ref=digitalpolitics.co)

---

## WHAT NO ONE SAYS ABOUT THE BRUSSELS EFFECT

**JUST LIKE CITY BUSES, I HAVE TWO DATA-HEAVY** trackers for you this week. The [second](https://www.digitalpolitics.co/tracker/) is based on the [monthly synopsis](https://www.digitalpolitics.co/digital-policy-regulatory-tracker-september-2026/) of global regulatory, legal and legislative digital policy actions, across 40 jurisdictions. The database is now live. It's built to answer a specific type of question like what's the latest filing in the New Mexico Meta lawsuit, or what Canada's competition regulator has said about a particularly platform. It points you at the source document, not a secondhand summary. 

**As I waded through this corpus** — one that still has kinks and skews overwhelmingly to Western jurisdictions (something I'm fixing) — one message came out loud and clear: the EU is way ahead on implementing its comprehensive digital rulebook compared to other jurisdictions.

**The numbers don't lie. Of the more than** 350 entries in the Digital Politics Tracker, just over 40 percent come from the 27-country bloc. Within that Europe-only subset, only a few are actual legislation. Most are compliance filings, secondary guidance documents and third-party audits that represent the mundane work underpinning the bloc's digital regulatory agenda.

**I don't say any of this in praise of Europe.** 

**As I put this database together, the litany of EU documents** made clear what few are willing to say out loud. It relates to the [Brussels Effect](https://en.wikipedia.org/wiki/Brussels%5Feffect?ref=digitalpolitics.co), or effort by Europe to nudge companies to adapt to the Western world's largest high-standard market, as well as provide other capitals a regulatory playbook that they could adopt for themselves.

**But the Digital Politics tracker points** to an uncomfortable truth: a regulatory template is only useful if a country can sustain the institutions, expertise and administrative work needed to make it stick. Few, if any, other countries combine the administrative muscle, fiscal resources and political appetite to match the EU's ever-growing volume of digital compliance work across so many policy areas. 

**There are some jurisdictions, like the** United Kingdom and Australia, that can go toe-to-toe with Europe in some areas. That's especially true on child safety issues where [Canberra](https://www.esafety.gov.au/newsroom/media-releases/tech-giants-to-face-uncomfortable-truths-as-esafety-compels-them-to-lift-their-game-tackling-child-abuse?ref=digitalpolitics.co) and [London](https://www.ofcom.org.uk/online-safety/illegal-and-harmful-content/online-safety-regulatory-documents?ref=digitalpolitics.co) have a track record of producing guidance, risk assessment obligations and other oversight that would make even the most ardent Eurocrat blush.

**But that's one (albeit important) digital policy** area. Where Brussels turns it up to 11 is that it produces the same widespread corporate obligations, regulatory guidance and enforcement actions across data protection, consumer protection, online safety, digital antitrust and artificial intelligence. That is just not something other countries are capable (or willing) to do.

## Sign up for Digital Politics

Thanks for getting this far. Enjoyed what you've read? Why not receive weekly updates on how the worlds of technology and politics are colliding like never before. The first two weeks of any paid subscription are free.

Subscribe 

Email sent! Check your inbox to complete your signup. 

No spam. Unsubscribe anytime.

**Now, I'm not saying the EU is wrong in its approach.** Regulators are going to regulate, amirite? But it does pose a problem for how Brussels envisions itself on the global stage. If the goal is to export its legislative and oversight vision to others, then the logical step is that the administrative burden that goes along with the Brussels Effect must also be exported.

**Again, this is not a fan-mail to Mario Draghi** and his deregulatory agenda. I am a firm believer that digital oversight, if based on evidence and proportionate to the potential harm, makes people safer and (dare I say it?) wealthier. 

**But Europe needs to take a hard look at itself** in the mirror with the ongoing regulatory guidance, compliance obligations and enforcement actions — across multiple digital policy areas, and often not coordinated. That may be how Brussels and (some) EU capitals want to go about their business. But if the secondary goal is to convince other countries to follow Europe's lead, this avalanche of regulatory paperwork, even well-intentioned, is not going to win hearts-and-minds on the global stage. 

**If Brussels wants its model to travel, exporting rules** is not enough. It also needs to export capacity: regulator-to-regulator training, shared technical tools, implementation guidance and sustained assistance for the institutions expected to enforce such rules.

---

## What I'm reading

**— Anonymizing prompts does not necessarily protect AI users’ identities,** as providers may reconnect supposedly private conversations to a person’s previous interactions, [finds](https://arxiv.org/pdf/2609.33903?ref=digitalpolitics.co) new research from the University of Massachusetts Amherst.

**— Europe's data-access rules risk becoming an administrative fight** rather than a research tool, based on an non-profit's experience of seeking information from major platforms, [writes](https://dsa-observatory.eu/2026/10/02/two-stories-of-dsa-data-non-access-from-an-ngo-perspective/?ref=digitalpolitics.co) Oliver Marsh.

**— Brazil’s Supreme Court has moved platform liability toward a more preventive, risk-based model,** with implications for digital governance beyond the country, [argue](https://tsjournal.org/index.php/jots/article/view/308?ref=digitalpolitics.co) Beatriz Kira and Ivar Hartmann.

**— US President Donald Trump and six leading** AI companies endorsed a voluntary “White House Accord on Super Intelligence” that promises internal controls, independent audits and board oversight for frontier models, according to a [Truth Social post](https://truthsocial.com/@realDonaldTrump/posts/117356435739432952?status%5Fid=117356435739432952&attachment%5Fid=117356434517531231&ref=digitalpolitics.co).

**— The UK needs a dedicated AI law and regulator** to address gaps in protections for equality, privacy and redress, [argues](https://publications.parliament.uk/pa/jt5902/jtselect/jtrights/160/report.html?ref=digitalpolitics.co) the UK Parliament’s Joint Committee on Human Rights.

---