X has rebuilt its advertising engine around AI at exactly the moment its core problem is not technology, but trust. For Elon Musk’s social network, advertising is still the only path to the kind of scale investors expect, yet many global brands remain wary of the platform’s content policies and volatility. An AI‑first ad stack promises better targeting, more automation and higher returns — but also more opacity. The real question is whether smarter algorithms can persuade marketers to forgive what’s happened on X over the last three years.
In this piece, we’ll unpack what X actually announced, why it matters for advertisers, how it fits into the broader AI ad boom, and what it means specifically for European users and brands.
The news in brief
According to TechCrunch, X has begun a phased rollout of a completely rebuilt advertising platform that leans heavily on AI. The company says the new system introduces modern “retrieval and ranking” technology designed to improve how ads are matched to users and how campaigns are optimized.
The move comes after X’s merger with Musk’s AI company xAI in 2025, and after several difficult years for its ad business. Market researcher eMarketer estimates that X generated about $2.26 billion in ad revenue in 2025, with a forecast of $2.46 billion in 2026 – still roughly half of Twitter’s ad revenue in 2021, but growing again.
X claims that the new stack will make it easier for marketers to create and control targeted campaigns, while AI handles optimization, placement relevance and more precise targeting. The rollout is described as gradual, with ongoing feature releases and improvements promised.
Why this matters
Rebuilding an ad stack from scratch is not a cosmetic update; it’s open‑heart surgery on X’s only scalable business model. If this works, X can argue that whatever one thinks about its politics, it is once again a high‑performing performance‑marketing machine. If it fails, the narrative that “advertisers have better places to spend their money” hardens into permanence.
The immediate winners are performance‑driven advertisers who are still on X or willing to experiment. AI‑driven retrieval and ranking, if competently implemented, should mean:
- more accurate matching between ad and user intent
- faster learning and optimization across thousands of micro‑segments
- less manual tweaking for small teams with limited expertise
For small and mid‑sized businesses, this could lower the barrier to running sophisticated campaigns, much like Meta’s Advantage+ or Google’s Performance Max have done.
But there are losers and risks. More automation usually means less transparency. Marketers get better numbers in dashboards, but less understanding of why certain users are seeing certain ads. That’s uncomfortable in a world where regulators, civil society and even some brands want more explainability in targeting, not less.
There’s also the unresolved brand‑safety issue. You can build the smartest ranking system in the world, but if the underlying content environment remains chaotic or toxic, the risk calculus for Coca‑Cola, Volkswagen or Unilever doesn’t change. AI can avoid some bad adjacencies; it cannot substitute for a credible, consistently enforced content and moderation policy.
Finally, this move pulls X deeper into the same AI arms race already dominated by Google, Meta and increasingly Amazon. To stand out, it will need not just comparable tooling, but a compelling reason for advertisers to tolerate the platform’s volatility.
The bigger picture
X’s announcement fits squarely into a broader shift: digital advertising is moving from manual control to AI‑orchestrated black boxes. Over the last two years, Google has pushed advertisers into Performance Max campaigns, Meta has leaned heavily on automated campaign types, and even smaller platforms tout “AI optimization” as a default.
The reasons are obvious. AI can:
- dynamically test thousands of creative/targeting combinations
- predict which users are most likely to convert, even with less granular tracking
- reallocate budget in real time across formats and placements
For platforms, this is a way to keep ad performance strong despite privacy changes (like Apple’s ATT) that broke older tracking models. For advertisers, it trades fine‑grained control for better aggregate performance — at least when the algorithms work.
Historically, big rewrites of ad systems have reshaped markets. Google’s Quality Score, Facebook’s shift to News Feed–centric targeting, and the programmatic RTB boom each created new winners and losers among agencies, adtech vendors and brands. X is now attempting a similar leap, but from a position of weakness rather than dominance.
The xAI connection is also strategic. By tightly coupling the social graph, real‑time conversations and proprietary AI models, X is trying to build a vertically integrated stack that can, in theory, personalize both content and ads more aggressively than competitors that separate these layers. That could yield powerful results — or create an even more opaque system that is difficult for regulators and independent auditors to scrutinize.
The European angle
For European advertisers and regulators, the question is not only "does it work?" but "is it compliant, explainable and safe?" X is already designated as a Very Large Online Platform (VLOP) under the EU’s Digital Services Act (DSA), which imposes extra obligations on ad transparency, risk assessments and mitigation of systemic risks like disinformation.
An AI‑heavy ad stack raises several flags:
- Transparency: The DSA requires clear information on why a user saw a particular ad. Deep learning systems are notoriously hard to explain in non‑technical terms.
- Profiling and consent: Under GDPR, granular profiling for advertising typically relies on valid consent. If xAI models infer sensitive traits (political views, health, religion) from behavior on X, regulators will ask tough questions.
- Political and issue‑based ads: With multiple national elections every year and EU‑level votes, micro‑targeting of political messages by opaque AI systems will attract intense scrutiny.
European brands, especially in privacy‑conscious markets like Germany or the Nordics, are already cautious with X. Many have pulled or reduced spend due to concerns about hate speech, disinformation and Musk’s own posts. A promise of “better AI” will not, on its own, overcome those concerns.
On the other hand, Europe’s large export‑oriented SMEs — from Italian fashion labels to Central European B2B manufacturers — are hungry for efficient global performance channels. If X can demonstrate robust brand‑safety controls, third‑party verification and DSA‑aligned transparency, it could slowly rebuild a European revenue base, particularly among smaller advertisers who care more about cost‑per‑lead than headlines.
Looking ahead
The next 12–18 months will determine whether this rebuild is a turning point or a footnote.
Indicators to watch:
- Case studies and benchmarks: X will highlight early success stories. The key question is whether independent advertisers corroborate meaningful lifts in ROAS (return on ad spend) versus Meta, Google or TikTok.
- Third‑party verification: Adoption of brand‑safety and measurement partners (IAS, DoubleVerify, Moat, etc.) will be critical for big brands. If these firms vouch for the new stack, some budgets may return.
- Regulatory friction in the EU: Any new DSA or GDPR enforcement actions related to ad targeting or transparency would be a major red flag for multinational brands.
- Product direction: Does X continue to add advertiser controls, transparency tools and APIs, or double down on "just trust the AI" black boxes?
There are also unanswered questions: How much xAI technology is truly novel versus rebranded standard machine learning? To what extent are user posts and interactions being used to train models, and can users in Europe meaningfully opt out under GDPR? How will X handle political ad targeting in a heated geopolitical climate?
For agile advertisers, there is opportunity in this uncertainty. Early adopters often benefit from lower competition and cheaper inventory. But they should treat X as a high‑beta channel: test with small budgets, demand transparency, and insist on independent brand‑safety checks before scaling.
The bottom line
X’s AI‑driven ad rebuild is strategically necessary and technically rational, but it does not solve the platform’s core issue: a deficit of trust among major brands, regulators and a portion of its user base. Smarter algorithms can improve campaign performance, yet they cannot mask the impact of content policy choices or governance style. If X pairs this new stack with serious investment in brand safety, transparency and regulatory compliance, it has a shot at a second act in advertising. If it treats AI as a shortcut around those issues, advertisers will enjoy the experiments — and keep their main budgets elsewhere.
