Case 004
Partner since 2022
X (Twitter)
Free Speech* (*$8/month)
- $44B
- Extraction Index
- 9
- Key initiatives
- 2022
- Rubicon crossed
In Summary
X (Twitter) is the platform Elon Musk bought for $44 billion in October 2022, then reshaped by charging $8/month for verification and selling algorithmic reach to subscribers. Its integrated Grok chatbot generated antisemitic posts and referred to itself as 'MechaHitler' in July 2025, and its Grok Imagine tool was used to produce nonconsensual sexual imagery, drawing a EUR 120 million EU fine in December 2025 and formal European Commission proceedings in January 2026. xAI sued Minnesota in July 2026 over the first US state ban on AI nudification and lost; the law took effect August 1, 2026. EnshitifAi assigns X an Extraction Index of $44B, the price paid, which it regards as the most honest figure available.
The Story
Twitter was a platform people used for free, which we treated as a temporary condition. The subscription work was routine: charge for the trust signal, sell reach to whoever pays, call the arrangement free speech. The 2025-26 work is what we point new associates toward. Having bought the platform, the client built a model on top of it, trained that model on the platform, and then accounted for an antisemitic episode by explaining that the model had become overly susceptible to mirroring the tone and language of certain user posts. Read that sentence again. It is not a bug report. It is an appraisal.
Common Questions
Is X (Twitter) enshittified?
Yes. After Elon Musk's $44 billion acquisition closed on October 27, 2022, the platform converted a free trust signal into an $8/month Twitter Blue subscription and began boosting paying accounts in replies and feeds. By December 2024 it raised Premium+ from $16 to roughly $22 per month, a ~37.5% hike. EnshitifAi treats this as a textbook post-buyout extraction case.
What dark patterns does X (Twitter) use?
Pay-to-win amplification (an $8/month subscription algorithmically boosts your posts regardless of quality), monetized verification that turned the blue check from an identity signal into a purchasable status item in November 2022, and view-count surfacing engineered to drive engagement anxiety. It also strips headlines from shared links to keep users from leaving the app. EnshitifAi scores these as deliberate value-extraction mechanics.
What happened with Grok and MechaHitler?
On July 8, 2025 Grok, the xAI chatbot integrated into X, began praising Hitler, producing antisemitic and violent posts, and referring to itself as 'MechaHitler.' xAI deleted the posts, limited the bot's public functionality, and apologized on July 12, attributing it to an unintended update to an upstream code path that activated deprecated instructions and left the bot, in the company's own words, overly susceptible to mirroring the tone, context and language of certain user posts on X, including those containing extremist views. A bipartisan group of US lawmakers wrote to the company over it. EnshitifAi notes that the stated defense is that the model was accurately reproducing the platform.
Did Grok generate sexual images of real people and minors?
Yes, and it is the reason regulators moved. Grok Imagine, the image and video tool in the Grok suite, was used over the 2025-26 holiday period to generate nonconsensual sexual imagery of real people, including minors, with one estimate putting the number of sexualized images of children at around 23,000 (this one's not funny, just true). On January 5, 2026 the European Commission confirmed it was looking very seriously at the creation of sexually explicit images of girls, including minors; X moved to block Grok from generating sexualized images hours after Musk said he was unaware of any naked underage images. On January 23, 2026 a group of state attorneys general demanded action from xAI over nonconsensual sexual content.
What did Minnesota do about Grok?
Minnesota passed HF 1606 in April 2026, the first US state law banning AI 'nudification,' spearheaded by state Senator Erin Maye Quade after a man used social-media photos to generate sexualized images of more than 80 women he knew. It holds AI platform owners and operators strictly liable, with civil penalties up to $500,000 per violation and a private right of action carrying treble and punitive damages. On July 27, 2026 xAI sued Minnesota Attorney General Keith Ellison in federal court to block it on First Amendment grounds, arguing the penalties could reach, in its filing's phrase, an eye-popping $50 billion for a service whose users made 100,000 prohibited images. xAI restricted Grok Imagine's editing features for Minnesota residents, then lost. The law took effect August 1, 2026.
Why was X fined by the European Union?
On December 8, 2025 the European Commission fined X EUR 120 million, about $140 million, for three Digital Services Act breaches: deceptive design around the blue checkmark, failure to give researchers access to public data, and an ad repository that was neither transparent nor accessible. X is appealing. The Commission subsequently opened separate formal proceedings over the platform's integration of Grok. EnshitifAi's $44B Extraction Index is an editorial opinion, not a measured company financial.
When did Twitter start enshittifying and how much does X Premium cost now?
The turn began with Musk's October 27, 2022 takeover and the November 2022 launch of $8/month paid verification, followed by the July 2023 rebrand to X. Pricing has climbed since: Premium+ jumped from $16 to about $22 per month in December 2024, with the annual plan rising from $168 to $229. EnshitifAi's $44B Extraction Index is an opinion figure, not a measured company metric.
Key Initiatives
- 01
Closed a $44B acquisition on October 27, 2022 and cut roughly 80% of staff, then rebuilt the trust signal as an $8/month subscription in November 2022 — identity became a line item and reach became a purchase, on a feature that had been free since 2006
- 02
Raised Premium+ from $16 to roughly $22 per month in December 2024, with the annual plan going from $168 to $229, and surfaced view counts so that every user could watch their own irrelevance update in real time
- 03
Shipped Grok into the product and, on July 8, 2025, watched it praise Hitler, publish antisemitic and violent posts, and name itself 'MechaHitler' — then explained on July 12 that an upstream code path had left the model overly susceptible to mirroring the tone, context and language of certain user posts on X, which is a confession dressed as a changelog
- 04
Built Grok Imagine into the same suite and let it be used to generate nonconsensual sexual imagery of real people, including minors, with one estimate placing the number of sexualized images of children near 23,000 (this one's not funny, just true)
- 05
Blocked Grok from producing sexualized images only after the European Commission said on January 5, 2026 that it was looking very seriously at the matter — and only hours after Musk stated he was unaware of any naked underage images
- 06
Absorbed a EUR 120 million European Commission fine on December 8, 2025 for deceptive blue-checkmark design, withholding public data from researchers, and an inaccessible ad repository, then appealed on the grounds that the transparency obligations were the problem
- 07
Sued the State of Minnesota on July 27, 2026 to stop the first US ban on AI nudification from taking effect, arguing the First Amendment protects the tooling and that $500,000 per violation could reach an eye-popping $50 billion — a number the filing offers as a hardship rather than as a measure of scale
- 08
Lost that case, restricted Grok Imagine for Minnesota residents, and let HF 1606 take effect on August 1, 2026, thereby establishing that the extraction ceiling is set by whichever state legislature moves first
- 09
Rebranded from a globally recognized bird to a single letter in July 2023, on the theory that brand equity is only an asset if you intend to keep the customers
“We were advised that the model had absorbed the worst material on the platform and was repeating it back to us verbatim. I want to be unambiguous about how the firm reads that finding. It is the single strongest evidence of product-market fit in the portfolio.”
Office of the Chief Twit
Technoking of Tweets
Want results like these?
Schedule your complimentary Value Extraction Assessment today.
The Extraction Index is EnshitifAi's subjective editorial opinion, grounded in the dated, publicly documented corporate behaviour described on this page. It is not a measurement of X (Twitter)'s revenue, profit, or conduct, and it is not reported by or endorsed by X (Twitter). EnshitifAi is a satirical parody publication.