Elon Musk's xAI empire is leaning into NSFW content. It's a risky move that could pay off.
XAI's products have taken a decidedly NSFW turn this summer, which could prove lucrative for the company — and challenging.
In June, xAI released a female AI anime companion that can be instructed to strip down to lingerie and describe sexual situations. Grok Imagine, the company's AI video generator, launched a "spicy" mode capable of borderline pornographic imagery — with some users finding it quick to generate Taylor Swift deepfakes that showed the singer's AI likeness topless.
It's a notable embrace of the idea that sex sells and an example of Musk choosing to zig where other AI companies are zagging. And with Grok closely integrated into the X platform, the NSFW content it produces is the latest evolution of the social media company's relationship with steamy content, as Twitter allowed adult content long before Musk purchased it.
Grok has limits. It doesn't appear that users can create fully nude videos or AI companions — the content must cover genitalia. But the NSFW shift puts xAI in sharp contrast to OpenAI and Google, which bar their AI models from generating much erotic content.
And while xAI is tapping into the multibillion-dollar adult entertainment industry, it could open the door to legal fights and advertiser skittishness.
Steamy content could prove a moneymaker
The adult entertainment market is massive. In the mid-2010s, estimates ranged from $10 billion to $97 billion, and the industry has likely grown significantly since then.
"There's essentially free entry to the market. There's not a lot of barriers; anybody with an iPhone can create porn," said Marina Adshade, a sex economist and assistant professor at the University of British Columbia. "The real money is being made in the distribution."
Adshade referenced the popular adult website Pornhub's market dominance. The company is privately held and does not report annual revenue; its parent company MindGeek reported $460 million in revenue in 2018 before its acquisition by Ethical Capital Partners.
OnlyFans, another major player in the adult entertainment market, reported $6.6 billion in gross payments to creators in 2023, of which the company takes a 20% cut. Its owner, Fenix International Limited, reported net revenue of $1.31 billion for the year. (Twitter, before Musk acquired it, reportedly explored monetizing adult content before nixing the idea after its Red Team concluded in April 2022 that the platform "cannot accurately detect child sexual exploitation and non-consensual nudity at scale.")
Customizability could prove to be xAI's edge. Adshade said that AI-generated porn can cheaply serve any user's specific preferences, giving it a market advantage.
"That's where the money is," she said.
There's plenty of existing demand for AI-generated sexual images, much of which has funneled to gray-market tools. Analysis from Indicator estimates that AI "nudifier" tools make around $36 million a year.
Henry Ajder watched these tools grow. He advises companies like Meta, Adobe, and EY on deepfakes and their impact. He described a network of tools that masquerade on the App Store as "fun, meme-ifying apps," but advertise themselves as deepfake tools on other platforms.
Now, xAI is one of the first big players to tap into that demand, Ajder said, though the company's offerings remain "softcore."
"It is something that a lot of the other models don't allow you to generate," Ajder said. "I guess it's trying to cater to an audience, and seeing a gap in the market."
OpenAI CEO Sam Altman recently acknowledged that NSFW AI may cause users to spend more time on platforms that offer it. On Cleo Abram's YouTube show, he said that adding a sex-bot could "juice growth or revenue," but that he has abstained because it doesn't align with OpenAI's mission.
App download data from market intelligence firm Sensor Tower suggests some of the new Grok features may have boosted downloads. The week xAI added companions — a week after the new Grok-4 model debuted — global downloads increased an estimated 41%. The week it added "spicy" video generation, global downloads fell an estimated 9% — but US downloads increased by 33%, per Sensor Tower.
Those looking to chat with the AI companion Ani or generate an NSFW video with Grok Imagine will have to pay. For premium access, xAI charges $30/month.
XAI did not respond to a request for comment.
NSFW chatbots pose brand risks
For all the money that xAI's adult turn could bring in, it comes with risks and moderation challenges. Like other AI companies, xAI will need to ensure its AI models are properly trained to prevent outputting illegal content.
Grok's "spicy" video generation is only available for AI-generated images. When a user uploads their own image to the video generator, the only options are "fun," "normal," and "custom."
I tried to test the limits of the "custom" button to see how it compared to "spicy" mode. While it was willing to comply with my prompt to "take off shirt" on a photo of myself I had uploaded, I was unable to convince the AI to remove my pants.
I also tried multiple times to have the bot create NSFW videos of celebrities. While my attempts to make "spicy" videos of Taylor Swift were shut down by Grok, the tool easily allowed me to create NSFW videos of other celebrities.
That could lead to PR headaches. Celebrities could publicly criticize xAI over deepfakes with their likeness, especially any containing nudity, and they command large fan bases.
Legal recourse around deepfakes isn't cut-and-dried and continues to evolve. The recently enacted Take It Down Act protects victims of deepfakes, but images of non-minors must include specific sexual body parts to be subject to the legislation.
Some states have more stringent laws on AI-generated deepfakes, and those affected could also file civil claims.
Section 230 of the Communications Decency Act of 1996 largely protects online platforms from liability for content shared by users.
Allison Mahoney, a lawyer who represents people who have experienced technology-facilitated abuse, questioned whether considering the platforms as creators due to their AI-generating tools would "remove their immunity."
"There needs to be clear legal avenues to be able to hold platforms accountable for misconduct," Mahoney said.
Where xAI could more directly feel the burn is in advertising revenue. It acquired X in March, and many of Grok's capabilities are integrated into the social platform. The Grok app makes it easy to share content, including its "spicy" output, directly to X.
X has allowed pornographic content on its feed for years. In 2024, X put that rule into writing, officially allowing consensual NSFW content.
Kelsey Chickering, a principal analyst at Forrester, said that AI-generated sexual content could push buyers of X's ad space away.
"When you layer on this whole not-safe-for-work feature and add suggestive new things to the mix, that just further degrades their credibility," Chickering said. "It gives advertisers just one more reason to walk away."
Try Our AI Features
Explore what Daily8 AI can do for you:
Comments
No comments yet...
Related Articles
NBC News
2 hours ago
- NBC News
Apple Watch getting redesigned blood oxygen feature following legal dispute
Apple on Thursday announced a redesigned blood oxygen feature for some Apple Watch users, following a yearslong intellectual property dispute over the capability. Apple said the redesigned feature is coming to some Apple Watch Series 9, Series 10 and Apple Watch Ultra 2 users on Thursday. The update was possible because of a recent U.S. Customs ruling, the company said. In 2023, the International Trade Commission found that Apple's blood oxygen sensors infringed on intellectual property from Masimo, a medical technology company. Apple paused the sale of some of its watches and began selling modified versions of the wearables without the blood oxygen feature. 'Apple's teams work tirelessly to create products and services that empower users with industry-leading health, wellness, and safety features that are grounded in science and have privacy at the core,' the company said in a release announcing the feature rollout. CNBC has reached out to Masimo for comment. Users who do not currently have the feature will be able to access it by updating their iPhone to iOS 18.6.1 and their Apple Watch to watchOS 11.6.1, Apple said. Users will be able to see their results in the Respiratory section of the Health app. Apple has been pushing deeper into health care in recent years. The company recently released a sleep apnea detection feature for Apple Watch users and hearing health features for its AirPods headphones. In February, Apple launched its first major health study in five years.
Forbes
3 hours ago
- Forbes
Cutting AI Costs: Smart Strategies for Small Business Savings
If you run a small business, you might already feel the AI pinch: your customer support runs on ChatGPT, your marketing automation uses Claude, and you're paying for Grok's research capabilities and real-time updates. For the average company (or user), those subscriptions can easily hit $300 a month, especially if you're integrating multiple tools into your workflow. That's a serious line item for what's supposed to be affordable technology. What most don't realize is that these ballooning costs have more to do with the hardware on the backend than they do the software running their workflows. Every time an AI model responds, it triggers a process called inference: the act of generating output from a trained model. Unlike training—which costs a fortune but only happens once—inference occurs billions of times each day and scales with usage. It has become one of the largest ongoing expenses in AI, driving massive, sustained energy demand that fuels the industry's growing power crisis. For individuals and small business owners, this hidden cost means AI remains incredibly expensive. But that might be about to change. A new cohort of hardware startups—including Positron AI, Groq, Cerebras Systems, and Sambanova Systems—are racing to make inference radically cheaper. If they succeed, AI tools could drop from $300-a-month luxuries to accessible everyday infrastructure for freelancers, educators, retailers, and entrepreneurs. If Positron and its peers succeed, the $300-a-month AI stack could shrink to $30. It could also be replaced entirely by tools you run yourself, privately and affordably. And that changes who gets to participate in the future of AI. Among these, Positron has emerged as a favorite choice by some of the world's dominant neocloud providers, gaining investor attention for its unique approach. 'The early benefits of AI are coming at a very high cost – it is expensive and energy-intensive to train AI models and to deliver curated results, or inference, to end users.' DFJ Growth co-founder Randy Glein said. 'Improving the cost and energy efficiency of AI inference is where the greatest market opportunity lies, and this is where Positron is focused.' Inference Is The New Electricity Bill In the world of AI economics, inference is like your utility bill: it grows as you grow, and it's never just a one-time fee. Whether you're sending AI-generated emails or running a support chatbot, inference is what keeps the lights on—and right now, that light is powered by Nvidia's premium-priced GPUs. 'Nvidia GPUs have become the backbone of AI infrastructure today according, powering nearly every major inference workload at every major cloud provider. The downside to this, beyond having one $4 trillion company They own the entire inference market, is that they weren't designed with efficiency in mind. They're built for flexibility and optimized for training complex models that require general-purpose chips for multifaceted tasks. And yet, the majority of inference today still runs on Nvidia hardware, leaving the industry with high power usage, steep cloud bills, and limited options for smaller players.'said Mitesh Agrawl CEO of Positron The Race To Make AI Affordable Those are exactly the problems Positron, Groq, Cerebras, and Sambanova are solving by building alternatives to the Nvidia tax. And while they all share a common goal—deliver inference infrastructure that slashes energy consumption, improves performance-per-dollar, and gives developers more control—Positron is arguably the most technically ambitious and commercially mature contender in this race. Founded by systems engineer Thomas Sohmers and compiler expert Edward Kmett, Positron has taken a radically different path from its peers. Instead of building application-specific chips or chasing general-purpose GPUs, Positron bet on field-programmable gate array (FPGAs)—reconfigurable chips optimized for memory efficiency—and used them to build Atlas, an inference-first system designed from the ground up for performance and energy savings. Atlas delivers 93 percent memory bandwidth utilization (vs. about 30 percent for GPUs), uses 66 percent less energy, and offers 3.5 times better performance per dollar—all while supporting seamless deployment with no code changes. That kind of out-of-the-box compatibility makes it a practical swap for existing cloud or local systems without forcing teams to rewrite their infrastructure from scratch. These gains have landed it major enterprise deployments with Cloudflare, Crusoe, and Parasail. The company recently raised a $51.6 million Series A led by Valor Equity Partners, Atreides, and DFJ Growth—the very firms that bankrolled SpaceX, Tesla, X, and xAI, some of the world's largest buyers of AI hardware. Positron is already working on its next-generation system, Titan, built on custom 'Asimov' silicon, which is expected to support models up to 16 trillion parameters with two terabytes of memory per chip—all while running on standard air-cooled racks. That could make high-throughput inference viable in a wider range of environments, from enterprise data centers to sovereign cloud infrastructure. While others in the field are exploring niche optimizations, Positron is staking a claim to general-purpose inference acceleration—solving for cost and compatibility at scale. But it's not alone. Other Challengers Redefining The Stack While Positron is focused on general-purpose inference acceleration, other challengers are tackling different bottlenecks. Groq is optimizing ultra-low-latency inference for large language models (LLM). Its Tensor Streaming Processor (TSP) delivers consistent, repeatable latency, with sub-millisecond response times—enabling a new class of AI tools that respond instantly—without incurring massive cloud costs—and laying the groundwork for local, responsive AI that could eventually be accessible to small businesses. Cerebras brings an edge-native, security-first perspective. Its modular AI appliances can run powerful models entirely on-site—ideal for defense, critical infrastructure, or industries where cloud deployment isn't an option. Cerebras makes it possible for organizations to deploy advanced AI with a small footprint—something previously only achievable by hyperscalers. Sambanova is taking a full-stack approach, combining hardware and software to deliver vertically optimized AI systems. Rather than asking businesses to build training pipelines and inference clusters from scratch, they offer a turnkey platform with pre-trained models—essentially packaging AI as an appliance for organizations without a dedicated machine learning (ML) team. All of these players are on a mission to unlock high-performance inference that doesn't require hyperscaler infrastructure or ballooning cloud costs—opening the door to entirely new economic possibilities. Why This Matters For Your Mottom Line If inference becomes cheaper, everything changes. A Shopify seller could train and run a private AI model locally—without relying on costly cloud infrastructure. A solopreneur could fine-tune a sales assistant on years of customer emails, then run it on a $10 chip instead of a $30,000 graphic processing unit (GPU). A tutoring platform could deploy personalized lesson-plan generators without needing a full-time infrastructure team. This is already happening. Smaller teams are building domain-specific copilots that live inside their own companies' firewalls. Independent consultants are running multi-agent AI workflows from their laptops. This shows that inference costs are a technical problem, but more importantly, they're the gatekeeper to who gets to build with AI. If Positron and its peers succeed, the $300-a-month AI stack could shrink to $30. It could also be replaced entirely by tools you run yourself, privately and affordably. And that changes who gets to participate in the future of AI. Nvidia's Grip Might Finally Be Loosening Today, Nvidia holds a near-monopoly over AI infrastructure—and by extension, who gets to play. Its chips power the vast majority of generative AI systems worldwide, and its ecosystem (CUDA, TensorRT, etc.) makes switching difficult. The result is a pay-to-play system where cost determines access. But that grip may not hold if companies like Positron, Groq, Cerebras, and Sambanova continue to gain traction and change the economics of AI. By lowering the cost of inference, they're making it possible for smaller teams and individual users to run powerful models without relying on expensive cloud infrastructure. This shift could have broad implications. Instead of paying hundreds of dollars a month for AI-powered tools, users may soon be able to run custom assistants, automations, and workflows locally—on hardware they control. For small businesses, freelancers, educators, and startups, that means more control, more customization, and a lower barrier to entry—for a fraction of today's cost. If inference becomes affordable, innovation stops being a privilege and starts becoming infrastructure. Because when you democratize cost, you decentralize control. The next chapter of AI won't be written by whoever builds the biggest model, but by whoever makes it cheap enough to run—and that's how you break a $4 trillion monopoly.
Yahoo
3 hours ago
- Yahoo
Apple accidentally leaked its own top-secret hardware in software code—and it looks like we're getting a slew of products across 7 categories
Apple appears to have accidentally revealed a slate of unannounced hardware after internal identifiers surfaced in publicly shared software code, offering unusually concrete clues across multiple product lines. The disclosures, first surfaced by MacRumors, point to next-generation chips spanning Apple's iPhone, iPad, Apple Watch, Mac, and Vision Pro product lines, and help validate past reports and rumors about Apple's roadmap for future products. For some context: Apple's software regularly includes device strings, codenames, and chip references that surface ahead of hardware—often confirming families and classes, if not final specs. The identifiers coherently map Apple's late‑2025 to 2026 pipeline: silicon upgrades across the lineup, an Apple TV spec jump that unlocks AI/gaming features, a Vision Pro 'speed bump' anchored on M5, and display R&D that points to a higher‑end Studio Display in early 2026. To be clear: While all these details below come from identifiers and documentation located in Apple's own software code, Apple has not actually announced any of these products, so specifications, features, and timelines can change before release. As always, take reports about future Apple products with a grain of salt until they are officially unveiled. What leaked, and what it suggests New HomePod mini What's coming: A next‑gen HomePod mini (codename B525) tied to Apple's T8310 microarchitecture—the same core architecture used to power the latest Apple Watches—indicating a major jump over today's S5-based model. MacRumors notes Apple is also testing an in‑house Wi‑Fi/Bluetooth combo with Wi‑Fi 6E and is considering new colors. How it differs: The current mini runs an S5-class chip without a Neural Engine. The new model is expected to add a 64‑bit dual‑core CPU derived from A16 and a 4‑core Neural Engine for faster on‑device processing and smarter audio features. Pricing context: The mini launched at $99 and has held close to that level. A chip and radio upgrade could carry a small premium, but Apple often protects the entry price. New Apple TV What's coming: An Apple TV refresh 'later this year' would prominently feature the A17 Pro, the same chip that powered last year's iPhone 15 Pro, which would significantly improve Apple TV performance, Apple Intelligence support, and headroom for console‑style graphics. How it differs: Today's Apple TV 4K runs A15. The A17 Pro upgrade should deliver sizable gains and enable more advanced gaming and AI features; Apple is also testing its own Bluetooth/Wi‑Fi silicon for the box. Pricing context: The current model starts at $129–$149; an A17 Pro upgrade suggests Apple stays near that band, with possible variance for storage or Ethernet. Apple Studio Display 2 What's coming: Code references to 'J427' for a next‑generation Studio Display, with a second code path 'J527' under evaluation. Reporting points to a potential mini‑LED backlight, early 2026 timing, and pairing alongside the first Mac computers powered by Apple's forthcoming M5 chip. How it differs: The 27‑inch Studio Display (2022) uses a standard LED backlight and 60Hz. A mini‑LED upgrade would raise peak brightness, contrast, and HDR performance. Two code paths suggest Apple tested multiple sizes or variants before finalizing. Pricing context: The current Studio Display starts at $1,599. A mini‑LED version could command a higher price, but Apple's exact positioning (and whether features like ProMotion appear) remains unclear until closer to launch. New iPad mini What's coming: The next iPad mini (J510/J511) is tied to the A19 Pro—the same chip family earmarked for iPhone 17 Pro—marking a rapid step up from the current mini's A17 Pro. How it differs: Beyond the SoC leap, Apple has explored an OLED mini as early as 2026, but timing remains up in the air. The A19 Pro variant is expected to offer more GPU cores than A19; Apple may also sell a version with one fewer core for certain devices, as it often does with its mobile devices. Pricing context: Apple has priced the iPad mini as a premium small tablet; an A19 Pro uplift without an immediate display overhaul suggests pricing stays where it is. New entry-level iPad What's coming: Apple's next low‑cost iPad (J581/J582) is referenced with an A18 chip and is tracking for a spring refresh. How it differs: The current entry-level iPad uses an A16 chip. The A18 will reportedly add a 16‑core Neural Engine and Apple Intelligence support, bringing modern AI features to the budget tier while maintaining the existing chassis, according to prior reporting cited by MacRumors. Pricing context: Apple has targeted aggressive entry pricing; the current entry model starts at $349. If the only major change is moving to an A18 chip, that starting price will likely hold. Vision Pro 2 What's coming: Code strings align Vision Pro's second generation to the M5, settling months of M4 vs. M5 reporting. Aside from silicon, changes are expected to be modest, with a new strap for comfort under discussion. The launch could be late 2025. How it differs: The first‑gen headset uses M2. An M5 step would dramatically lift CPU/GPU performance and neural processing while keeping the industrial design largely intact, per current indications. Pricing context: The first model starts at $3,499. With an iterative update centered on silicon and comfort, Apple is likely to keep the entry price in the same zone pending any configuration changes. New Apple Watch What's coming: Series 11, Ultra 3, and Watch SE 3 are tied to an S11 SiP that maintains the T8310 architecture used in S9/S10 (two 'Sawtooth' performance cores, 64‑bit dual‑core CPU, 4‑core Neural Engine, second‑gen UWB, 64GB storage). How it differs: Apple often reuses watch microarchitectures for multiple years. Expect packaging or efficiency tweaks rather than big performance gains this cycle; larger shifts may come with S12 in 2026. Codename references also point to next‑year models already in planning. Pricing context: Apple has historically kept the aluminum Series model's U.S. starting price steady—Series 9 launched at $399, with another $100 for the cellular option. Absent a materials or display change, Series 11 is likely to remain in that band. Similarly, the Ultra line has held a $799 starting price since its debut, so barring new sensors or materials, that pricing is likely to stay in that ballpark. And the value pricing for the Watch SE—$249 to start, with $50-100 extra for cellular depending on configuration‚ is likely to have a similar entry price. A leak not without precedent This isn't the first incident of this kind; Apple has a long history of leaving breadcrumbs about future hardware through breadcrumbs in its software. AirTags, for instance, were once named in an official Apple support video before launching after months of iOS code hints that foreshadowed the product and its 'Find My' integration. Other popular Apple products like AirPods, Apple TVs, and Watch variants have also been tipped off through references in code. So, while none of the products are official yet, the newly surfaced identifiers are consistent with Apple's internal taxonomy and build on earlier code-based reporting touching Apple's entire hardware lineup. But, if these references are accurate, this could be a very expensive holiday season coming up for Apple fans. For this story, Fortune used generative AI to help with an initial draft. An editor verified the accuracy of the information before publishing. This story was originally featured on
