Tech

robopop

Streaming views for hire

Robot
Robot

The viewbots have come for Twitch

We’ve seen examples of 5,000% spikes in viewership, many times for no apparent reason.

Ryan Broderick, Adam Bumas

If attention is the main currency of the modern web, that means dealing with counterfeit attention is an increasingly existential problem. 

Across the internet, there are dozens of services and programs designed to create artificial engagement and activity on social platforms, adding hundreds or thousands of followers, likes, and more to accounts. While every platform is fighting a war against the bots, it seems like the streaming service Twitch is getting hit hard right now.

Over the past few months, Twitch has shown multiple signs of a rise in what’s commonly called “viewbotting,” the ostensibly banned practice of having thousands of mass-controlled accounts watch livestreams. These fake views push the streams up in the platform’s recommendation service and, if they’re already popular enough to have a Partner account, generate real money from the site’s revenue-sharing program.

Twitch HQ
A sign outside of Twitch headquarters in San Francisco (Smith Collection/Getty Images)

Viewbotting is against Twitch’s rules, but it’s hard to prove, which is why it’s been an issue on the site for years. Some of the biggest accounts on Twitch have accused their rivals of the practice, while smaller channels often get much less attention for more substantiated claims. Garbage Day has identified a larger pattern to the most suspicious cases — one that suggests the practice has become so common, it’s being used as a weapon. 

We reached out to Twitch for comment weeks ago when we started reporting this story, but haven’t gotten a response. The company did make several public statements in late July about cracking down on the practice.

Viewbotting occupies a complicated spot in the culture and infrastructure of Twitch. The platform’s moderation guidelines specify that fake engagement in general is “a violation of our policies” and “not permitted on Twitch services.” Even so, the ban comes with a lot of caveats, because it’s very difficult to definitively prove viewbotting is taking place without direct admission from the one responsible. The platform’s last widely reported ban for viewbotting happened in March, after streamer QueenGloriaRP accidentally displayed an active viewbotting program during a stream. 

Twitch’s especially opaque moderation system seems to operate on the assumption “innocent until proven guilty,” meaning consequences will be rare without some kind of ironclad evidence. For example, rapper Ray J said during a stream last month, “Viewbotting… so what? It’s legal... Why wouldn’t I?”

Two associates of streaming superstar Kai Cenat were accused of viewbotting in April by Felix “xQc” Lengyel, one of Twitch’s biggest streamers. In response, Cenat said, “I don’t give a f---. They’re all viewbotters — I don’t care,” dismissing Lengyl’s allegations.

Viewbotting is hard to distinguish from authentic popularity. Streams rapidly gain and lose viewers all the time, without any artificial boost needed. For example, last month’s NFL Draft was streamed by Shilo Sanders, brother of top prospect Shedeur Sanders, who was on the stream. When Shedeur was finally selected, over 40,000 viewers started watching the stream within 10 minutes, according to statistics from TwitchTracker, so they could see Shedeur’s reaction live. Even without that clear flashpoint of interest, there are many other elements of Shilo’s stream that indicate the influx of viewers was likely authentic. Most obviously, there were already more than 19,000 viewers before the upswing. Many people who follow college football and the NFL were interested in Shedeur’s fate, so it’s not crazy to think that viewer count could have doubled.

If views become meaningless, so do the ad dollars they represent. 

Over the past several months, however, Garbage Day has found over 20 examples of suspicious spikes over 5,000%, where within 10 minutes, at least 50x as many viewers start watching the stream. Some of the spikes coincided with a major moment of interest like the NFL Draft, but many didn’t have any apparent reason, going from dozens or hundreds of viewers to tens or hundreds of thousands with no ramp-up of increased activity.

A handful of streams in the first few days of May, concentrated on May 2, all saw huge surges. These were smaller channels and the engagement went largely unnoticed by the Twitch community, another sign the engagement might not be authentic. Two of these streamers changed the title of their streams to mention viewbotting. Many of the spikes occurred at the same time on May 2, which was also true of several others on July 3. 

Interestingly, there are also signs that some of these potential cases of viewbotting were carried out by someone other than the streamer. 

One of May’s most watched Twitch streams was by the streamer JonahVeil, who went from less than 200 concurrent viewers to over 150,000 instantly. When the flood of viewers continued, JonahVeil changed the stream title to “Please Read Pinned Message,” and later posted on X that he was “being viewbotted” by “odd fans.” 

Twitch has a history of users gaming the system. Among its community, this is often referred to as a “Twitch meta.” So it’s not surprising, in the wake of xQc’s high-profile accusations going nowhere, that Twitch users may feel emboldened to take advantage of one of the platform’s blind spots.

But this particular meta strikes right to the core of Twitch’s business model. If views become meaningless, so do the ad dollars they represent. 

The difference between authentic and inauthentic content online has never felt more nebulous, thanks to the rise of generative AI. But at the same time, creators are more desperate than ever for views and engagement. It’s an arms race that every platform has to be actively engaged in fighting. And if Twitch can’t keep up, the whole platform could easily fall apart.


Garbage Day is an award-winning newsletter that focuses on web culture and technology, covering a mix of memes, trends, and internet drama. We also run a program called Garbage Intelligence, a monthly report tracking the rise and fall of creators and accounts across every major platform on the web. We’ll be sharing some of our findings here on Sherwood News. You can subscribe to Garbage Day here.

More Tech

See all Tech
tech

New study finds AI doesn’t reduce work — it intensifies it

The rapid adoption of AI by businesses was fueled by the promise of huge productivity boosts that could supercharge workers. A new study has found that while it did indeed boost workers’ productivity, the use of generative AI at work also made work more intense and creep into workers’ downtime.

Researchers Aruna Ranganathan and Xingqi Maggie Ye followed about 200 workers at a US tech company for 8 months. They found AI did speed up work, allowing employees to take on more responsibilities. But after the novelty of their newfound superpowers wore off, workers reported “cognitive fatigue, burnout, and weakened decision-making.”

The researchers noted that to avoid AI-inspired burnout and turnover, organizations should adopt an “AI practice,” spelling out how the technology is expected to be used, and what kinds of limits are in place.

Researchers Aruna Ranganathan and Xingqi Maggie Ye followed about 200 workers at a US tech company for 8 months. They found AI did speed up work, allowing employees to take on more responsibilities. But after the novelty of their newfound superpowers wore off, workers reported “cognitive fatigue, burnout, and weakened decision-making.”

The researchers noted that to avoid AI-inspired burnout and turnover, organizations should adopt an “AI practice,” spelling out how the technology is expected to be used, and what kinds of limits are in place.

tech

Report: Anthropic staffing up to build as much as 10 gigawatts worth of data centers

Anthropic has been hiring a team of executives with a very particular set of skills — building huge data centers. The Information is reporting that Anthropic may be planning to build up to 10 gigawatts of AI computing capacity over several years.

According to the report, Anthropic has hired several former Google executives with deep experience building data centers, which aligns with Anthropic’s heavy use of Google’s tensor processing units.

10 gigawatts would be incredibly expensive. OpenAI executives reportedly have said that building a 1-gigawatt data center costs about $50 billion — putting the cost of 10 gigawatts in the ballpark of $500 billion. But Anthropic told investors it would only spend $180 billion on AI computing servers through 2029, according to the story.

In November, Anthropic announced a deal with Fluidstack to build its first data centers, based in New York and Texas, investing $50 billion in the projects. Anthropic is racing alongside OpenAI to pull off an IPO later this year.

10 gigawatts would be incredibly expensive. OpenAI executives reportedly have said that building a 1-gigawatt data center costs about $50 billion — putting the cost of 10 gigawatts in the ballpark of $500 billion. But Anthropic told investors it would only spend $180 billion on AI computing servers through 2029, according to the story.

In November, Anthropic announced a deal with Fluidstack to build its first data centers, based in New York and Texas, investing $50 billion in the projects. Anthropic is racing alongside OpenAI to pull off an IPO later this year.

tech

Report: OpenAI tells employees it is growing again, with Codex eating into Claude Code’s market share

The competition between OpenAI and Anthropic continues to intensify. Last night during the Super Bowl, a comedic Anthropic ad poked fun at OpenAI’s plans to add advertisements to ChatGPT, something it says it will not do to its Claude chatbot. And both companies released new models last week with improved coding capabilities.

In case OpenAI employees were beginning to sweat from all the pressure, CEO Sam Altman sought to assure the team that the company has gotten its mojo back.

According a new report from CNBC, Altman told employees in an internal Slack group that the company is “back to exceeding 10% monthly growth” and is seeing “insane” growth in its Codex coding tool.

A chart circulated among OpenAI employees shows that this new tool is winning market share from Claude Code, per a screenshot viewed by CNBC.

Per the report, Altman said another new model was coming this week. The company is reportedly working on what could end being a $100 billion investment round.

In case OpenAI employees were beginning to sweat from all the pressure, CEO Sam Altman sought to assure the team that the company has gotten its mojo back.

According a new report from CNBC, Altman told employees in an internal Slack group that the company is “back to exceeding 10% monthly growth” and is seeing “insane” growth in its Codex coding tool.

A chart circulated among OpenAI employees shows that this new tool is winning market share from Claude Code, per a screenshot viewed by CNBC.

Per the report, Altman said another new model was coming this week. The company is reportedly working on what could end being a $100 billion investment round.

tech

Google plans $15 billion US bond sale as capex surges

Alphabet is preparing a roughly $15 billion US investment-grade bond sale, Bloomberg reports, citing people familiar with the deal. The offering is expected to be split into as many as seven tranches, with initial price talk for the longest maturity — a 2066 bond — at about 120 basis points over Treasurys. JPMorgan is leading the sale alongside Goldman Sachs and Bank of America.

In a sign of just how attractive lending money to Alphabet is to investors, the bond sale has already attracted more than $100 billion in orders.

The sale follows Google parent Alphabet’s $17.5 billion US bond deal in November and underscores how even tech companies flush with cash are turning to the bond market to finance their huge AI ambitions. Alphabet expects its capital spending to balloon to $175 billion to $185 billion this year, as it races other tech giants shelling out record sums to get ahead in artificial intelligence. In 2025, the company’s total operating income was $129 billion.

In a sign of just how attractive lending money to Alphabet is to investors, the bond sale has already attracted more than $100 billion in orders.

The sale follows Google parent Alphabet’s $17.5 billion US bond deal in November and underscores how even tech companies flush with cash are turning to the bond market to finance their huge AI ambitions. Alphabet expects its capital spending to balloon to $175 billion to $185 billion this year, as it races other tech giants shelling out record sums to get ahead in artificial intelligence. In 2025, the company’s total operating income was $129 billion.

tech

EU wants Meta to open up access to rival AI chatbots through WhatsApp

The European Union threatened Meta with “interim measures” if it continues to block rival AI assistants from operating on WhatsApp, escalating regulatory pressure on the company’s strategy in Europe. Meta said there was “no reason for the EU to intervene in the WhatsApp Business API,” arguing that its policy does not stifle competition because WhatsApp is not a “key distribution channel” for chatbots.

The dispute comes as Europe seeks to reduce its reliance on US technology companies by fostering homegrown alternatives. Europe accounted for roughly a quarter of Meta’s total advertising revenue last quarter, underscoring the region’s importance to the company’s business.

The case could test how far EU regulators are willing to go to prevent Big Tech firms from favoring their own AI tools inside widely used platforms — a stance that could ultimately reshape how AI services are distributed and monetized across Europe.

The dispute comes as Europe seeks to reduce its reliance on US technology companies by fostering homegrown alternatives. Europe accounted for roughly a quarter of Meta’s total advertising revenue last quarter, underscoring the region’s importance to the company’s business.

The case could test how far EU regulators are willing to go to prevent Big Tech firms from favoring their own AI tools inside widely used platforms — a stance that could ultimately reshape how AI services are distributed and monetized across Europe.

Latest Stories

Sherwood Media, LLC produces fresh and unique perspectives on topical financial news and is a fully owned subsidiary of Robinhood Markets, Inc., and any views expressed here do not necessarily reflect the views of any other Robinhood affiliate, including Robinhood Markets, Inc., Robinhood Financial LLC, Robinhood Securities, LLC, Robinhood Crypto, LLC, or Robinhood Money, LLC.