Matthew Prince thinks the web has already crossed a line it cannot uncross. In a long interview with Nilay Patel on The Verge’s Decoder podcast, published on 26 September 2026, the Cloudflare co-founder and CEO said automated traffic overtook human traffic in May, years ahead of his own forecasts. He expects bots to generate 1,000 times as much traffic as people within five years. His conclusion is blunt: advertising cannot pay for a web whose main visitors never see an ad.
The question in the headline is fair because few people are better placed to answer it. Cloudflare sits in front of more than a fifth of the web, and Matthew Prince has spent two years trying to turn that position into leverage over the AI companies that scrape it. His plan has three parts. Block training crawlers by default, including Google’s if it will not separate search from training. Charge bots a fraction of a penny per page through the long-dormant HTTP 402 status code. Then pool the money the way Spotify pools subscriptions for musicians.
This article tests that plan against the record. We read the full Decoder transcript, Cloudflare’s own posts from July and September, its earnings, and the coverage since. Some of it has already worked. The 15 September switch moved Apple, Google and Microsoft. Some of it has not been built yet, and a few of the figures Matthew Prince quoted on air need correcting. We also ask the question Patel kept pressing: whether one company should hold this much power over who reads the web.
Table of contents
- What Matthew Prince Told Decoder About the Bot-Majority Web
- Why Matthew Prince Says Bots Break the Web’s Business Model
- How Matthew Prince Used 15 September to Pressure Google
- Matthew Prince’s Payment Plan: HTTP 402 and a Spotify for Text
- What Matthew Prince Thinks AI Companies Will Pay For
- Is Matthew Prince’s Leverage Over the Web a Risk?
- How Matthew Prince Is Rebuilding Cloudflare With AI
- Can Matthew Prince Save the Web From AI? Our Verdict
- What UK Site Owners Should Do Now
- References
What Matthew Prince Told Decoder About the Bot-Majority Web
The interview appears to have been recorded in early September, before Cloudflare’s new crawler defaults took effect, and it covers more ground than the headline suggests. Patel opens by reminding his guest that two years earlier he described Cloudflare as a service that “makes the internet faster and protects it from bad guys”. Matthew Prince now calls that his cocktail-party answer for people he does not want to keep talking to.
The longer version, he says, is that “Cloudflare is trying to rebuild the internet the way it should have been built from the beginning if we knew how important it was going to be.” The company runs a network in more than 350 cities. It secures sites and users, hosts applications, and has “recently started to think about how the business model of the internet is changing dramatically”. That last job is the one this article is about.
The forecast that kept arriving early
Matthew Prince told the story of the bot crossover as a series of forecasts that kept being overtaken. At Web Summit in November 2025 Cloudflare’s data pointed to automated traffic passing human traffic in the second half of 2027. At South by Southwest in March 2026 the estimate moved up to the first half of 2027. Then, in May, “the team came to me and said, ‘You won’t believe it, but automated traffic has now passed human traffic online.'”
Matthew Prince posted the news on X on 3 June: “Welp, that happened faster than I predicted.” Cloudflare Radar, the company’s public traffic dashboard, put bots at 57.4% of the HTTP requests it sees in mid-September, against 42.6% from humans. Fortune’s own interview with Matthew Prince on 18 September described bots as “almost 60% of requests”.
| When | What Cloudflare said about bots versus humans | Source |
|---|---|---|
| November 2025 | Automated traffic forecast to pass human traffic in the second half of 2027 | Prince on Decoder (Web Summit answer) |
| March 2026 | Forecast brought forward to the first half of 2027 | Prince on Decoder (SXSW answer) |
| May 2026 | Automated traffic passes human traffic | Prince on Decoder |
| 3 June 2026 | “Bots have now passed human traffic online for the first time” | Prince’s post on X |
| 18 September 2026 | Bots 57.4% of HTTP requests, humans 42.6% | Cloudflare Radar via Startup Fortune |
| About 2031 | Automated traffic at 1,000 times human traffic, human traffic roughly flat | Prince on Decoder, with his own caveat |
Why 1,000 times more traffic breaks the ad model
The five-year number is the one that matters, and Matthew Prince hedged it himself: “with the giant caveat that I’ve gone wrong so far in every prediction that I’ve made on this”. He does not expect human traffic to fall. He expects everything else to explode around it, which means someone has to pay for the bandwidth and servers.
The arithmetic shows why he calls that a business-model problem rather than a capacity one. If human traffic stays flat and automated traffic grows, the human share of requests shrinks fast. The chart below uses only the Radar split above and the multiples Matthew Prince mentioned. At 1,000 to one, people make up about a tenth of one per cent of requests. Cloudflare’s chief financial officer, Thomas Seifert, put it more starkly on the August earnings call: “humans will be a rounding error on the internet”.
Why Matthew Prince Says Bots Break the Web's Business Model
“For at least the last at least 30 years, the business model of the internet has been advertising,” Matthew Prince told Patel. He credits Google with building most of it: search, DoubleClick and Google Analytics, which let publishers see who was visiting. He calls Google “the hero of what I call the first generation of the internet”.
The trouble is simple. “Bots don’t click on ads,” Matthew Prince said. “They don’t respond to pretty swirls of paint and what we traditionally think of as a brand.” If most requests come from software, the money that pays for pages has to come from somewhere else.
The give-to-get deal that search made
The old bargain let search engines crawl pages because they sent readers back. AI answer engines break it. They read the page, answer the question and often send nobody. Our earlier report on how AI crawlers are eating website traffic set out the click-through data behind that shift.
Matthew Prince frames the fix as restoring the exchange rather than ending crawling. Publishers can give content away free to humans, he says, but “if a bot is coming for it, then bots have to pay for it because they don’t have that traditional give-to-get.” Patel, a former copyright lawyer, pushed on whether anyone will publish at all when most readers are machines. That tension runs through the rest of the conversation.
The lunch-menu problem
Matthew Prince’s clearest example is a lunch order. A person choosing a restaurant might look at two or three menus. An AI agent “goes and scans every single menu in the local area”, and only one restaurant gets the money. Scale that up with no cost per request and the agent reads everything on the web before telling you to go to Wendy’s.
“That’s an enormous waste,” he said. Matthew Prince’s answer is a price on every page load, even “a thousandth of a penny or something like that”, to cover the bandwidth and servers behind it. It is an argument about congestion as much as fairness. A tiny price makes an agent choose what it reads, the way the “tragedy of the commons” he mentioned is usually solved.
What AI companies actually buy
Matthew Prince listed three inputs every AI lab needs, and argued that only one of them is getting scarcer. He also said Anthropic “added $10 billion of revenue in a single month”, a figure he gave as evidence of demand; we could not find it confirmed in a filing.
| Input | Prince’s view today | Where he thinks it goes |
|---|---|---|
| Talent and researchers | Scarce | Easing, because universities are rebuilding AI departments and labour markets adjust |
| Chips | Short supply, with Nvidia leading | Easing, as AMD, Qualcomm and the hyperscalers ship their own silicon |
| Data | Historically free, because the web was open to bots | Tightening, as creators start charging bots while staying free for humans |
How Matthew Prince Used 15 September to Pressure Google
The sharpest exchange in the interview is about Google, and it is also where the record has moved since the recording. Patel asked whether Cloudflare needed to “turn the screws all the way and actually block the Google crawlers to get them to pay”.
What he threatened on the podcast
Matthew Prince’s answer was direct. “As of September 15th, we’re going to set the defaults across all of our free customers for it to be that Google is going to be blocked for AI training, but if they won’t differentiate between AI training and search engine training, we’re just going to block them across the board.”
That matched what Cloudflare had published on 1 July. Its Content Independence Day post said that from 15 September, multi-purpose crawlers “such as Googlebot, Applebot, and BingBot will be blocked by customers who have selected to block Training”. Search Engine Watch warned on 9 September that the change “could shut out Google next week, even if you change nothing”.
What Cloudflare actually shipped
On 15 September Cloudflare published “Have it both ways“, and the blanket block did not arrive. Instead the company launched a setting called Disallow AI Training. It publishes a no-training line in robots.txt, keeps approved mixed-use crawlers such as Googlebot in for search, and blocks every other training crawler outright.
Crucially, customers who had used the old “Block AI” switch were migrated to Disallow AI Training, not to a full block. Google stays in their search results. Cloudflare says Apple, Google and Microsoft “honor or have committed (in a specified time frame) to honor this setting”. Fortune reported that the three companies “have agreed to the deal”.
| Training setting (from 15 September) | Googlebot for search | Google training use | Training-only crawlers (Amazon, Anthropic, Meta, OpenAI) |
|---|---|---|---|
| Allow | Allowed | Allowed | Allowed |
| Disallow AI Training | Allowed | Refused through robots.txt (Google-Extended) | Blocked |
| Block on pages with ads | Blocked on ad pages only | Blocked on ad pages only | Blocked on ad pages only |
| Block | Blocked, so the site leaves Google Search | Blocked | Blocked |
The “Accountable” test
To stay on the right side of the new setting, a crawler operator has to earn a label Cloudflare created, “Accountable”. Talks with operators began in July. To qualify, a company must offer an opt-out from AI training and an opt-out from AI summaries. It must also give URL-level visibility into which pages were available for training, and promise that opting out of training will not hurt normal search results.
Apple, Google and Microsoft qualified partly on promises. Microsoft’s robots.txt no-training support is “targeted for early 2027”, and Google says more URL-level reporting for Google-Extended is due “in the weeks to come”. Cloudflare’s own numbers show why the split matters. Fewer than 1% of its sites block search bots, while 17% use some mechanism to block training.
The British lever behind Google’s move
Matthew Prince gave the UK a large share of the credit. He said Google had pushed back on splitting its crawler because it would mean crawling twice, and offered instead to announce what its crawler was doing on each page. It agreed to do that, he said, “largely in response to what was a ruling out of the United Kingdom”, and not only in Britain.
The UK background is the Competition and Markets Authority. It designated Google with strategic market status in general search in October 2025, citing a 90% share, and in January 2026 it consulted on conduct requirements for publishers. In a January response Cloudflare said Google’s opt-outs, “including Google-Extended and ‘nosnippet’, have failed to prevent content from being utilized in ways that publishers cannot control”. It argued that only a separate crawler would work.
That is the tension in September’s result. Disallow AI Training relies on Google honouring Google-Extended, the same opt-out Cloudflare criticised eight months earlier. What has changed is that Google has made commitments in public, and Cloudflare can now block any “Accountable” operator that breaks them.
Matthew Prince's Payment Plan: HTTP 402 and a Spotify for Text
Blocking is the stick. The money is supposed to come from a much older idea, which Patel greeted as “the 1996 dream of micropayments on the web”. Matthew Prince pointed to HTTP status code 402, “Payment Required”, which the web’s standards reserved in the 1990s for future use and then largely left idle.
x402 and the payment rails
Cloudflare has been working with Coinbase and Stripe, Matthew Prince said, so that a site can take “a fraction of a penny every time somebody actually accesses that information”. Coinbase’s protocol for this, x402, now has a governing body. In July the Linux Foundation formally launched the x402 Foundation, with members including American Express, Mastercard, Visa, Stripe, Google and Amazon Web Services, according to Payments Dive.
The rails are the unfinished part. Cloudflare launched its Pay Per Crawl marketplace in July 2025, and in July 2026 said it was evolving into “Pay Per Use”, charging AI companies when content creates value rather than when it is fetched. Neither the interview nor Cloudflare’s recent posts says how much money has moved through it. In his Fortune interview Matthew Prince stuck to the principle. “Companies need to get paid by the AI companies for what is the fuel that runs these AI systems,” he said.
The Spotify model, with the numbers corrected
The model he wants is Spotify. Users pay a subscription for access to everything, and the money is pooled and paid out to rights holders by formula. In Matthew Prince’s version, part of what people pay for AI agents would go to the creators of the knowledge those agents use, alongside the money that goes to researchers and chips.
Matthew Prince said Spotify “sent something like $12 billion back into the music creator ecosystem” last year. Spotify’s own figure, from its Loud and Clear report, is a record of more than $11 billion paid out in 2025. He also said the Anthropic book-piracy case settled “for $2 billion”. A federal judge gave final approval to a $1.5 billion settlement in July. Neither correction weakens his argument, but both show how loosely the numbers fly in this debate.
Premium prices for scarce knowledge
The base rate would be tiny. On top of it, Matthew Prince imagines a premium for material that is expensive to make. “If you have incredibly valuable content, if you’re a news publisher or you’re an academic, then maybe there’s a premium on top of that,” he said. The important claim is that both sides of the market say they want this.
| Layer | How it would work, as Prince describes it | Status on 26 September 2026 |
|---|---|---|
| Identify the bot | Crawlers declare who they are and why they are visiting | Live: Search, Training and Agent classes, the Accountable label |
| Block by default | Stop training crawlers unless the site says yes | Live as the preset for new ad-funded domains and for legacy “Block AI” users |
| Charge per request | HTTP 402 with a fraction of a penny per page | x402 standardising under its own foundation; Pay Per Crawl becoming Pay Per Use; no payout figures published |
| Premium tier | Higher prices for news and research | Handled today by bilateral licensing deals |
| Pooled payouts | A share of AI subscription fees paid out to creators, Spotify style | An idea, with no scheme announced |
What Matthew Prince Thinks AI Companies Will Pay For
The most interesting part of the interview is about which information will be worth money. Matthew Prince thinks it will not be the content that wins today’s attention economy.
The Swiss cheese theory of knowledge
“For the first time in human history, we’ve built a mathematical model of human knowledge,” he said. “That’s what the LLMs are.” He pictures that model as a block of Swiss cheese with a lot of holes. The labs, he says, do not want yet another political story. “What they want is to fill in the holes in the cheese. They want new knowledge that no one ever knew about before.”
He tied this to Spotify publishing searches it cannot answer well, such as a disco song about dancing with a cat, back to songwriters. He said one songwriter in Denmark makes 40 million euros a year writing for those gaps. We could not verify that anecdote, and it should be read as his illustration rather than a fact.
Reddit versus The New York Times
Matthew Prince’s evidence is the licensing market. The New York Times and Reddit hold roughly similar amounts of content measured in tokens, he said, yet Reddit is paid more for its tokens, “by at least seven, or by some measures, 14” times. His explanation was provocative. A model can imitate one national paper from another, but “if you don’t have Reddit, there’s no substitute for it.”
The Park Record test
Matthew Prince has a personal test case. He and his wife, Tatiana, bought The Park Record, the newspaper of Park City, Utah, which has been published since 1880, from Swift Communications. “I think we will make more money off AI licensing deals this year than we do off digital advertising,” he said.
That is his own forecast for a paper he owns, not an audited figure. It fits the thesis, though. A travel agent built on AI needs to know about the new restaurant in Park City, and only the local paper reliably knows. Local news, he argued, “was completely decimated” by the scale-driven ad market and could be worth more in a market that pays for what is unique.
Nilay Patel’s objection
Patel did not let this pass. “Is the outcome you’re describing good,” he asked, “where we’re just making an infinite flood of information GEO-optimized for some human to consume, digested by a chatbot in the middle?” His point was that distribution shapes content, and a market whose buyers are “a bunch of model companies” would shape it too.
Matthew Prince answered with a critique of the current system. He argued that engagement-driven media has made headlines more inflammatory, and that “Google begets Facebook, which begets TikTok”. He told Fortune the same thing more briefly: the last 30 years “really rewarded popularity”, and the next 30 “might really reward credibility”. He also wants a Nobel Prize or Academy Awards for knowledge, so that creators are recognised as well as paid.
Is Matthew Prince's Leverage Over the Web a Risk?
Patel’s hardest question was about power. Cloudflare’s ability to stop AI crawlers comes from the same position that let it drop the Daily Stormer and Kiwi Farms. It has also drawn criticism in the wider scraping debate, including from a Microsoft executive who called AI scraping “the largest theft of labor in human history”.
A hole in the middle of Google’s tree
Matthew Prince explained exactly why Cloudflare has leverage. Google’s ranking, he said, “builds a tree” of reputable sites and their links. “With Cloudflare being 20-plus percent of the internet, if that just disappears, that’s in a giant hole in the middle of the tree. It doesn’t just break it for this, it breaks it for everything.”
That is a remarkable thing for an infrastructure company to say out loud. Matthew Prince describes it as acting “on behalf of and in conjunction with a lot of our customers”. Cloudflare, not a regulator, set the terms Google agreed to on 15 September.
Kiwi Farms, Daily Stormer and the same switch
Asked whether he would make the Daily Stormer and Kiwi Farms decisions again, Matthew Prince said each was “a moment in time”. He admitted that dropping the Daily Stormer was partly a way to force a policy debate. “We very much offended Kant in that,” he said, because the site was used “as a means to an end”.
Matthew Prince’s defence of the current push is that customers can walk away. “It takes five minutes to sign up for Cloudflare… It takes 30 seconds to leave,” he said. He recalled rolling his eyes the first time a media company called the AI firms a threat, then changing his mind after looking at the data. “Is that power, or is that just being a good steward to the internet?”
Both sides of the market are customers
There is a second complication. Fortune reports that Cloudflare counts 80% of the top AI companies as customers. The company sells bot defences and cybersecurity to publishers and infrastructure to the crawlers it is pricing. That can make it a neutral clearing house, or it can make it a toll collector. The difference depends on how open the standards around x402 and bot verification stay.
| Argument | The case for Cloudflare’s role | The case for caution |
|---|---|---|
| Scale | Only a network in front of a fifth of the web can move Google | The same scale lets one firm set terms for everyone |
| Exit | Sites can leave in 30 seconds | Few alternatives offer the same defences at the same price |
| Standards | robots.txt signals, x402 and IETF ai-prefs are open | The “Accountable” label and bot verification are Cloudflare’s own |
| Neutrality | Serves publishers and AI firms alike | Earns money from both sides of the deal it brokers |
How Matthew Prince Is Rebuilding Cloudflare With AI
The second half of the interview turned inward. On 7 May, alongside first-quarter results, Cloudflare said it would cut about 20% of its staff, roughly 1,100 people. It was the first mass layoff in the company’s 16-year history, according to TechCrunch. In a Wall Street Journal op-ed that followed, Matthew Prince explained how he had chosen which roles to replace with AI.
Builders, sellers and measurers
The framework comes from Peter Drucker. Companies have people who build, people who sell and people who measure. Matthew Prince said AI made builders about ten times as productive, so he wants more of them, and “we’re going to hire as many builders and as many sellers as we possibly can”. Most of the cuts came from measurement roles such as internal audit and middle management.
He said the company had tried the gentler route first. From mid-2025 it trained staff on AI tools, using his “electric screwdriver” analogy. On the earnings call in May he said Cloudflare’s AI usage had risen more than 600% in three months.
Two weeks of investor documents in three minutes
Matthew Prince’s example was an investor-relations team of about 20 people who spent two weeks before each earnings release assembling documents. With AI tools, he said, “we took what used to take two weeks and we reduced it down to three minutes”, with fewer errors. He also described managers moving from about six direct reports towards 12, though he admitted “we’re not there yet”.
Matthew Prince was unapologetic about moving first. He described death threats from people opposed to AI, and said peers who agree but wait are, in his word, “chickenshit”. Cloudflare now employs “somewhere between 4,500 and 5,000” people.
Flatter teams and Cloudflare OS
The other change is a shared internal platform, Cloudflare OS, which pulls data from systems of record such as Salesforce and Workday into one interface for staff and their agents. Cloudflare open-sourced it in August; our review of Cloudflare OS covers what it does. Patel’s objection was that this style of management only works where every job leaves a digital trail. Matthew Prince replied that “every company has signal in various ways”.
| Measure | Figure | Source |
|---|---|---|
| Q1 2026 revenue | $639.8m, up 34% year on year | TechCrunch, 8 May 2026 |
| Q1 2026 net loss | $62.0m | TechCrunch, 8 May 2026 |
| Q2 2026 revenue | $696.1m, up 36% year on year | Q2 results, 6 August 2026 |
| Q2 2026 operating result | GAAP loss $205.7m; non-GAAP income $96.1m | Q2 results, 6 August 2026 |
| Layoff | About 1,100 roles, roughly 20% of staff | Announced 7 May 2026 |
| Headcount now | 4,500 to 5,000 | Prince on Decoder |
| Network | More than 350 cities | Prince on Decoder |
| AI companies as customers | 80% of the top AI companies | Fortune, 18 September 2026 |
Can Matthew Prince Save the Web From AI? Our Verdict
The honest answer is that he can save part of it, and he has already changed the terms for the biggest crawler on the web. Whether that becomes a working economy for small publishers depends on pieces that are not finished. First, the claims from the interview, checked against the record.
| What Prince said on Decoder | What the record shows |
|---|---|
| Automated traffic passed human traffic in May | Consistent: his 3 June post and Radar’s 57.4% bot share |
| Google blocked across the board from 15 September if it would not separate training | Did not happen: Google was rated Accountable and stays in search under Disallow AI Training |
| Cloudflare sits in front of more than 20% of the web | Cloudflare’s own long-standing figure, repeated by Fortune |
| Anthropic settled the books case for $2 billion | The approved settlement is $1.5 billion |
| Spotify paid about $12 billion to music last year | Spotify reports a record of more than $11 billion for 2025 |
| Layoffs made “back in April” | Announced on 7 May with first-quarter results |
| The Park Record will earn more from AI licensing than digital ads this year | His own forecast; not independently verifiable |
Where he is right
Matthew Prince’s diagnosis holds up. Ads do not work when most visitors are software, and the free crawl-for-referrals bargain has broken. The 15 September change also proves that coordinated defaults can move the largest platforms. Google, Apple and Microsoft now publicly honour a training opt-out that is separate from search, which publishers could not rely on a year ago.
Matthew Prince is also right that agents need a price signal. An agent that pays nothing for each page has no reason to read less, and site owners pay the bandwidth bill. Even critics of Cloudflare’s power tend to agree on that part.
Where the plan is still thin
The payment layer is mostly a promise. HTTP 402 has been “the future” since the 1990s, and the x402 Foundation is only months old. The pooled Spotify-style payout has no operator, formula or timetable. Until money actually flows to ordinary sites, the plan protects big publishers who can sign licensing deals better than it protects the long tail.
Two other risks stand out. The September settlement leans on Google-Extended, an opt-out Cloudflare itself called ineffective in January. And the more successful Cloudflare is as a gatekeeper, the more Patel’s power question matters. As Matthew Prince put it to Fortune, “The world that I’m playing for is not one where there’s five AI companies. It’s where there’s 500,000.” That world needs open rules, not just one company’s settings page.
| Problem | Prince’s lever | Verdict today |
|---|---|---|
| Training without consent | Default blocks and Disallow AI Training | Largely working |
| Google’s search-plus-AI crawler | Threatened block, Accountable label | Partly working, relies on Google’s commitments |
| AI summaries that replace visits | Summary opt-outs, finer controls promised | Opt-outs exist; finer control targeted for early 2027 |
| Paying creators | HTTP 402, x402, pooled payouts | Not yet delivering at scale |
| Recognising creators | A prize for knowledge | An idea only |
What UK Site Owners Should Do Now
You do not need to agree with the whole vision to act on the parts that are live. If your site runs on Cloudflare, the 15 September change has already altered your settings or your defaults.
Check your Training setting
Open the security settings for each domain and look at the Search, Training and Agent controls. If you previously switched on “Block AI”, you have probably been moved to Disallow AI Training, which keeps Google search and refuses training. Choose Block only if you are willing to leave Google Search, Bing and Apple’s search features entirely.
Decide what AI summaries are worth to you
Training and summaries are separate decisions. Cloudflare cites research that summary readers are more likely to end a search without clicking, while visitors referred by AI search can convert at several times the usual rate. Google offers a toggle in its webmaster tools to keep a site out of generative results. Weigh that against how much of your revenue depends on visits. Our GEO services team helps businesses measure how they appear in AI answers before they decide.
Keep your own audience
The strongest point in the whole conversation came from Patel’s summary of what Google’s chief would like to say to publishers: “You got super addicted to my fire hose of traffic.” Whatever happens to HTTP 402, email lists, communities and direct relationships are revenue that no crawler setting can take away. The same agents that read your pages are also misbehaving on the open internet, which is one more reason to know exactly who is visiting.
References
Can Cloudflare CEO Matthew Prince save the web from AI? (The Verge, Decoder)
Can Cloudflare CEO Matthew Prince save the web from AI? (The Verge via MSN)
Have it both ways: stay discoverable in search while disallowing AI training (Cloudflare)
Your site, your rules: new AI traffic options for all customers (Cloudflare)
Cloudflare on the CMA’s proposed Google conduct requirements (Cloudflare)
Matthew Prince: bots have passed human traffic online (X)
Cloudflare Radar: bot versus human traffic
Cloudflare’s Matthew Prince Says Bots Now Outnumber Humans Online (Startup Fortune)
Cloudflare’s new policy pushes AI companies to pay for publishers’ content (TechCrunch)
Cloudflare’s AI block could shut out Google next week (Search Engine Watch)
Cloudflare says AI made 1,100 jobs obsolete, even as revenue hit a record high (TechCrunch)
Cloudflare Announces Second Quarter 2026 Financial Results (Cloudflare)
Visa, Stripe join AI protocol project (Payments Dive via Yahoo Finance)
Judge approves Anthropic’s $1.5 billion settlement with authors (Mashable)
Spotify paid a record $11 billion to the music industry (IBTimes UK)
Tatiana and Matthew Prince set to acquire the 143-year-old Park Record (Newspapers.org)
Cloudflare says humans could become a “rounding error” (TechSpot)
More AI coverage: explore Progressive Robot's AI Models, Tools & Releases hub — hands-on reviews, setup guides and benchmarks in one place.
