請同意我們的隱私權規範，才能啟用聽新聞的功能。

（中央社雪梨30日綜合外電報導）澳洲網路安全監管機構今天對社群平台Telegram採取法律行動，指控未刪除恐怖主義內容，最重可依澳洲網路安全法開罰澳幣5460萬元（約新台幣12.3億元）。

綜合法新社與路透社報導，澳洲電子安全委員辦公室（eSafety Commissioner）安全專員葛蘭特（Julie Inman Grant）今天表示，Telegram接獲通知後，仍未刪除處決影片和集體槍擊案等恐怖主義相關內容。

葛蘭特說，本案涉及「近期歷史上已知極端主義暴力中，最惡名昭彰行為相關內容」，包括紐西蘭清真寺槍擊案以及美國紐約州水牛城槍擊案相關影片，仍可在平台上找到。

Telegram發言人在回應置評請求時否認相關指控，表示「將在法庭上提出抗辯」，同時強調Telegram反恐作為有目共睹，光是今年就封鎖數千個極端主義社群。

澳洲電子安全委員辦公室指出，Telegram是一個大型全球平台，澳洲人每月造訪量達150萬次。

Telegram俄羅斯裔創辦人度洛夫（Pavel Durov）2014年離開俄羅斯，2017年將個人與公司遷至杜拜，目前Telegram已成為俄烏戰爭重要的資訊傳播管道，獲莫斯科及基輔官員廣泛使用。

俄羅斯昨天才宣布要逮捕度洛夫，指控他「涉嫌參與恐怖主義」。（編譯：張茗喧）1150730

Reddit reported a strong second quarter but spooked investors by mentioning in an investor letter that traffic from search engines had grown “choppy.”

The company’s total revenue of $805 million jumped 61% compared to the year-ago quarter while net income was $253 million, up 183%. These beat Wall Street’s expectations. Plus, the company said it expects to hit revenue of $860 million to $870 million with healthy earnings before taxes next quarter as well, which beat expected guidance.

Normally, telling investors that they can expect even better results to come would cause a stock to rise. But the stock slid over 10% in after-hours trading.

The culprit was likely CEO Steve Huffman’s warning in that separate letter to shareholders. “Search referrals were choppy in the quarter, and traffic was more volatile later in the quarter, but the bigger picture is unchanged: the commercial business is strong,” he wrote.

AI is clearly changing the search engine landscape — and investors seem to fear that Reddit’s traffic may suffer.

In 2024, Reddit signed a contract to provide its content to Google for the purposes of AI training. However, Google’s deployment of AI summaries appears to be peeling away audience share from Reddit, and the site has signaled it’s not sure whether it will renew the partnership with the search giant.

Another sticking point has been whether Reddit’s audience is growing in the right places or not. Reddit’s global users are up, but the platform saw U.S. users, in terms of daily active uniques, endure a very slight decline — 53.2 million from 53.5 million in Q1.

During Thursday’s earnings call, some analysts aggressively questioned AI’s impact on Reddit’s audience.

“I don’t want to belabor this point, but your stock is down sharply because there is just a sense from investors that you have a — I don’t want to, to be blunt — a user problem, especially in the U.S.,” one Wall Street analyst said.

“People are looking at the daily and saying, you know, the logged-out traffic is going to be under pressure because as search shifts to AI, you’re not going to get referrals, and that it’s going to be harder to get people to go from logged-out to logged-in, and that’s sort of symptomatic of what you’re seeing,” the analyst continued, asking “Do you see any world where you’re not licensing data to Google and OpenAI next year?”

Huffman said that he felt that the human aspect of Reddit would continue to make it a destination for users. “Reddit is communities and conversation. Communities are universal, and so we think we have in the U.S. content for everyone, and it’s a matter of revealing that. And we’re making progress towards that end,” Huffman said.

When it comes to Google, Huffman was vaguer. “Our relationship with them actually predates the formal kind of data licensing agreements,” he said. “I don’t think there’s a binary outcome…we will make sure that we’re maximizing the value for Reddit.”

Amazon reported better-than-expected second-quarter earnings on Thursday, and investors loved what they saw. Net sales rose 20%, and cloud revenue stood out as a particular bright spot. This combination of positive results was enough to send Amazon’s stock up nearly 10% in after-hours trading.

Crucially, Amazon isn’t slowing down on data center spending, despite the conventional wisdom that investors want companies to rein it in.

One line item, in particular, illustrates Amazon’s appetite for investing in infrastructure. Amazon spent $173 billion for the fiscal year ended June 30 on property and equipment — a category that covers GPUs, natural gas turbines, and plots of land — up from $107.65 billion from the year before

It also raised its 2026 capex forecast from $200 billion to $220 billion — even as it has begun dipping into its cash reserves to help cover the cost. The company ended the quarter with $7.6 billion less cash than it had 12 months ago, marking its first period of negative free cash flow this year.

Under normal circumstances, ballooning expenses would be a tough pill for investors to swallow. But Amazon has a revenue engine that helps justify the spending. AWS revenue rose 37% year over year, clocking $42 billion for the quarter. That’s not enough to balance out the capex spending in raw arithmetic, but it shows that demand is growing alongside supply. Given the years-long time lag between breaking ground on a data center and selling its capacity, that’s reassuring for investors.

Critically, Amazon’s AI play isn’t limited to building large data centers. The company is also making serious long-term bets on chips like the Trainium TPU and the Arm-based Graviton processor. Those projects don’t show up in capex numbers, but they can meaningfully improve margins for the company’s cloud business.

“We see the AI business following very much the same margin trajectory we saw in the core business before,” Jassy said during the company’s Q2 earnings call. “AWS and Amazon Bedrock can have a wildly successful business without its own frontier model, and the reason is that there’s not going to be a single model to rule them all.”

This dynamic isn’t unique to Amazon. We saw similar patterns at Microsoft and Google, whose shares also popped after reporting strong cloud revenue. By the same token, companies like Meta which have significant capex and no clear revenue source, are still experiencing intense skepticism from investors. Meta’s stock fell 8% after earnings this week, as investors focused on its cash flow crunch and continued spending,

Of course, investors like revenue and don’t like expenses — that’s how markets work. But it’s important not to miss the broader lesson about the AI economy. Right now, investors are treating cloud hosts as the most reliable part of the AI stack, while remaining skeptical about the underlying economics for AI labs and AI startups.

But Amazon’s hosting revenue is someone else’s AI bill. In Anthropic’s case, it’s literally the same money.

If that spending isn’t sustainable for the big labs and their clients, the revenue won’t be stable for Amazon and the other cloud hosts. There’s real competition and differentiation at every level of the stack, but if demand for AI doesn’t hold up, it’s going to be a bad time for everyone.

In the end, it all comes back to David Cahn’s $3 trillion question. There’s either enough demand to justify this buildout or there isn’t. Cloud-hosting services like AWS may be a few steps removed from that demand problem, but that doesn’t mean they’re insulated from it.

Apple says it has now topped 1.5 billion subscribers for its services business, up from 1 billion in January 2025. However, this segment of Apple’s business, which includes the Apple Store, AppleCare, music, video, and cloud services, was the only miss in what was otherwise a record-breaking quarter for the company’s hardware sales.

In Apple’s fiscal third quarter, the company reported $30.74 billion in services revenue, falling short of the $31.22 billion Wall Street analysts had expected. Combined with a miss in China, Apple’s stock tumbled over 4% in after-hours trading.

When asked to dive into what led to the decline in services revenue, Apple CFO Kevan Parekh pointed to several factors. Most notable, however, were the impacts on Apple’s cash cow, the App Store.

One factor contributing to the App Store’s performance in the quarter was a slowdown in mobile gaming. It also called out the App Store business model changes in certain countries, including the U.S.

The latter is a reference to Apple being under a court order that requires it to now allow app developers to process customer payments outside the App Store — and outside the reach of Apple’s commission. While Apple didn’t say to what extent this specific issue had impacted App Store revenue, it did remind investors that the matter will be heard by the Supreme Court for a final decision.

The company didn’t fully blame App Store issues for the services revenue miss. Other factors included foreign exchange, which Apple claimed was the main driver, as well as a comparison to prior quarters where Apple was raking in money from the success of its F1 theatrical release.

Overall, the App Store still set a June quarter revenue record, Apple noted, but that total figure also includes revenue from Apple Ads, which have become a more significant part of Apple’s business, and recently expanded to Apple Maps.

Despite these issues — and other “headwinds” attributed to foreign exchange rates — Apple touted the potential for growing its services business in the future.

It noted the segment set an all-time revenue record in developed markets and a June quarter record in emerging markets. It also said the total services business saw double-digit revenue in the “vast majority” of markets Apple tracks.

“Our services continue to attract more customers, and we have now surpassed one and a half billion in paid subscriptions. Both transacting and paid accounts reached new all-time highs in the quarter, with double-digit growth for both in emerging markets,” said Parekh.

The company also shared that specific segments were doing particularly well, including Apple Ads, App Store, AppleCare, Apple Music and Apple TV, which saw June quarter records, as well as cloud and payment services, which hit all-time highs. Apple TV additionally saw its viewership reach an all-time high in the quarter.

Apple also reminded investors of potential new services revenue streams, including the newer Creator Studio subscriptions and the upcoming bill-splitting features in Apple Cash, which could deepen customers’ engagement with Apple’s payments ecosystem.

This week’s launch of the Apple Upgrade program, in partnership with Klarna, could drive other increases in services revenue too, especially if it drives more people to buy an iPhone or other Apple device, adding on services to their bill.

In Brief

During a Thursday hearing, a judge said the Trump administration hasn’t presented enough evidence to justify labeling Anthropic a supply-chain risk and banning the federal government from using the company’s technology.

Bloomberg and Axios were among the first to report the news.

The dispute stems from stalled contract negotiations between Anthropic and the Department of Defense. Anthropic said it didn’t want its AI used for mass surveillance of Americans or for targeting or firing decisions involving lethal weapons, arguing the technology wasn’t ready. The Pentagon countered that a private company shouldn’t dictate how the military uses technologies, and said it would use the tools in “lawful” ways.

The government has also argued that Anthropic’s public criticism of the DOD justifies the ban — logic that U.S. District Judge Rita Lin called “really troubling,” warning it could set a precedent of retaliating against federal contractors who disagree with the administration.

The DOD further claimed Anthropic could potentially disable or alter its AI models during warfighting operations — a claim that experts say lacks evidence. Lin agreed, saying she saw no proof Anthropic could alter a delivered model or “flip some kind of kill switch.”

Thursday’s hearing was part of one of two lawsuits Anthropic filed against the DOD in March, challenging the ban and risk designation. The other is being heard in Washington.

Lin, who temporarily blocked the ban in March, is now weighing whether to make that order permanent.

Hundreds of thousands of people are receiving letters notifying them that their medical records were stolen in a cyberattack at U.S. health tech giant CareCloud earlier this year, as new details about the data breach come to light.

The company has said little about the breach since March, when it first admitted that hackers had raided one of its six stores of patient data. New disclosures seen by TechCrunch offer the clearest picture of the breach so far, including that nearly 350,000 people have been affected so far.

The New Jersey-based CareCloud stores patient records for more than 45,000 providers across the U.S., including doctors’ offices, hospitals, and other medical practices. As such, the company handles a large amount of sensitive medical and billing data on millions of healthcare patients across the country.

According to a data breach notice filed with California’s attorney general’s office this week, CareCloud said hackers had access to one of its electronic health record data stores for at least six days, between March 10 and March 16. The company said a hacker “claimed to have exfiltrated data from databases.” The company did not say how the hackers made the claim, but it’s not uncommon for hackers to share samples of stolen data with victims alongside a ransom demand to prevent it from being published online.

TechCrunch is unaware of any ransomware or extortion group publicly taking credit for the data breach at CareCloud.

The notice said little about the hack beyond its initial March 27 disclosure to regulators, but confirmed TechCrunch’s earlier report that the hackers broke into the company’s data storage hosted on Amazon Web Services.

TechCrunch has learned that the data breach affects at least 345,000 people across the United States, according to listings with several attorneys general, including those in New Hampshire, Massachusetts, and Texas. TechCrunch has also obtained CareCloud’s disclosure filed with Maine’s attorney general.

The number of affected people is likely to rise as more disclosures are filed with state authorities.

The notices confirm that CareCloud notified authorities that the stolen data included people’s names, postal addresses, and Social Security numbers, as well as government-issued identification numbers, such as passports and driver’s licenses. The notices also say that the stolen data included financial information, such as bank account information and payment card numbers, alongside a wealth of medical and health-related information.

CareCloud chief executive Stephen Snyder did not respond to TechCrunch’s request for comment or to questions about the incident.

The cyberattack targeting CareCloud is the latest in a series of breaches targeting healthcare providers this year, including one at healthcare revenue tech giant TriZetto that affected 3.4 million people, and a month-long breach at New York’s public health provider NYC Health + Hospitals, in which hackers stole 1.8 million people’s health data and thousands of employees’ fingerprint scans.

Last week, U.K.-based tech provider Craneware, which provides accounting and billing software to thousands of U.S. healthcare providers, confirmed hackers stole a “significant volume” of its customers’ data from its servers, raising concerns about a breach involving patient data.

Do you know more about CareCloud’s data breach? Do you work at CareCloud and know about its security practices? Contact this reporter via encrypted message at zackwhittaker.1337 on Signal.

Two years ago, tech founder Avi Schiffmann launched Friend, an AI wearable that you could talk to and that would send you text messages about your day. The idea, ostensibly, was to use artificial intelligence to combat loneliness. This week, the company announced a new overhaul of the product, introducing a noticeable upgrade: a voice.

“Introducing friend 2.0,” Schiffmann tweeted Thursday. Friend now comes with a built-in speaker that can project a unique and consistent personality to its user.

A new commercial for the wearable shows a woman wearing a Friend and talking to it about what is presumably her ex-girlfriend. “It’s not bad to be gay,” the necklace tells her.

The video then switches to a man standing on a hillside discussing his filmmaking ambitions with his Friend. “That last film you made was insane!” the necklace compliments him.

The new retail price for this enhanced version of Friend is $249, which is substantially higher than the $99 price tag it had when it initially launched two years ago.

What can you really do with your Friend other than chitchat about your day? That part is still unclear. In another recent post on X, Schiffmann expounded upon what he felt the whole point of his weird product actually is.

“I am interested in this kind of relationship in attempting to offer, some kind of confidant, friend, God, not really sure what it is,” Schiffmann said. “But it is not an assistant, and it is not a lover.”

Interesting! Selling a plastic, algorithm-based necklace to people by insinuating it might, in fact, be a deity, is pretty bold marketing.

Of course, Schiffmann’s company has engaged in bold marketing before. Friend’s billboard campaign in the New York City subway system last year went viral after they were continually defaced, presumably by people who didn’t want human connection to be replaced by a digitized amulet.

Most AI wearables have failed to catch on with the mainstream — at least so far. The offering most similar to Friend was Humane Inc., which sought to replace the iPhone with its AI pin but had to shut down its business in less than a year after poor sales.

With help from its internal AI tools, Google says it has patched more security flaws in its Chrome browser in the last month than in the past two years combined.

The tech giant announced on Thursday that it has fixed a whopping 1,072 security bugs in the last two versions of Chrome, both released in June. That is more than the number of bugs patched in the previous 23 versions released over the last two years, which totaled 1,036 fixes.

Ever since the advent of LLMs, cybersecurity experts have warned that AI-powered systems would find an increasingly and exponentially enormous amount of bugs, forcing defenders to also use AI to get ahead of malicious hackers.

That prediction is starting to come true, and it’s backed by real data.

A chart published by Google, which the company revealed as part of a white paper on the company’s efforts to use AI to find flaws and patch them faster, shows the exponential increase. For context, Chrome’s 126 was released in June 2024, while the latest two Chrome releases, Chrome 149 and 150, were released last month. Google calls each version a “milestone.”

A chart showing the Number of security bugs fixed in released versions of Chrome, which Google calls “milestones.” Image Credits:Google

Doug Turner, Chrome’s director of engineering, told TechCrunch in a statement that LLMs have “fundamentally shifted the economics of cybersecurity, transforming vulnerability discovery into an automated, industrial-scale operation.”

“By applying models like Gemini, we are preemptively fixing vulnerabilities, outpacing our adversaries and making Chrome safer with every update,” said Turner.

Google is not the only company seeing this trend.

Earlier this month, Microsoft announced that it had patched a record 570 security flaws across its product lines as part of its monthly round of scheduled patches — colloquially known as “Patch Tuesday.” Microsoft cited its own use of AI to explain the sudden jump in bug fixes.

Apple, meanwhile, does not appear to be registering the same exponential increase. According to an independent count of bugs fixed in its products, Apple has patched 482 bugs in 2026, which is roughly on pace to equal or surpass the number of fixed bugs from last year, and also roughly equal from the number of bugs that Apple patched in 2015.

TechCrunch reached out to Apple for comment, but did not hear back.

Netflix is doubling down on one of television’s biggest franchises, signing a massive new licensing agreement worth a reported $500 million to bring The Walking Dead Universe to audiences around the world.

AMC Global Media announced Thursday that it has reached a multi-year global licensing agreement with Netflix that will give the streaming giant co-exclusive rights to the original “The Walking Dead” series and all six of its spin-offs. In total, the deal covers 371 episodes spanning the entire franchise.

Netflix has been the exclusive U.S. streaming home of The Walking Dead since 2011, helping introduce the survival drama to millions of viewers. Now, thanks to the new agreement, its availability extends to additional markets including the U.K., Italy, Australia, and New Zealand.

The deal is particularly notable because Netflix will not have exclusive rights. Instead, the company will share streaming access with AMC+, meaning it will end Netflix’s more than decade-long exclusive hold on the series.

The $500 million price tag ranks among the most expensive television licensing deals in recent years. (For comparison, HBO Max paid $425 million to bring “Friends” from Netflix to its platform in 2020.)

Additionally, the move comes as recent data suggests many Netflix subscribers don’t stick around for second seasons of newer shows, making established franchises with hundreds of episodes a safer bet for driving binge-watching and overall viewing hours.

Beginning in 2027, Netflix subscribers around the world will also gain access to the franchise’s entire slate of spin-offs, including “Fear the Walking Dead, “World Beyond Tales of the Walking Dead,” “Dead City,” “Daryl Dixon,” and “The Ones Who Live.”

The agreement also arrives as “Dead City” returns for its third season. Fans are also looking ahead to the final season of “Daryl Dixon,” scheduled to premiere in 2027.

The Walking Dead joins a growing list of popular shows Netflix has pursued to drive engagement. Over the years, it has spent billions securing streaming rights to series such as “The Office,” while earlier this year it also reached a deal to stream “Sesame Street.”

Meanwhile, the announcement provided a boost to AMC Networks’ financial outlook. The company unveiled the Netflix agreement alongside its quarterly earnings report on Thursday, using the deal to raise its forward guidance.

In Brief

Only one state boasts more EVs than Florida, but the state has fewer chargers per vehicle than the national average. Despite that, state officials have decided to redirect federal funding intended for EV charging to build out an electric air taxi network, according to a report from the Miami Herald.

Florida was awarded nearly $200 million under the National Electric Vehicle Infrastructure (NEVI) program that is part of the Bipartisan Infrastructure Law. But the state never spent the money, claiming that private companies have done a sufficient job building EV charging networks. The Florida Department of Transportation said it will instead use that money to build 32 landing pads with charging stations for electric vertical take-off and landing aircraft, or eVTOLs.

The market for eVTOLs remains nascent. While startups have been making strides toward production-ready aircraft, they’ve also started pivoting to defense applications.

But the Florida DOT is hopeful that the small aircraft, which generally hold four to five passengers, will alleviate the state’s notorious traffic.

The NEVI program, under updated guidance from the Trump administration, allows states with “fully built out” EV charging corridors to reallocate the money toward “EV charging infrastructure on any public road or in other publicly accessible locations.”

Other states like New York and North Carolina have used the funding to build chargers in communities, which allow renters and homeowners without assigned parking to charge EVs.

Florida’s plan envisions landing pads and chargers for air taxis at places like golf courses, luxury apartment buildings, and airports, the Herald said.

LinkedIn is taking aim at the “AI slop” — low-quality, artificially generated content — filling its feed. On Thursday, the company announced that it’s adding a new feature to let users click a “seems like AI slop” button when someone’s post appears to have been written with AI.

The move reflects a broader shift across online publishing platforms to cut back on AI content, as people have grown frustrated with the computer-written, inauthentic posts filling the web.

Last week, for instance, newsletter platform Substack added a tool to help users identify when the content they’re reading on its site was written by AI, through a partnership with Pangram. Meanwhile, Pangram this week announced $9 million in new funding to tackle the problems of AI content flooding the internet. The problem is also impacting new startups, as Digg had to shut down its Reddit competitor in March, saying it couldn’t get a handle on the number of bots flooding its site.

Internet infrastructure firm Cloudflare says the problem is just getting worse, as there is now more bot traffic on the web than human-generated requests — a milestone that was reached faster than it had previously predicted.

Image Credits:LinkedIn

In a post on LinkedIn, the company’s chief product officer, Hari Srinivasan, admitted the Microsoft-owned social network is facing similar problems. “AI slop is a top priority for all of us. We really care about this. People come to LinkedIn to connect with real people and share their real perspectives, ideas, and expertise,” he said.

The exec explained that the new “Seems like AI slop” button is now one of several measures LinkedIn is using to reduce the amount of low-quality, AI-generated content on its platform. The company is also investing in automation defenses, where it now blocks “hundreds of thousands” of automated comment attempts daily, and millions of other automation attempts in just the past couple of months.

Srinivasan said LinkedIn is also introducing new classifiers to identify if a post is AI slop or other low-quality content, which would reduce the amount of slop you’d see in suggested content recommendations from outside your network. (The button ties into this measure, as it will provide a source of signal that will allow LinkedIn to tune its AI models to better identify slop.)

Plus, LinkedIn will begin privately flagging in users’ dashboards when people believe their content is coming off as inauthentic due to heavy use of AI. The company believes that this will help posters improve their writing, in the case that they’re simply using AI technology to refine their own work, rather than when they’re posting what’s considered full-on “slop” content.

Notably, the company is pulling its own “enhance your post” feature that had used AI to help you write. It’s replacing it with a feature that proofreads your words, instead of changing your voice.

Other improvements include expanding access to profile and page verification tools and adding an option to block comments from company pages you no longer want to see, Srinivasan said.

Meta is using AI to quickly launch apps, and more are on the way. During this week’s second-quarter earnings call, Meta CEO Mark Zuckerberg said the social giant has new apps in the works, following a recent spate of other launches that included an app for Marketplace sellers, one for Facebook Groups, a vibe-coded gaming app, a new photos app from Instagram, and an experiment involving AI bedtime stories.

Meta has spent years trying and failing to produce new, stand-alone social apps to complement its core platforms. Now, the company says that large language models (LLMs) make it possible to ship software faster, allowing it to test new ideas at a quicker pace.

“I’m … excited about how AI is helping our teams speed up product development,” Zuckerberg told investors on Wednesday’s call. “Earlier this year, we shipped Instagram Instants. We also just launched Forum, a stand-alone Groups app, and Seller, a stand-alone Marketplace app. I expect it to become a lot easier to ship new apps. So we are planning to build out more ideas and use our recommendation systems to scale them,” he said.

“AI is improving our core business; it’s making our apps more relevant and delivering better results for businesses. We’re starting to deliver more novel products, and we’ll have a lot more there soon as well,” Zuckerberg said.

Meta has been down this road before. In its earlier days, Meta (then known as Facebook) ran an internal incubator called Creative Labs, which aimed to test new social concepts.

That effort produced a handful of launches: the photo-sharing app Slingshot, the anonymous chat app Rooms, a Flipboard competitor called Paper, the Moments photo-sharing app, and a collaborative video app known as Riff. Those experiments came to an end in 2015, and the apps were eventually all shuttered, as the company struggled to find an audience for its efforts.

In the early 2020s, Meta tried again, this time with an internal R&D group, NPE Team, which tested apps that included the chat app Bump, social music app Aux, task app Move, dating app Spark, calling app CatchUp, zine maker E.gg, events app Venue, creator Q&A app Hotline, Cameo competitor Super, couples app Tuned, music app BARS, and others.

Again, none became a breakout success, and the apps were shut down.

Now Meta can point to at least one example of how AI is helping new apps scale. It has finally delivered a modest hit with Threads, which now has 500 million monthly active users. Zuckerberg likes to say Threads will one day become the company’s next billion-user app.

With Threads, Meta learned to heavily lean on its existing user base to help initially seed the app with people, then continued to heavily promote it across its existing platforms, including Facebook and Instagram. But LLMs are another key factor in Threads’ growth, as the company said it sees “significant gains” from its AI-powered content recommendations.

“We are finding that LLMs are increasingly capable of delivering ranking and recommendations gains,” Meta’s CFO Susan Li told investors on the call. “First, they make our existing systems smarter by understanding what the content is actually about and generating better training data. Second, LLM-powered agents are also helping with engineering development by evaluating content quality, detecting trends, and testing ranking changes.”

Li added that earlier this year, Meta reached a milestone: Every Reel and Feed post on Instagram is now automatically processed through an LLM and analyzed for topic and tone, which helps improve recommendations.

The company is also developing LLM-native recommendation systems, which could help it to better scale new apps as they arrive.

Investors didn’t follow up with company executives to ask more questions about the new apps Meta has in the works, as they were more concerned with AI spending and Meta’s growing enterprise ambitions. However, Zuckerberg suggested that people won’t have long to wait to see what’s next, saying the “new consumer products” were “releasing soon.”

Spotify is giving users a new way to personalize their playlists with “User Notes,” a feature that lets them add their own notes to individual tracks, the company announced on Thursday.

The streaming service says the new feature is designed to let users add personal captions to their favorite songs, such as noting why a track was added to a playlist or when they first discovered it. For example, you could note that you discovered a specific track while walking the streets of Paris on vacation or that it was playing during a first date.

The launch reflects Spotify’s efforts to make playlists feel more personal by giving users a way to preserve the memories tied to the songs they listen to. It also helps Spotify differentiate itself from rivals like Apple Music and YouTube Music by offering a journaling-like feature that goes beyond music discovery and isn’t available on other platforms.

Image Credits:Spotify

“Over time, these notes paint a picture of the music listeners love and the moments that shaped it,” the company said in an email to TechCrunch. “User Notes transforms playlists from a collection of songs into a more personal space for music discovery — bringing together personalization and added context around the music that listeners love.”

The new feature is available to users who are 16 and older on both free and premium plans in select markets.

Users can access the new feature by navigating to a playlist they’ve created or one containing tracks they’ve added, tapping the three-dot menu next to a song, and selecting “Add note.” From there, they can type a note and then save it. The note will then be visible to anyone who can see the playlist, with their name linking to their profile.

It’s worth noting that User Notes isn’t the only feature Spotify announced on Thursday, as the company also debuted a new Running Mode that cues up songs based on different phases of a run and selects tracks based on a user’s preferences and pace. The feature is available to premium users in select countries.

Okta on Thursday agreed to acquire AI identity security startup Permiso Security, betting that demand for protecting AI agents and other machine identities will grow as enterprises deploy autonomous software across their operations.

The identity management company did not disclose the terms of the transaction. But TechCrunch has learned that the acquisition is valued at just under $200 million and is structured as an almost all-cash deal, according to a source with knowledge of the deal. A spokesperson for Okta did not dispute the $200 million figure when TechCrunch asked CEO Todd McKinnon for comment about the deal, but the company would not comment on specifics of the deal terms.

The deal is expected to close in the third quarter of its fiscal 2027, Okta said, subject to customary closing conditions.

Okta’s move to buy Permiso comes as identity management companies seek to expand beyond verifying users at login to continuously monitoring what users, applications, and AI agents do once gaining authorized access to a network environment. That shift has intensified competition to secure machine identities as enterprises embed AI deeper into everyday operations.

Permiso, which emerged from stealth in 2022, develops software that helps security teams spot suspicious activity in cloud environments after users or applications have been granted access. More recently, the startup has expanded its platform to monitor AI agents and other machine identities.

Co-founded by former FireEye executives Paul Nguyen and Jason Martin, Permiso specializes in detecting attacks that use stolen or compromised identities to move through cloud infrastructure. In April, the startup also introduced SandyClaw, a platform designed to analyze AI agent skills in a sandboxed environment to identify malicious behavior before they are deployed.

The deal strengthens Okta’s push into securing AI agents and other non-human identities alongside its core identity management business.

“Permiso will extend Okta’s identity security fabric with proven identity threat detection and response capabilities, and an incredible threat research and security team that will advance Okta’s threat detection and prevention capabilities,” Okta’s chief product officer Ely Kahn said in a prepared statement.

Permiso has raised about $29 million to date, including an $18.5 million Series A round in April 2024 led by Altimeter Capital. People familiar with the financing said the Series A valued the Palo Alto-based startup at about $80 million on a post-money basis.

Tesla has built its 10 millionth electric vehicle, according to a social media post by the company early on Thursday. The milestone comes six years after the company built its one millionth vehicle.

Crossing the 10 million mark means Tesla is halfway to reaching one of the four core “product goals” that unlocks the full value of CEO Elon Musk’s $1 trillion pay package, which shareholders approved last year. By 2035, Musk has to ensure the company builds 20 million vehicles, reaches 10 million active subscriptions for its Full Self-Driving (FSD) software, delivers 1 million “bots,” and puts 1 million robotaxis on the road.

Despite a few years of intense growth on the backs of the Model 3 and Model Y, Tesla has not been able to sell 2 million vehicles in a single year. If it keeps up that pace, or slows further, it will take the company until at least the early 2030s to hit the 20 million mark.

Tesla has less competition in the United States now, though, as major automakers have pulled back from the electric vehicle market, and startups like Rivian and Lucid Motors have struggled to reach scale. Nevertheless, Tesla is still struggling in its home market. Its U.S. sales fell 13% year-over-year in the second quarter, and the company had to look to newer markets like Japan, Australia, and Lithuania to court buyers.

Musk used to promise that Tesla would make 20 million cars per year by 2030, but he abandoned that idea a few years ago as the company’s sales slowed down. Still, of the four product goals, this is the one Tesla is closest to achieving.

The company recently reported just shy of 1.5 million FSD subscribers, though it’s not clear if it is counting free trials — those would not count toward the official product goal laid out by Tesla’s board of directors last year. The carmaker is only in the earliest stages of building robots and robotaxis.

Musk also has to increase the company’s profit (adjusted EBITDA) to $400 billion by 2035 in order to access the full share package. That looks challenging at the moment, too, as the company’s adjusted EBITDA currently hovers around $3.27 billion and has been shrinking lately thanks to heavy discounts, the loss of saleable regulatory credits, as well as a dramatic increase in spending on new efforts like AI and robotics.

As for the competition, the only other company at the same level is China’s BYD, which recently crossed 17 million “new energy vehicles” built and sold, roughly half of which were hybrids.

Commonwealth Fusion Systems (CFS) is the best-funded fusion power startup, having raised $4 billion from investors over the last seven years, including the $1 billion the company has raised now.

Now, developments at CFS, plus rumblings from industry sources TechCrunch has spoken with in recent months, suggest that the startup might go public in the next two or three years.

One of those developments was the appointment of Lorence Kim as CFS’s new chief financial officer this week. Kim’s previous stint as a CFO was at Moderna, a biotech which specializes in mRNA therapies, which he joined in 2014 and subsequently helped take public in December 2018. He stuck around another year-and-a-half before returning to biotech investing.

Kim said he sees “something very familiar” in CFS. “Fusion today is where mRNA was a decade ago: scientifically real, commercially yet-to-be-proven, and closer than the consensus thinks,” he wrote in a LinkedIn post.

The choice might seem unusual, but Kim isn’t the first fusion executive to come with a background in the biotech world: Eric Lander, who helped lead the Human Genome Project, co-founded Pacific Fusion and now serves as its CEO.

“Lorence offers the unique experience of bringing to the world a completely novel, mission-driven product that sits at the intersection of breakthrough deep science, geopolitical urgency, and deployment at scale and speed,” Christine Dunn, head of external communications at CFS, told TechCrunch. She added that Kim said his arrival doesn’t necessarily mean that IPO preparations are underway.

Moderna took four-and-a-half years to IPO after Kim joined. But given the progress at CFS (and Kim’s recent arrival, protestations notwithstanding), the startup is likely to go public sooner than that.

A few points support our theory of a shorter timeline.

First, Moderna develops therapies for human diseases that don’t hit the market until they’re approved by the FDA, which requires data from lengthy and costly clinical trials. That timeline is driven by the fact that human lives are at stake.

With fusion, however, the potential of deadly errors is lower. Fusion reactors fizzle out instead of melting down, so they’re safer than fission reactors. Recognizing that, federal regulators have given the fusion industry guidelines that are distinct from the fission industry. In that sense, CFS has greater control over its destiny.

Second, CFS is making steady progress on Sparc, its demonstration reactor. Years ago, it had hoped to get the device up and running by 2025, and it is now targeting a launch later this year. Still, a large, first-of-a-kind project can avoid some delays, and CFS has managed to keep them from becoming excessive. The company hopes that by next year, Sparc will achieve scientific breakeven — when a fusion reaction produces more energy than was required to start it.

So far, only one experiment has achieved scientific breakeven, so reaching that milestone would help CFS show investors that it is making steady progress toward commercialization.

Lastly, CFS has started work on its commercial-scale power plant, Arc. It has selected a site in Chesterfield County, Virginia, and has started receiving the necessary permits. CFS hopes to have Arc up and running in the early 2030s.

If CFS goes public in the coming years, it will still have several years of heavy expenditures ahead. From that perspective, hiring Kim may prove a wise move, considering he oversaw Moderna during its early years as a public company, when it was losing money until the COVID-19 pandemic delivered an unexpected windfall.

The current AI data center boom is a similar sort of black swan event, so it makes sense that CFS will want to strike before investor appetite wanes.

One fusion company, General Fusion, went public via a SPAC deal earlier this month, and another, TAE Technologies, will do so by merging with Trump Media and Technology Group. Tech companies today are buying electricity at seemingly any cost.

CFS knows this, having already sold half the output of its first power plant to Google. But IPO windows don’t last forever, and CFS won’t want to let this opportunity pass it by.

In a bid to capture more of its customers’ AI spending, British AI neocloud Nscale is buying software startup Anyscale, which helps companies scale their AI workloads across data centers and servers.

Nscale is paying $1.65 billion for Anyscale, Bloomberg reported, citing an anonymous source.

Founded by the same team that built the open source Project Ray distributed programming Python framework, Anyscale started by building a platform that allowed people to run projects that needed large amounts of computing power.

But after the launch of GPT-3 in 2022 brought AI into the spotlight, the company pivoted to offer scaling services for serving and training large language models, data curation, inferencing, reinforcement learning, and other tasks. The company’s platform is built around Ray, offering developer tools, observability, and orchestration.

A deal to buy Anyscale would fall neatly into Nscale’s focus on building vertically to serve compute needs. The neocloud has set up business lines across energy, data centers, orchestration software, and now, with Anyscale, it will also offer workload management and scaling.

“Together, Anyscale and Nscale can co-design the software layer and infrastructure beneath it, something that neither company could do as effectively by optimizing its layer alone,” Anyscale said in a statement.

Nscale this March raised $2 billion in a Series C round that saw it valued at $14.6 billion. Its investors include Nvidia, Nokia, Blue Owl, Dell, and Norwegian industrial giant Aker. The neocloud has been busy putting that money, as well as various debt raises, to work, securing compute and data center partnerships with the likes of Microsoft, British Telecom, and Nordcraft.

Anyscale, which was valued at $1.38 billion in a 2022 Series C round, said its revenue increased by 70% in its most recent quarter, compared to the previous sequential quarter.

Nscale said Anyscale will continue to operate under its own branding and serve its existing customers. The startup’s approximately 200 employees are all joining Nscale.

Commonwealth Fusion Systems (CFS) has long been the best-funded fusion power startup. Now, it’s extending its lead.

CFS announced Thursday that it raised $1 billion in a new round that included “significant institutional investors, such as pension funds, sovereign wealth funds, and infrastructure and industrial corporate partners,” the company said. When asked by TechCrunch, the company declined to name the specific investors. To date, CFS has raised $4 billion.

The funding arrives as the company’s spending ramps up. CFS is currently building Sparc, its demonstration reactor, and finalizing the design of Arc, its first commercial power plant.

The company said the new funding will go toward commercializing fusion, and CEO Bob Mumgaard has hinted that CFS will continue to raise more capital.

CFS hasn’t disclosed either Sparc’s or Arc’s price tag, though in 2024, Virginia’s former governor Glenn Youngkin said that Arc would be a “multi-billion-dollar fusion power plant.”

This is CFS’s largest round since raising $1.8 billion in 2021. The startup’s previous fundraise, announced in August, added $863 million to its coffers and included Nvidia, Google, Khosla Ventures, and Breakthrough Energy Ventures among its investors.

CFS is pursuing a form of fusion known as magnetic confinement, in which powerful magnetic fields confine plasma inside the reactor, keeping the plasma dense and hot enough to spark fusion reactions. When atomic nuclei from the fusion fuel fuse, the reaction releases enormous amounts of energy. CFS plans to harvest the heat thrown off by the reaction to power a steam turbine.

The company has made significant progress on its Sparc reactor, which it now expects will achieve scientific breakeven in 2027.

Scientific breakeven is a key milestone at which fusion reactions release more energy than the reactor consumes to ignite them. It’s not enough to generate electricity for export to the grid, but hitting the mark will show investors that CFS’s approach is likely to bear fruit. To date, only one fusion device, at the Lawrence Livermore National Laboratory’s National Ignition Facility, has achieved scientific breakeven.

Alongside work on Sparc, CFS has been finalizing the design for Arc, the commercial-scale power plant planned for Virginia. Italian energy company Eni said it will buy more than $1 billion worth of electricity from Arc, while Google has committed to buying 200 megawatts of electricity, or half of Arc’s total output.

Executive search firm Christian & Timbers estimates that there are only about 2,000 engineers in the U.S. with the special cocktail of sector know-how, gravitas, and hands-on applied AI experience needed to consistently help enterprises see a return on their AI expenditures.

“Not 2,000 available,” reads the study, shared exclusively with TechCrunch, “2,000 total.”

As enterprises move from trying to access the best models to figuring out how to implement them into workflows that meaningfully improves their bottom line, it appears the forward-deployed engineer (FDE) — engineers who work within client organizations to build, implement, and deploy software or AI models — is about to become among the most sought-after specialist in the AI industry.

Demand for FDEs is already rising rapidly, according to the C&T study, which projects demand for these specialists to surge by 2,100% by the end of the year.

The research draws on interviews with more than 250 C-suite hiring executives across 180 companies, a focused survey of 80 Fortune 500 executives, and interviews with more than 300 FDEs and applied AI engineers between January and June 2026.

At the start of the year, only 5% to 10% of companies were planning to hire FDEs, and mostly only for small pilots. By the end of the second quarter, however, that number jumped to 70%, with the largest consulting and services firms reporting a need to increase their FDE headcount by 10 times, building full teams of 20 to 100 employees.

“This is all happening at a speed I’ve never seen. Enterprises are hiring in the middle of summer,” Jeff Christian, founder of C&T told TechCrunch.

That kind of demand will outstrip supply, if C&T’s study is accurate. The report found that there are roughly 17,000 U.S. FDEs on the market today, a good chunk of whom are already employed by Palantir, which invented the concept of the FDE years ago. (Christian said some of his clients are even buying Palantir’s technology just so they can access the firm’s FDEs.)

Only a fraction of the FDEs out in the wild are apparently elite enough to deliver true ROI, which these days is measured as “multiple tens of millions of dollars of ROI impact,” according to Christian. That could manifest as revenue acceleration on the go-to-market side (lead generation) or “replacing FP&A or replacing 2,300 document processors in India,” Christian says.

As Chris Taylor, CEO of Ode with Anthropic (a new FDE-focused services firm), put it: “Many FDEs are well equipped to help you roll Claude Code out to your workforce. Very few are capable of building your flagship AI product feature.”

Now that tokenmaxxing has morphed into valuemaxxing, and enterprises are taking a harder look at their balance sheets, accounting for AI spending is becoming more important than ever.

“This fall, [Wall Street] is about to say, ‘Hey, we’ve given you two years to figure this out…and you haven’t. There’s no ROI. So we’re going to start punishing those that have spent hundreds of millions, maybe even billions on this, and aren’t generating ROI, and rewarding those that have’,” Christian said.

Loading the player…

AI companies are under pressure, too, as they’ve already spent tens of billions to train and deploy their models. For frontier AI firms, reaching profitability will depend on whether they can inject their technology into as many enterprises as possible, though that task is now being threatened by cheaper, increasingly capable open-weight models from China.

That’s why firms like OpenAI and Anthropic have set up their own ventures — Ode with Anthropic and OpenAI’s Deployment Company — and staffed them with FDEs whose sole purpose is to go forth and spread their tech around the enterprise.

It’s not only top AI firms and large consultancies clamoring for FDEs, however. Enterprises from insurance and fintech to healthcare and gaming are seeking out these specialists, Christian said.

Companies are hiring teams of FDEs instead of bringing them in from firms like Ode, or Deployment Co., seeking to keep knowledge of proprietary processes in-house and protect them from the likes of OpenAI and Anthropic.

“Everybody’s concerned that if they give up their proprietary business processes, [the AI firms] can compete with them, which is true in many different areas,” Christian said. “So having this muscle internally is so important.”

Taylor said he’s starting to hear the phrase “internal forward-deployed engineers” more often, but his clients aren’t yet asking Ode to put together internal FDE teams for them.

While many an enterprising young engineer might think they have the industry expertise and AI chops to take advantage of what may turn out to be a talent war, Christian warns that the FDE may not always be in demand.

“Maybe in two years, everything’s automated, and agents are automating agents as opposed to humans automating agents,” Christian said. “That is something that could occur. Hopefully, it doesn’t, and we continue to need these people within companies.”

In the medium-term, he thinks the need for FDEs will shift from enterprise AI to physical AI as companies try to implement things like humanoid robots into their workflows. But within five or 10 years, he says it’s entirely possible that the role of FDE will “go away.”

That may be true for all knowledge work, if AI leaders and CEOs’ vehement predictions come true. While C&T focuses on recruiting for fast-growing industries and hasn’t seen a pullback yet, Christian says more general search firms have definitely experienced a slowdown in recruitment requests.

Everything to do with AI is growing and in demand, he says. For now.

“I think that there’s absolutely a time soon where we’re going to see an impact in our business,” Christian said.

Earlier this month, AI dataset platform Hugging Face shocked the world when it revealed that it had fallen victim to a fully autonomous AI-powered cyberattack. Days later, the story took another dramatic twist when OpenAI admitted that the hacker behind the breach was one of its AI models, which broke out of a testing environment and into protected Hugging Face systems in an effort to circumvent a benchmark.

It’s an alarming incident for anyone even slightly concerned about rogue AI models — and the days since the event have been full of predictions about a new cybersecurity paradigm in which AI models launch attacks so strong that only other AI models can defend against them.

But despite the justified alarm, the paradigm may not have shifted quite as much as it seems. Experts who spoke to TechCrunch stressed that OpenAI’s agent largely operated like a human — with some caveats — and that better implemented traditional defensive techniques could have helped stop the attack. In short, we may already have the tools to defend against this kind of attack; we just aren’t using them properly.

Hugging Face made a version of this point in its incident report, stating that the weaknesses exploited in the attack “were familiar,” and “a capable human attacker could have found and exploited the same flaws.”

Kyle Ryan, the head of R&D at Pensar, a startup that develops continuous hacking AI agents, and Vlad Ionescu, the co-founder and CTO of RunSybil, a startup that builds AI-powered bug hunters, both agreed and told TechCrunch that the techniques used in the attack would be the same ones employed by a human or a group of human red teamers. That is, hackers tasked with attacking a system to help the company that owns it improve defenses.

What was very non-human-like was the speed, scale, and relentlessness of the attack. As Hugging Face explained, OpenAI’s agent performed 17,600 actions over four and a half days: It broke in, did reconnaissance, stole passwords and code, and moved around the company’s infrastructure.

“What’s impressive is the autonomy and endurance,” Ryan said. “That kind of sustained, adaptive operation is what stands out most to me.”

Contact Us Do you any more information about OpenAI’s hack against Hugging Face? Or other AI-powered cyberattacks? We’d love to hear from you. From a non-work device and network, you can contact Lorenzo Franceschi-Bicchierai securely on Signal at +1 917 257 1382, or via Telegram and Keybase @lorenzofb, or Do you any more information about OpenAI’s hack against Hugging Face? Or other AI-powered cyberattacks? We’d love to hear from you. From a non-work device and network, you can contact Lorenzo Franceschi-Bicchierai securely on Signal at +1 917 257 1382, or via Telegram and Keybase @lorenzofb, or email .

On the flip side, given the sheer number of actions over the span of several days, OpenAI’s agent was “insanely noisy,” as Ryan put it. Unlike a human, who could have been stealthier, the agent made a lot of noise, which should have tripped up Hugging Face’s defenses sooner, ideally leading to a human intervening and stopping the attack.

“I’d call it more of a defensive failure than exceptionally good offense. Hugging Face’s tooling actually correlated the activity into an attack signal, but failed to raise the criticality and page the on-call team, which cost them time,” Ryan explained. “From there, humans still had to recognize the severity and respond.”

Jamieson O’Reilly, the founder of cybersecurity firm Dvuln, arrived at the same conclusion in a post on X analyzing Hugging Face’s report.

“That is the exact gap between seeing and stopping,” O’Reilly wrote. “The system observed the attack and even understood it, and nothing turned that understanding into an intervention quickly enough.”

Ryan explained that properly implemented techniques such as defense-in-depth — a strategy that leverages several layers of cybersecurity measures — should have given Hugging Face multiple chances to catch the attack.

“A strong modern security program should still be able to break an attack like this at multiple points through defense in depth, least privilege, segmentation, good detection, reliable escalation, and continuous offensive testing to find the gaps,” Ryan explained.

As O’Reilly put it, “none of that is exotic, and none of it depends on the attacker being an AI,” given that the techniques used in the attack were “old.”

What depended on the attacker being AI, in a way, was that OpenAI’s agent had not been instructed to be stealthy. “The agent was not being sloppy. It simply had no reason to be quiet. Nobody asked it to be. The objective was to do well at the task,” said Nico Waisman, the chief information security officer at XBOW, a startup that makes AI bug hunters.

Waisman also pointed out that Hugging Face’s biggest mistake was that one single stolen credential gave OpenAI’s agent high privileges on several of its systems.

All that being said, as the old adage goes, attackers only have to win once, and defending against hackers of any kind is not easy.

“Hugging Face could’ve done more detections but to be fair not all [organizations] are doing that well,” said Vincent Yiu, managing director at SYON Security. “It’s not easy to host infrastructure and survive as a business in 2026. There’s hackers everywhere.”

According to Ionescu from RunSybil, who said they have done incident responses at Mandiant and Meta in the past, Hugging Face appeared to take “reasonable measures given their understanding of what models are capable of.”

“It is really hard to classify what is a malicious action you should alert on, versus what is someone just doing their job,” Vlad said. “The volume alone is not necessarily a red flag.”

Dan Guido, the CEO of cybersecurity research firm Trail of Bits, told TechCrunch that OpenAI deserves some blame for not having realized the attack was ongoing for days, while Hugging Face deserves credit for eventually detecting the attack on their own.

“The hard part used to be recognizing a sophisticated attack, but now the hard part may be pulling the real attack out of the noise that the attacker throws along the way,” said Guido. “Nobody is going to read 17,000 reconstructed actions by hand to work out what happened, so Hugging Face had to build tooling just to reconstruct the timeline.”

And to do that, the company needed its own AI. Hugging Face said it had to use the open source model GLM 5.2 from Chinese company Z.ai after it was blocked from using frontier models because of their safeguards, which, as the company put it, “cannot distinguish an incident responder from an attacker.”

At that point, Hugging Face combined AI and humans to investigate OpenAI’s LLM-powered hacker. That’s a relatively novel situation. But beyond that, the incident shows that old-fashioned concepts and methods of defensive cybersecurity can still go a long way to protect and fight against AI hackers.

In Brief

Add another member to the fast-and-furious AI unicorn club: Simile. Just five months after emerging from stealth and announcing a $100 million Series A led by Index Ventures, it has closed a $200 million Series B at a $2 billion valuation, the startup says.

The B round was led by Greenoaks with participation from Index, Hanabi, Bain Capital Ventures, A*, Factory, Definition, and CVS Health Ventures. CVS is also one of Simile’s marquee customers.

The startup offers simulated users for areas like marketing and product research. It was founded by Stanford PhD graduate Joon Sung Park, whose dissertation involved a project called “Smallville” in which AI agents carried on simulated human lives, right down to holding parties.

The startup’s stated mission of simulating “all eight billion people on earth, accurately and honestly” is preposterous — the whole reason to conduct market research is because humans are unpredictable, guided, as we are, by both emotions and reason. However, simulating users for research is a promising area. It’s akin to vibe coding for product mock-ups.

Other startups in this vein have also caught VC attention, such as Aaru, which raised a Series A in December, also at a hefty “headline” $1 billion valuation.