請同意我們的隱私權規範，才能啟用聽新聞的功能。

（中央社紐約/舊金山22日綜合外電報導）根據路透社取得的內部備忘錄，Meta表示，公司在美國員工的電腦中安裝新的追蹤軟體，用於記錄滑鼠移動、點擊與鍵盤輸入情況，以協助訓練其人工智慧（AI）模型。

路透社報導，Meta指出，這是公司推動打造可自主執行工作任務的AI代理（AI agents）整體計畫的一部分。

根據其中一份備忘錄，這個名為「模型能力計畫」（MCI）的工具將在工作相關應用程式（App）與網站上運行，也會不定期截取員工螢幕畫面，目的是讓AI模型學習如何像人類一樣操作電腦，例如熟練地在下拉選單中點選內容、使用鍵盤快捷鍵，以及在不同工作流程之間切換等。

備忘錄寫道：「這是所有Meta員工都能透過日常工作幫助模型進步的方式。」

臉書（Facebook）與Instagram母公司Meta正積極將AI導入工作流程，並重塑人力結構，主張此舉能提升營運效率。

Meta技術長博斯沃斯（Andrew Bosworth）對員工說：「我們的願景是由AI代理負責主要工作，而人類的角色是指導、審核並協助其改進，透過觀察人類介入行為，讓AI在下一次表現更佳。」

他並未明確說明將如何訓練這些AI代理，僅表示Meta會「嚴謹地建立各類工作互動所需的數據與評估機制」。

Meta發言人史東（Andy Stone）證實，MCI蒐集的數據將成為訓練資料來源之一。

●AI重塑勞動力

史東表示，MCI蒐集的數據僅限用於訓練AI模型，不會用來評估員工績效或做其他用途，並稱公司已設置保護「敏感內容」的機制，但未說明哪些數據將被排除。

他說：「若要打造能夠協助人們完成日常工作的AI代理，模型必須理解真實使用情境，例如滑鼠操作、點擊按鈕與瀏覽下拉選單等。」

今年以來，美國大型企業普遍出現以AI取代部分人力的趨勢。AI工具能在有限人力監督下完成應用程式開發與大量數據整理等複雜任務，已在矽谷引發熱潮，也促使部分企業高層計劃大規模裁員。

Meta計劃自5月20日起裁減全球約10%人力，並考慮今年稍晚將進一步大裁員。

電商與雲端運算巨擘亞馬遜（Amazon）近月裁撤約3萬名企業員工，占其白領人力近10%；金融科技公司Block則在2月裁減近半數員工。

●監控員工工作疑慮

耶魯大學（Yale University）法學教授阿瓊瓦（Ifeoma Ajunwa）指出，鍵盤記錄與螢幕監控技術過去多用於調查員工不當行為或非工作相關活動，如今擴展至AI訓練用途，意味白領員工正面臨更高程度的即時監控。

她表示，「在美國聯邦層級，企業對員工監控並無明確限制」，僅部分州法要求雇主需事前告知。

多倫多約克大學（York University）法學教授史特法諾（Valerio De Stefano）則指出，歐洲法律可能禁止此類監控。在義大利等國，以電子監控追蹤員工生產力屬違法行為；德國法院亦規定，僅在涉及重大犯罪嫌疑等特殊情況下，雇主才可使用鍵盤記錄技術。

他補充說，這類做法可能違反歐洲「一般資料保護規則」（GDPR）。（編譯：劉文瑜）1150423

請同意我們的隱私權規範，才能啟用聽新聞的功能。

（中央社紐約22日綜合外電報導）美國Axios新聞網獲悉，過去一週，人工智慧（AI）軟體業巨擘OpenAI持續向美國聯邦機構、州政府以及五眼聯盟（Five Eyes）盟友介紹其新網路安全產品的能力。

報導指出，OpenAI昨天在美國華府舉辦活動，向聯邦政府大約50名負責網路防護的人員，展示其新一代GPT-5.4-Cyber模型的功能。該模型已於上週依分級存取計畫推出。

知情人士透露，若想加入OpenAI的「網路安全的受信任存取計畫」（Trusted Access for Cyber），政府申請者需通過與商業客戶相同的審核流程。

昨日活動的與會者包括來自不同政府部門及國安機構的官員，多數負責日常網路安全任務。

OpenAI推出新網安模型不久前，AI公司Anthropic亦發布了Mythos Preview，兩家公司都在跟政府機構合作，以確認誰能獲得存取權限。

Anthropic因網安風險並未公開發布Mythos，僅提供給大約40個企業及組織，當中至少有兩個屬於聯邦政府。

OpenAI則採取雙軌策略，一方面推出安全防護較強、可廣泛使用的版本，同時透過「網路安全的受信任存取計畫」，向資安防護人員推出另一個使用限制較寬鬆的版本。

OpenAI全球事務長黎漢（Chris Lehane）在昨天的活動中表示，這種做法將使更多公司，例如地方供水單位，也能取得先進AI工具。

OpenAI也在跟州政府合作，協助他們取得GPT-5.4-Cyber的使用權限。

值得注意的是，Axios獲悉，OpenAI本週開始向五眼聯盟成員簡報，協助其完成審核並獲得模型存取權。這個情報共享聯盟的成員包括美國、澳洲、加拿大、紐西蘭及英國。

透過 Google News 追蹤中央社

Axios指出，目前多數已獲得OpenAI模型及Anthropic Mythos Preview存取權的企業，正在利用這些工具，尋找自家內部系統的潛在資安漏洞。

此外，許多政府機關面臨舊有系統資安難以維護或修補的困境，先進AI工具有望協助他們更快發現重大漏洞。（編譯：楊昭彥）1150422

Tesla saw an uptick in revenue and profit year-over-year, figures buoyed by an increase in automotive revenue and other services, including active subscriptions to its Full Self-Driving (Supervised) advanced driver assistance system, which reached 1.28 million.

Tesla shares rose 4% in after-hours trading following the release of its first-quarter earnings report, driven largely by a jump in its free cash flow, as well as increases in revenue and profit on a year-over-year basis.

The company reported Wednesday revenue of $22.38 billion, a 16% increase from the $19.3 billion it generated in the first quarter of 2025. Its automotive revenue also rose to $16.2 billion, compared to $13.96 billion in the same year-ago period. Notably, the company reported positive free cash flow ‌of $1.44 billion, more than double what it held in the first quarter of 2025. The figure surprised analysts who had expected the company to burn through more cash in the first quarter.

That pop in revenue, which met expectations of analysts’ surveyed by Yahoo Finance, provided a bit of good news for the company, which has grappled with lagging EV sales. Tesla delivered 358,023 EVs globally in the first three months of the year, below analysts’ expectations of around 368,000. The company also produced 408,386 vehicles during that same period, far more than it delivered.

The company’s first-quarter revenue got a bump from higher average vehicle prices, services, and active FSD subscriptions, which grew 51% year-over-year to 1.28 million.

Tesla’s business hit considerable headwinds in 2025 causing profits to fall 46% year-over-year to $3.8 billion. The dip was primarily due to lower EV sales — a problem other automakers also faced after the Trump administration ended the $7,500 federal tax credit for electric vehicles.

Tesla’s first-quarter results, while positive in year-over-year terms, still shows some weakness when the previous three quarters are taken into account. The company’s fourth-quarter revenue was $24.9 billion and its third-quarter revenue was $28 billion, a figure propped up by consumers who bought an EV before the tax credit expired.

Techcrunch event Meet your next investor or portfolio startup at Disrupt

Your next round. Your next hire. Your next breakout opportunity. Find it at TechCrunch Disrupt 2026, where 10,000+ founders, investors, and tech leaders gather for three days of 250+ tactical sessions, powerful introductions, and market-defining innovation. Register now to save up to $410. Meet your next investor or portfolio startup at Disrupt

Your next round. Your next hire. Your next breakout opportunity. Find it at TechCrunch Disrupt 2026, where 10,000+ founders, investors, and tech leaders gather for three days of 250+ tactical sessions, powerful introductions, and market-defining innovation. Register now to save up to $410. San Francisco, CA | REGISTER NOW

The first-quarter results also illustrate a company that still relies on its traditional EV business, along with service and subscriptions, and is not yet benefiting from its future bets on AI and robotics.

Tesla’s net income was $477 million, compared to the $409 million in the first quarter of 2025. That Q1 2025 profit figure was notably off the mark, a 71% drop from the same period in 2024. Like the revenue story, Tesla’s first-quarter profits are still notably lower than the past three quarters. The company’s fourth-quarter profit was $840 million and its third-quarter income was $1.37 billion.

Tesla said a higher vehicle average selling price combined with an increase in vehicle deliveries, growth in services, and curiously, an increase in automotive one-time benefits related to warranty and tariffs boosted its bottom line.

Tesla CEO Elon Musk has repeatedly warned that the company is in an awkward and potentially financially painful transition from its core EV business to an AI and robotics company. It has yet to scale production of its Optimus humanoid robot, which will be produced at its Fremont, California factory, or meaningfully ramp up its robotaxi service. The company said preparations for its “first large-scale Optimus factory” will begin shortly in the second quarter.

The company currently operates a limited robotaxi service without a human safety operator in Austin. It recently started operating that service in Dallas and Houston, but access to those vehicles remains severely limited.

Earlier this month, NASA’s Artemis II mission sent four astronauts to orbit the moon and used new laser communications systems to beam dramatic images back to Earth.

One of the receivers, though, wasn’t hosted by the U.S. space agency. A low-cost terminal built by the companies Observable Space and Quantum Opus, and operated by the Australian National University, pulled down data broadcast from a spacecraft at the moon at a rate of 260 megabits per second.

That success proves that high-throughput connections between Earth and space can be done on the cheap, the companies say.

The terminal used Observable Space’s software and telescope to capture and lock onto the transmissions from the Orion spacecraft, and a photonic sensor built by Quantum Opus to decode the data. Their terminal cost less than $5 million, compared to more bespoke solutions that cost tens of millions of dollars.

NASA has been testing deep space laser communications for several years, including a demonstration of data links with a spacecraft 218 million miles from Earth on its way to an asteroid. Artemis II was its most comprehensive demonstration yet: NASA’s primary receivers in California and New Mexico, as well as the low-cost experimental terminal in Australia, all collected 4K video from the trip around the moon.

Though laser communications boast much higher throughput than radio frequency transmissions that remain the primary communications option for space, lasers are more vulnerable to disruptions from cloudy weather, and they must be within line of site of their target — hence the importance of having a reception site on the other side of the world from the U.S.

Josh Cassada, a former U.S. astronaut who co-founded Quantum Opus, pointed out that Australia was the first continent to appear in the first Earthrise photo captured by the Artemis II astronauts.

Techcrunch event Meet your next investor or portfolio startup at Disrupt

Your next round. Your next hire. Your next breakout opportunity. Find it at TechCrunch Disrupt 2026, where 10,000+ founders, investors, and tech leaders gather for three days of 250+ tactical sessions, powerful introductions, and market-defining innovation. Register now to save up to $410. Meet your next investor or portfolio startup at Disrupt

Your next round. Your next hire. Your next breakout opportunity. Find it at TechCrunch Disrupt 2026, where 10,000+ founders, investors, and tech leaders gather for three days of 250+ tactical sessions, powerful introductions, and market-defining innovation. Register now to save up to $410. San Francisco, CA | REGISTER NOW

Observable Space CEO Dan Roelker said the mission proves that space-to-Earth laser downlinks are ready to scale. Already in wide use for satellite-to-satellite connections, the technology hasn’t previously been used to transmit back to Earth because of the cost, but now he envisions a global network of these terminals to receive data sent from satellites of all kinds.

“We can scale this over the next year or more,” Roelker told TechCrunch, though he said the company isn’t ready to reveal its full strategy yet.

“We will partner with a lot of people around this,” he said, “whether this is something we’re going to do ourselves, or partner with other ground station-as-a-service companies, or work with extremely large constellation providers that are going to want to own their own infrastructure.”

Until a few hours before SpaceX announced its deal, giving it the option to acquire Cursor — the maker of AI-powered coding software — for $60 billion, Cursor was on track to close a $2 billion funding round later this week, according to a person familiar with the matter. The round would have valued the company at $50 billion. SpaceX said it would either buy the company at some point later this year or pay $10 billion to Cursor to collaborate on AI development.

Cursor was apparently running a parallel process, negotiating a potential acquisition by SpaceX while simultaneously finalizing a private funding round with investors that include Andreessen Horowitz, Thrive, Nvidia, and Battery Ventures, details of which were first reported by TechCrunch last week.

It is not uncommon for startups to engage in acquisition discussions while simultaneously raising new capital. While many private companies prefer to remain independent, Cursor’s $2 billion raise would have fallen short of the capital needed to reach cash-flow breakeven, likely forcing the company to raise substantial funding later, the person said.

SpaceX, which recently merged with xAI, has been aiming to beef up its AI capabilities to better compete with leaders like Anthropic and OpenAI. Acquiring Cursor gives Elon Musk’s company a better chance of challenging rivals in AI coding, currently the most lucrative application of the technology.

However, SpaceX is delaying the potential acquisition of Cursor until after its IPO this summer. This is largely because the company wants to avoid updating its confidential financial filings before the listing, and it will be easier to finance the $60 billion purchase using its new, publicly traded stock.

The deal appears to benefit both sides for several reasons.

Despite fast revenue growth, Cursor is facing fierce competition from Anthropic’s Claude Code and OpenAI’s Codex. Given that threat, the startup could face challenges in continuing to raise private capital to finance its massive computing needs. Even if SpaceX doesn’t go through with the acquisition, Cursor is receiving a $10 billion capital injection paid out over time from Elon Musk’s company.

Techcrunch event Meet your next investor or portfolio startup at Disrupt

Your next round. Your next hire. Your next breakout opportunity. Find it at TechCrunch Disrupt 2026, where 10,000+ founders, investors, and tech leaders gather for three days of 250+ tactical sessions, powerful introductions, and market-defining innovation. Register now to save up to $410. Meet your next investor or portfolio startup at Disrupt

Your next round. Your next hire. Your next breakout opportunity. Find it at TechCrunch Disrupt 2026, where 10,000+ founders, investors, and tech leaders gather for three days of 250+ tactical sessions, powerful introductions, and market-defining innovation. Register now to save up to $410. San Francisco, CA | REGISTER NOW

Additionally, if SpaceX goes through with the acquisition, the space giant will likely keep the entire Cursor team intact. Unlike Google’s purchase of Windsurf, which was structured as an acqui-hire of key individuals, SpaceX currently lacks a meaningful AI workforce and is widely seen as not having a significant AI business.

Meanwhile, SpaceX has access to vast computing capacity at its data centers in Mississippi and Tennessee, which it can offer Cursor, potentially in lieu of part of the $10 billion “collaboration” payment promised the coding startup.

The company would also like public investors to value it as more than just a space and satellite business. By promising to potentially acquire Cursor, SpaceX positions itself as an AI company, giving it a chance to garner the much higher valuation multiple that Wall Street currently assigns to AI companies.

In Brief

The French government agency that handles the issuing and management of citizens’ identity documents, including national IDs, passports, and immigration documents, confirmed Wednesday that it experienced a data breach.

In an announcement, the Agence Nationale des Titres Sécurisés (ANTS) said the data stolen in the breach could include full names, dates and places of birth, mailing and email addresses, and phone numbers on an undisclosed number of citizens. ANTS said the investigation to determine how the breach happened and its impact is ongoing, and people whose data was affected are being notified.

ANTS, which said it detected the attack on April 15, did not specify how many people were affected by the breach. But some reporting suggests millions may have had some of their personal information stolen.

According to Bleeping Computer, a hacker has advertised the stolen data on a hacking forum, claiming to have a database with 19 million records. The hacker’s forum post referenced the same kind of stolen information as mentioned in ANTS’ announcement and was published before ANTS publicly disclosed the breach on April 20.

At Google Cloud Next this week, the tech giant announced a bevy of new updates to Workspace, its subscription-based productivity suite aimed at professionals. As you might expect, the updates are heavy on AI, integrating new automation tools into various workflows—everything from drafting emails to organizing Google Sheets. Overall, the changes are clearly designed to give office workers a leg up in their pursuit of less busy work.

Here’s some of what is new:

Workspace Intelligence. Workspace Intelligence, a new AI system built into Google’s office suite, is designed to automate assistance across various tasks. The system draws on a user’s Workspace data, including their Gmail, Calendar, Chat, and Drive (Docs, Slides, and Sheets). Google has given users administrative control over what the AI system can see and access. Users can disable Workspace Intelligence’s access to particular data sources at any time. The tradeoff: the more data the system has access to, the more it’s able to assist in those particular areas.

Build and fill out Google Sheets with Gemini. A number of new features allow users to both build and fill out Google Sheets, the company’s spreadsheet tool. Users can construct sheets by prompting Gemini to construct them. Prompts can include things like formatting and data retrieval, allowing Google’s AI system to do much of the work a human would’ve previously needed to do. At the same time, Gemini also helps with data entry, automatically filling out Sheets with “prompt-based” filling. Google claims that its new feature allows users to populate the spreadsheets “9x faster” than manual entry, because the system is designed to infer what you’re going to enter. Another new Sheets feature allows users to convert unstructured data into organized tables.

AI writing capabilities. Google has also brought new new AI writing tools to Google Docs. Users can now use Gemini to “generate, write, and refine” documents. The feature is powered by the company’s Workspace Intelligence system, which draws on data from a user’s Drive, Chat, and Gmail archives, as well as the internet, to assist with editorial tasks. Users merely prompt Gemini to help them write or edit their documents. Users can prompt Gemini to “help me write” or ask it to “match” their writing style so that it can effectively mimic their voice.

Realizing that enterprise customers are where the money is, tech companies are racing to deploy the most convenient and efficient office tools—applications that can make the average worker’s life a matter of degrees easier. Google has a certain advantage; its office products are already deeply embedded in workplaces worldwide, giving it a built-in audience for these AI upgrades. But Microsoft, Apple, and a growing field of startups are all competing for the same turf.

Elon Musk said Wednesday that millions of Tesla owners will need hardware upgrades to run a future version of its Full Self-Driving software that doesn’t require human supervision — after he and Tesla spent years promising that wouldn’t be the case.

The admission could open the company up to legal challenges from customers who bought Teslas on the belief that the cars were just one software update away from being able to drive themselves.

Musk said on Tesla’s quarterly earnings call Wednesday that cars with the third version of Tesla’s driver-assistance tech, known as “Hardware 3,” will need a new computer and new cameras. He also said these replacements will be so burdensome that Tesla is looking at building “micro-factories” in “major metropolitan areas” so the company’s service centers don’t get overwhelmed.

“If it’s done just at the service center, it is extremely slow to do so, and inefficient. So we basically need, like many production lines, to make the change,” Musk said on the call.

Tesla owners with Hardware 3 cars have spent years bugging the company and Musk for a straight answer about whether they would be able to run this advanced version of Full Self-Driving — which, it should be noted, Tesla has not yet released, or proven it can. Tesla sold these cars between 2019 and 2023.

Musk finally copped to the reality that Hardware 3 was insufficient to run more advanced versions of Full Self-Driving in January 2025.

“I think the honest answer is that we’re going to have to upgrade people’s Hardware 3 computer for those that have bought Full Self-Driving,” he said. “That’s going to be painful and difficult, but we’ll get it done.”

Techcrunch event Meet your next investor or portfolio startup at Disrupt

Your next round. Your next hire. Your next breakout opportunity. Find it at TechCrunch Disrupt 2026, where 10,000+ founders, investors, and tech leaders gather for three days of 250+ tactical sessions, powerful introductions, and market-defining innovation. Register now to save up to $410. Meet your next investor or portfolio startup at Disrupt

Your next round. Your next hire. Your next breakout opportunity. Find it at TechCrunch Disrupt 2026, where 10,000+ founders, investors, and tech leaders gather for three days of 250+ tactical sessions, powerful introductions, and market-defining innovation. Register now to save up to $410. San Francisco, CA | REGISTER NOW

However, the company’s chief financial officer Vaibhav Taneja made a different claim just six months ago in October 2025 when he said Tesla may still be able to make it work. “We have not completely given up on HW3,” he said at the time.

Tesla said Wednesday it will still release slightly more advanced versions of the current Full Self-Driving software to Hardware 3 owners. But Musk was unequivocal that upgrades will be needed for anything more.

“Hardware 3 simply does not have the capability to achieve unsupervised FSD,” he said.

Bluesky isn’t the only company leaning into AI to help build custom feeds, it seems. Amid a slate of recent product releases, X this week announced the launch of Grok-powered Custom Timelines, which let you dive into one of over 75 specific topics through curated feeds that can be pinned to your home tab.

The company touted the feature as one of the “biggest changes” to the app to date, saying it uses Grok’s AI to not only build these custom timelines but also personalize them for individual users.

The custom feeds are arriving at the same time as X has announced it is shutting down X Communities, a feature that had allowed people to create their own member-based communities around various topics, but saw declining use.

Today we're announcing two product changes for organizing communities on X:



1. XChat now supports joinable links for groupchats. Create a public link & share direct to Timeline. With support for 350 members per chat (and growing), Groupchat Links are the fastest way to bring… pic.twitter.com/GNcRB99Opc — Nikita Bier (@nikitabier) April 22, 2026

On X, the company’s head of product, Nikita Bier, noted that the custom timelines work even better for topics you already engage with. A rep for X explained to TechCrunch that the custom timelines aren’t based on traditional signals like keywords or hashtags. Instead, the company said, Grok reads every post, understands it, and then adds topic labels. This is made possible by the AI models from Grok owner xAI, the company that acquired X last year, tying the two services even closer together.

At launch, the custom timelines are available only to Premium subscribers on iOS. Android support is in the works. All Premium subscription tiers can access this feature.

Image Credits:screenshot from X

To use the feature, simply scroll to the right past your “For You” and “Following” feeds on X, as well as any other personal lists you may have pinned. Then tap the plus ( + ) sign to choose which custom timelines you want to pin to your home tab. (Choose wisely, because you can only pin up to 10 topics or lists!)

You can also reorder your selected topics from the same screen.

Image Credits:screenshot from X

Once pinned, you can tap on any of the feeds from your home tab across platforms to browse your pinned custom feeds.

Notably, the second position in each feed was filled by an ad — which suggests X just found a way to increase its ad inventory. That matters: X’s ad business has reportedly been struggling since Musk’s acquisition, with conflicting reports about whether things have improved.

Ladies and gentlemen, today we're launching one of our biggest changes to 𝕏



Introducing Custom Timelines



This feature allows you to pin a specific topic to your home tab. With support for over 75 topics, you can dive deep into your favorite niche on X.



It's powered by Grok's… pic.twitter.com/9jkIEXvubj — Nikita Bier (@nikitabier) April 21, 2026

X’s custom timelines offer 75+ category options

The initial topics are broad and fairly standard — high-level categories similar to the type of sections you might find on news sites. These include subjects like Business & Finance, Sports, Technology, Politics, Stocks & Economy, News, Science, Movies & TV, Food & Drink, Art, Real Estate, Home & Garden, Beauty, Education, Gaming, and others.

Beyond the broader sports category, there are also options for following specific sports, including American football, baseball, basketball, boxing, soccer, golf, MMA & wrestling, racing & motorsports, rugby, snow sports, ice hockey, tennis, cricket, Formula 1, cycling, and the Olympics. (Oh, and e-sports, if you want to count that.)

Pop culture and tech topics also make up many of the available categories, with the former allowing you to pin topics like celebs, music, concerts, country music, dance, electronic music, fashion, pop, K-pop, J-pop, podcasts, hip hop, and jazz.

Alongside the Technology category, you can also follow special interests like Artificial Intelligence and Cryptocurrency — two perennially popular topics on X. There are also categories for things that overlap with Elon Musk’s various businesses and interests, like robotics, software development, space, and biotech.

Image Credits:screenshot from X

Other general categories include things like anime, digital art, photography, career, pets, design, marriage & family, shopping, mental health, and more.

News Categories: War, Crime, and Elections

Worth flagging: The initial set of news-related topics leads with the Iran Conflict, Crime, and Elections at the top of suggestions.

While this likely reflects the current conversations taking place on X, it’s also an example of how a product decision can influence what news people see. A cleaner solution might be to organize the dozens of options into larger high-level categories listed in alphabetical order, with the subcategories appearing when you tap each. That would allow X to greatly expand its “news” categories beyond these big three.

There could also be concern about these timelines being built by Grok, which was ostensibly created to be politically neutral and “truth-seeking,” but in practice has often skewed right or amplified misinformation.

In our own testing though, the custom timelines did not seem to lean obviously right or left. In a handful of test scrolls, the feeds drew from a range of outlets like ABC, CBS, CSPAN, AP, Reuters, AFP, Daily Beast, The Hill, Foreign Policy, Puck, The Atlantic, The Economist, Bloomberg, Al Jazeera, Forbes, and the BBC (not all of which I follow), alongside commentary from various pundits.

Whether these custom feeds will dramatically change how people use X remains to be seen.

For the most part, people tend to want to see the things they care about appear in their main algorithmic feed. But the custom feeds do allow for exploring new interests or dipping into topics only when they’re relevant — like pulling up a sports feed when the game is on. Combined with X’s new “Snooze Topics” option for the For You feed, you can more precisely tailor your X to your liking.

In Brief

Ryan Roslansky stepped down as LinkedIn’s CEO on Wednesday after six years running the world’s largest professional network. Dan Shapero, the company’s COO, takes over immediately.

Roslansky’s tenure deserves attention. He joined LinkedIn in 2009 — he was one of Jeff Weiner’s first hires — and spent more than a decade working through nearly every corner of the business before taking the top job in June 2020, at the height of pandemic-era labor market chaos. Microsoft acquired LinkedIn in 2016.

He inherited a platform with 700 million members and roughly $8 billion in annual revenue. He’s leaving with 1.3 billion members and more than $17 billion in revenue. A solid run by any measure.

That revenue growth may be inseparable from a transformation he accelerated — turning a glorified jobs board into something closer to a full-blown social network, where executives share personal essays, post career advice, and occasionally sob on camera. (You may have feelings about that; millions of people apparently do, and they’re not shy about sharing them.)

Apparently, you needn’t feel too bad for Roslansky, who also holds the title of EVP at Microsoft, and who suggested in a LinkedIn post that his role inside Microsoft is only expanding, with Shapero now reporting directly to him.

Apple released a software update on Wednesday for iPhones and iPads fixing a bug that allowed law enforcement to extract messages that had been deleted or disappeared automatically from messaging apps. This was because notifications that displayed the messages’ content were also cached on the device for up to a month.

In a security notice on its website, Apple said that the bug meant “notifications marked for deletion could be unexpectedly retained on the device.”

This is a clear reference to an issue revealed by 404 Media earlier this month. The independent news outlet reported that the FBI had been able to extract deleted Signal messages from someone’s iPhone using forensic tools, due to the fact that the content of the messages had been displayed in a notification and then stored inside a phone’s database — even after the messages were deleted inside Signal.

After the news, Signal president Meredith Whittaker said the messaging app maker asked Apple to address the issue. “Notifications for deleted messages shouldn’t remain in any OS notification database,” Whittaker wrote in a post on Bluesky.

Contact Us Do you have more information about how authorities are using forensic tools on iPhones or Android devices? From a non-work device, you can contact Lorenzo Franceschi-Bicchierai securely on Signal at +1 917 257 1382, or via Telegram and Keybase @lorenzofb, or Do you have more information about how authorities are using forensic tools on iPhones or Android devices? From a non-work device, you can contact Lorenzo Franceschi-Bicchierai securely on Signal at +1 917 257 1382, or via Telegram and Keybase @lorenzofb, or email .

It’s unclear why the notifications’ content was logged to begin with, but today’s fix suggests it was a bug.

Apple did not immediately respond to a request for comment asking why the notifications were being retained. The company also backported the fix to iPhone and iPad owners running the older iOS 18 software.

Privacy activists expressed alarm when they learned that the FBI had found a way around a security feature that is used daily by at-risk users. Signal, like other messaging apps such as WhatsApp, allows users to set up a timer that instructs the app to automatically delete messages after a set amount of time. This feature can be helpful for anyone who wants to keep their conversations secret in the event that authorities seize their devices.

Some of the most consequential companies in tech history didn’t launch with a splashy fundraising announcement. They started with a pitch. Dropbox demoed to a room of skeptics. Cloudflare took the stage before most people understood what edge networking meant. Discord was a scrappy game developer called Hammer & Chisel. Mint, Trello, Forethought, N26 — all of them passed through the same crucible: TechCrunch Startup Battlefield.

That’s not a coincidence. Battlefield isn’t just a competition. It’s a launchpad, and the numbers back it up. More than 1,700 companies have competed on the Battlefield stage. Together, they’ve raised $32 billion in total funding and generated over 250 exits — including acquisitions by Microsoft, Google, Yahoo, Salesforce, Twitter, Uber, and Amazon. The Startup Battlefield network runs so deep that alumni have even acquired each other: Dropbox acquired fellow Startup Battlefield alum DocSend in 2021. For thousands of founders, it’s become a defining milestone — not just a pitch competition, but the moment the world started paying attention.

We wanted to show you what happens after the confetti falls. We checked in with some of our recent alumni, many of whom have sat down with us on Build Mode: The Founder Survival Guide, TechCrunch’s podcast for founders at every stage. Here’s what they’ve been building, in their own words.

About Build Mode

Each season goes deep on a different chapter of startup life. Season 1 covered go-to-market. Season 2 — out now — is all about building your team. And mark your calendars: Season 3 drops in June, tackling the most requested topic we’ve ever gotten: fundraising.

Subscribe now so you don’t miss it.

The champions and runners-up

From military logistics to Startup Battlefield 2025 champion

Kevin Damoa, founder of Glīd — 2025 winner

Kevin Damoa didn’t come from Sand Hill Road. He came from military logistics — a background that turned out to be ideal training for building under pressure, with constrained resources and real stakes. Damoa’s path to the Startup Battlefield 2025 championship is the kind of origin story that makes you reconsider where the next generation of great founders is actually coming from.

Techcrunch event Meet your next investor or portfolio startup at Disrupt

Your next round. Your next hire. Your next breakout opportunity. Find it at TechCrunch Disrupt 2026, where 10,000+ founders, investors, and tech leaders gather for three days of 250+ tactical sessions, powerful introductions, and market-defining innovation. Register now to save up to $410. Meet your next investor or portfolio startup at Disrupt

Your next round. Your next hire. Your next breakout opportunity. Find it at TechCrunch Disrupt 2026, where 10,000+ founders, investors, and tech leaders gather for three days of 250+ tactical sessions, powerful introductions, and market-defining innovation. Register now to save up to $410. San Francisco, CA | REGISTER NOW

→ Listen to Kevin’s Build Mode episode

From the Startup Battlefield stage to the International Space Station

Capella Kerst, founder and CEO of geCKo Materials — 2024 runner-up

Capella Kerst didn’t set out to reinvent adhesion. She set out to solve a problem that has stumped engineers for decades: How do you make things stick — reliably, repeatedly, and without residue — in the most extreme environments imaginable? geCKo Materials, spun out of Stanford, has developed gecko-inspired adhesive technology with applications ranging from manufacturing floors to, quite literally, the International Space Station.

Kerst’s Startup Battlefield moment was a signal to the market that the science was ready for the world. What’s happened since is proof that runner-up isn’t a consolation prize — it’s a credential. Hear how she got there:

→ Listen to Capella’s Build Mode episode

How Forethought AI found product-market fit — before it was obvious

Deon Nicholas, co-founder of Forethought AI — 2018 winner (acquired by Zendesk)

Few Startup Battlefield stories have a more complete arc than Forethought AI. Deon Nicholas took the stage with a conviction that AI could fundamentally transform customer support — before that was an obvious bet. Before the term sheets and the headlines, there was a pitch and a thesis. Forethought was recently acquired by Zendesk — the latest example of what the Startup Battlefield stage can set in motion. His Build Mode episode is essential listening, and a perfect primer for Season 3’s deep dive on fundraising.

→ Listen to Deon’s Build Mode episode

Top 20 finalist stories

The danger of fundraising before finding product-market fit

David Park, founder of Narada

Raising before product-market fit doesn’t speed things up — it speeds up your mistakes. Park doesn’t sugarcoat the lessons.

→ Listen to David’s Build Mode episode

Using AI to hire for compatibility, not just skill

Sarah Lucena, founder and CEO of Mappa

Skills get people in the door. Compatibility determines whether they stay. Lucena is using AI to fix the part of hiring nobody talks about.

→ Listen to Sarah’s Build Mode episode

More from the Startup Battlefield alumni community

These founders competed on the Startup Battlefield and sat down with us on Build Mode to tell their story. All worth a listen.

Anna Sun of Nowadays and Hala Jalwan and Alessio Tresanti of Rivio — On what happens when a startup becomes a family business, and the community that forms around Startup Battlefield. → Listen

Kyle Rudolph and Jon Walburg, co-founders of Alltroo — On why your network is your first go-to-market strategy. → Listen

Jas Schembri-Stothart of Luna and Andre Peart of Untapped Solutions — On reaching the markets everyone else ignores and building for underserved communities without the typical growth playbook. → Listen

The milestone is real

Every generation of Startup Battlefield alumni adds a new chapter to the same story. But behind every one of those data points is a founder who made a bet on themselves — publicly, in front of people who were paying attention. The stage matters. The community lasts. The milestone is real.

Applications for Startup Battlefield 2026 are open. If you’re building something that deserves a stage, this is yours.

→ Apply now

Know a founder who’s ready for the spotlight? Investors, operators, and fellow founders can nominate companies directly.

→ Nominate a founder

Not ready to apply yet? Build Mode is where we meet you. Season 2 is live now. Season 3 — all about fundraising — drops this summer.

→ Subscribe to Build Mode

Google Cloud on Wednesday announced that its eighth generation of custom-built AI chips, or tensor processing units (TPUs), will be split in two. One chip, named the TPU 8t, will be geared for model training and another, the TPU 8i, is aimed at inference.

Inference is the ongoing usage of models, aka what happens after users submit prompts.

As you might expect, the company touts some impressive performance specs for these new TPUs compared to the previous generations: up to 3x faster AI model training, 80% better performance per dollar, and the ability to get 1 million+ TPUs to work together in a single cluster. The upshot should be a lot more compute for a lot less energy — and cost to customers — than previous versions. It calls these chips TPUs, not GPUs, because its custom low-power chips were originally named Tensor.

But Google’s chips are not a full frontal assault on Nvidia’s future, at least not yet. Like the other giant cloud providers, including Microsoft and Amazon, Google is using these chips to supplement the Nvidia-based systems it offers in its infrastructure. It is not flat-out replacing Nvidia. In fact, Google promises its cloud will have Nvidia’s latest chip, Vera Rubin, available later this year.

One day the hyperscalers building their own AI chips (which includes Amazon, Microsoft, and Google) may grow to need Nvidia less, as enterprises move their AI needs to their clouds and port their apps to these chips.

Still, as things stand today, it’s not profitable to bet against Nvidia. As notable chip market analyst Patrick Moorhead jokingly posted on X, he had predicted that Google’s TPU could be bad news for Nvidia (and Intel) back in 2016 when the search giant launched its first one. Nvidia is now a nearly $5 trillion market cap company, meaning that prediction didn’t exactly hold up to the test of time.

If all goes according to Nvidia’s plan, Google’s growth as an AI cloud provider would result in more business for the chip maker not less, even if many a workload runs on Google’s chips.

Techcrunch event Meet your next investor or portfolio startup at Disrupt

Your next round. Your next hire. Your next breakout opportunity. Find it at TechCrunch Disrupt 2026, where 10,000+ founders, investors, and tech leaders gather for three days of 250+ tactical sessions, powerful introductions, and market-defining innovation. Register now to save up to $410. Meet your next investor or portfolio startup at Disrupt

Your next round. Your next hire. Your next breakout opportunity. Find it at TechCrunch Disrupt 2026, where 10,000+ founders, investors, and tech leaders gather for three days of 250+ tactical sessions, powerful introductions, and market-defining innovation. Register now to save up to $410. San Francisco, CA | REGISTER NOW

In fact, Google also says it has agreed to work with Nvidia to engineer computer networking that allows Nvidia-based systems to perform even more efficiently in its cloud. In particular, the two tech giants are working to beef up the software-based networking tech called Falcon, which Google created and open sourced in 2023 under the godfather of all open source data center hardware organizations, the Open Compute Project.

As part of its slate of Google Cloud Next announcements on Wednesday, the company shared plans to bring “auto browse” agentic capabilities to Chrome users in the enterprise, along with enhanced security measures.

With auto browse, Chrome users can take advantage of Gemini to understand the live context in their open browser tabs, and then use the AI to handle various tasks like booking travel, inputting data, scheduling meetings, and others related to web-based work.

Image Credits:Google

Google suggests the tool could be used for things like inputting information in the company’s preferred CRM system based on content in a Google Doc, comparing vendor pricing across tabs, summarizing a candidate’s portfolio before an interview, pulling key data from a competitor’s product page, and more.

The company notes that its workflows will still require a “human in the loop,” meaning that the user will have to manually review and confirm the AI’s input before any final action takes place.

However, the idea is to help speed up these types of more tedious tasks to free up people to focus on what Google refers to as more “strategic work.”

Image Credits:Google

This is the larger promise from AI advocates: that you’ll get your time back by using this new technology. But in practice, studies have shown that AI isn’t reducing work — it’s intensifying it. It remains to be seen how this will play out at the enterprise level as AI becomes a standard part of the workflow. Presumably, that could mean managers will expect that people can get more tasks done in less time.

Google says the new feature will initially be available to Workspace users in the U.S., as a part of Google’s push to infuse its AI into one of its most-used apps in the workplace, the web browser nearly everyone uses. It can be enabled via a policy, and Google states that an organization’s prompts won’t be used to train its AI models. (A disclosure that is increasingly necessary these days, given that Meta is even using its own employees’ keystrokes to train its AI.)

Like the consumer-facing version of the feature, Workspace users will be able to save their most common workflows for later use. These “Skills,” as they’re called, can be pulled up by either typing a forward slash (” / “) or by clicking the plus sign to access the needed Skill.

In addition to the infusion of AI into Chrome, Google is touting its ability to detect unsanctioned AI tools in the workplace via Chrome Enterprise Premium. Now, it’s expanding those capabilities to help IT teams look for compromised browser extensions or other AI services — specifically “anomalous agent activity.”

Google is correct to position this as a security feature, but it has another advantage, too. The tech giant is essentially leveraging corporate IT to shut down any other AI agents that could be taking root in the enterprise world organically. Years ago, this was how many web services established themselves in the workplace, amid an employee-driven “Enterprise 2.0” rush to adopt new technology like cloud storage, collaborative docs, or file sharing.

This new feature, which Google somewhat ominously dubs “Shadow IT risk detection,” will give IT teams visibility into the usage of both sanctioned and unsanctioned GenAI and SaaS sites across their organization.

Image Credits:Google

IT teams will also receive a “Gemini Summary” of the Chrome Enterprise release notes and other AI-powered suggestions. This will surface critical changes, new policies, and upcoming deprecations, along with recommendations about things like configuring new settings or reviewing managed browsers.

The company also announced an expanded partnership with Okta to secure the agentic workplace with added features to reduce session hijacking and other protections. It’s also upgrading its security controls for extensions and introducing Microsoft Information Protection (MIP) integration to help organizations enforce consistent security policies.

Threads is launching “Live Chats” to allow for real-time conversations during cultural events, the Meta-owned platform announced on Wednesday. The company says the idea behind the new feature is to help make the app feel more timely and relevant.

Live Chats are launching first within the NBA Threads community during the Playoffs. Media personalities, including Malika Andrews, Rachel Nichols, and Da Kid Gowie will host Live Chats during games.

The company told TechCrunch via email that it’s initially rolling out the feature to a small group of creators. Not all users will be able to start a Live Chat right away, but Meta plans to expand access over time.

Users can send messages, photos, videos, links, and emoji reactions. Up to 150 participants can actively send messages in a chat. Once this limit is reached, additional users can still view the conversation, react to messages, and participate in polls in “spectator” mode.

When Threads first launched, it lagged behind X when it came to relevance and timeliness; it was harder to follow what was currently happening in the world. X had already established itself as a global “town square” of sorts, and was known for real-time commentary and breaking news. Threads, on the other hand, was struggling because it lacked many of the features that X had, such as robust search, hashtags, and a chronological feed.

Over time, Threads has built out its platform with these tools, and now it’s looking to better compete with X through the addition of a feature that even the Elon Musk-owned app doesn’t have, one specifically designed for real-time engagement.

Image Credits:Meta /

“It’s a new way to build community with others around shared interests like an album drop or a big game as it unfolds,” Meta explained in the blog post. “Live Chats are an extension of what’s already happening on Threads — and a new way for creators and fans to connect over what matters to them in real time. Live Chats are more dynamic than traditional group chats because they’re designed for real-time conversations around cultural moments as they’re happening.”

Techcrunch event Meet your next investor or portfolio startup at Disrupt

Your next round. Your next hire. Your next breakout opportunity. Find it at TechCrunch Disrupt 2026, where 10,000+ founders, investors, and tech leaders gather for three days of 250+ tactical sessions, powerful introductions, and market-defining innovation. Register now to save up to $410. Meet your next investor or portfolio startup at Disrupt

Your next round. Your next hire. Your next breakout opportunity. Find it at TechCrunch Disrupt 2026, where 10,000+ founders, investors, and tech leaders gather for three days of 250+ tactical sessions, powerful introductions, and market-defining innovation. Register now to save up to $410. San Francisco, CA | REGISTER NOW

Users can join Live Chats from the top of a Community feed, through a shared post in their main feed, or by tapping the red live ring around a host’s profile photo. Although Live Chats end after a certain time, they remain open and publicly discoverable after they end, Meta says.

Threads will automatically detect and take down messages that violate its policies, and anyone in a chat can report messages, the company told TechCrunch. Additionally, hosts have real-time moderation controls and can demote users to spectator mode or remove them from the chat.

As for the future, Meta says it is going to update Live Chats with several new features, including co-hosting, real-time play-by-play updates, lock-screen widgets that highlight live chat activity, and the ability to quote and share chat messages directly to Threads feeds.

While Meta noted that the feature could also be used for things like album drops, there’s potential for Live Chats to be used in numerous other settings — such as the upcoming FIFA World Cup games, awards shows, TV show finales, and more.

In Brief

Google CEO Sundar Pichai opened the Google Cloud Next conference on Wednesday with a video in which he announced one of the company’s biggest new products: Gemini Enterprise Agent Platform.

Google’s tool is intended for building and managing agents at scale. This is Google’s answer to Amazon’s Bedrock AgentCore and to Microsoft Foundry.

Given that AI, and agents in particular, are furthest along for technical tasks like coding, and that the tech is so new to the enterprise that security remains a real concern, Google has made an interesting choice with this tool. Agent Platform is particularly geared at IT and technical teams.

The business folks, meanwhile, are directed toward what Google calls its Gemini Enterprise app, introduced in the fall. They can work with agents built by IT or build their own for tasks like scheduling meetings, performing trigger-based processes, creating shortcuts for repetitive tasks, or creating and editing files without needing to switch apps, Google says.

Google also underscored that the underlying models these tools tap into include Google’s own Gemini LLM and Nano Banana 2 image generator, as well as Anthropic’s Claude. The company announced support for Claude Opus, Sonnet, and Haiku — in other words, flagship, reasoning, and lower-cost models, including the new Opus 4.7 that launched last week.

During its Google Cloud Next conference on Wednesday, the company announced a slew of Workspace-focused updates, including the addition of its AI Overviews feature to Gmail. The feature, which today uses AI to summarize Google Search results, will now do the same for Gmail users in the workplace.

According to Google, this will allow Gmail users to ask questions in search using natural language and then get concise answers without having to open and read different emails.

The company suggests the feature could be used to ask business-related questions about topics typically shared in emails, like those about performance improvements, project milestones, invoices, comments on decks, trip details, and more with straightforward answers.

The AI Overview will create an instant summary pulled from across multiple emails and conversations.

Image Credits:Google

While not everyone prefers to have AI as their first step to finding an answer, it is rapidly becoming the norm, both within Google’s products and elsewhere on the web.

In this case, Google says the AI Overviews in Gmail will be the default setting if the company has Gemini for Workspace in Gmail enabled, and if Workspace Intelligence access to Gmail is enabled. (End users must have “Smart features in Gmail, Chat, and Meet” and “Google Workspace smart features” enabled, too.)

The feature was previously available to consumers with Google AI Pro and Ultra subscriptions. Google says it will also now come to business, enterprise, and education customers as well through the following products:

Techcrunch event Meet your next investor or portfolio startup at Disrupt

Your next round. Your next hire. Your next breakout opportunity. Find it at TechCrunch Disrupt 2026, where 10,000+ founders, investors, and tech leaders gather for three days of 250+ tactical sessions, powerful introductions, and market-defining innovation. Register now to save up to $410. Meet your next investor or portfolio startup at Disrupt

Your next round. Your next hire. Your next breakout opportunity. Find it at TechCrunch Disrupt 2026, where 10,000+ founders, investors, and tech leaders gather for three days of 250+ tactical sessions, powerful introductions, and market-defining innovation. Register now to save up to $410. San Francisco, CA | REGISTER NOW

Business: Business Starter, Standard, and Plus

Enterprise: Enterprise Starter, Standard, and Plus

Consumers: Google AI Pro and Ultra

Other Editions: Frontline Plus

AI Add-ons: Google AI Pro for Education

Alongside the launch, Google said it’s also making AI Overviews in Drive broadly available to eligible Workspace and Google AI plans. It was previously in beta.

Rivian has rolled the first customer-ready R2 SUVs off the production line at its factory in Normal, Illinois, just days after it was hit by an EF-1 tornado that tore off part of the roof.

Despite the damage, founder and CEO RJ Scaringe told Bloomberg Television on Wednesday morning that Rivian doesn’t expect any delays to the R2’s rollout, which is crucial to the company’s survival.

“The tornado went through the south end of the plant, and ripped the roof off the building, and knocked down some of the plant as well, and so the last 72 hours have been around the clock,” he said. Scaringe explained that Rivian has had to change how and where it brings some materials into the factory to build the R2.

But “we’re not making any changes to the plan,” he said, referring to the company’s production roadmap.

Scaringe wasn’t asked when Rivian will make the first R2 deliveries during the interview. The company has previously said it will start shipping R2 SUVs before the first half of 2026 comes to an end.

Getting the R2 into production is a major milestone for the company. It’s the first production vehicle Rivian has made that has a chance to reach mass-market customers, as it costs far less than the company’s current R1 EVs. It’s also supposed to help the company finally reach profitability after years of losing money on every vehicle it sold.

The company has big expectations for the R2. Rivian told investors earlier this year that it expects to deliver between 20,000 and 25,000 of the SUVs by the end of 2026. If Rivian achieves that, it would become one of the fastest-scaling new EVs ever launched in the U.S., second only to Tesla’s Model Y.

Techcrunch event Meet your next investor or portfolio startup at Disrupt

Your next round. Your next hire. Your next breakout opportunity. Find it at TechCrunch Disrupt 2026, where 10,000+ founders, investors, and tech leaders gather for three days of 250+ tactical sessions, powerful introductions, and market-defining innovation. Register now to save up to $410. Meet your next investor or portfolio startup at Disrupt

Your next round. Your next hire. Your next breakout opportunity. Find it at TechCrunch Disrupt 2026, where 10,000+ founders, investors, and tech leaders gather for three days of 250+ tactical sessions, powerful introductions, and market-defining innovation. Register now to save up to $410. San Francisco, CA | REGISTER NOW

That said, Rivian is launching with a version of the R2 that costs nearly $13,000 more than the $45,000 price tag the company spent years promoting. The launch edition R2 starts at $57,990, with a slightly cheaper $53,990 variant coming by the end of this year. Rivian won’t sell an R2 for under $50,000 until the first half of 2027, and a true base model starting at $45,000 won’t hit the market until late 2027.

And that’s if the $45,000 R2 ever arrives at all. When Rivian announced pricing for the SUV in March, the company said the base model price will start “around $45,000” — not “at $45,000” as it had promoted on its website as recently as February.

Netherlands-based cosmetics giant Rituals has confirmed a data breach affecting customers’ personal information after hackers stole reams of data from its membership database.

The company disclosed the breach on Wednesday, according to an email sent to customers that TechCrunch has viewed and verified.

Rituals said it identified an “unauthorized download” of members’ data in April that contained customers’ full name, date of birth, gender, postal and email address, and phone number, as well as their preferred Rituals store and account type.

When reached by TechCrunch, Rituals spokesperson Eline van Malssen said the hacker stole membership data about customers in Europe and the United Kingdom.

TechCrunch has learned that some customers notified by Rituals are based in the United States. The spokesperson confirmed the incident also affects some U.S. customers.

Rituals did not describe the nature of the cyberattack and the company said its investigation was underway to understand how the data breach happened.

The cosmetics giant is the latest retailer to have customer membership data stolen in the past year, following a string of intrusions at U.K. grocery and shopping chain Co-op and Marks & Spencer, among others. Customer records can be attractive targets for hackers who steal the data and extort the company for a ransom in exchange for not publishing the information online.

When reached with questions about the incident, a Rituals spokesperson declined to comment on whether the company received any communication from the hackers, to share a more precise timeline of the breach, or to provide the exact number of affected members, citing unspecified “security reasons.”

According to its website, Rituals has over 41 million customers in its membership database. The retail giant made €2.4 billion euros ($2.8 billion) in revenue in 2025.

ARK Invest Venture Fund has made its first-ever lead investment in an early-stage startup called Lucra, firm founder Cathie Wood told TechCrunch.

“We feel pretty excited about it,” Wood (pictured above) said in the recent interview regarding the investment in the startup.

Lucra developed a software platform that reimagines corporate loyalty programs into interactive, esports-like events such as tournaments where customers can play each other, even betting or winning cash or company giveaways. The startup said its customers include Five Iron Golf, Chess Kings, and Dave & Buster’s.

Lucra announced on Wednesday that it raised a $20 million Series B, led by the ARK fund, with participation from Alumni Ventures, Astralis Capital, Harlo Equity Partners, Simplex Ventures, SeventySix Capital, and WTI.

There are a few reasons why the famed financial company has never led a startup deal before. For one, the ARK Invest Venture Fund is not a typical VC fund. It’s an SEC-regulated interval fund (also known as a closed-end mutual fund), meaning anyone can invest in it, for as little as $500. However, it is not traded on a public exchange, so investors cannot sell shares at will. They can sell limited shares on specific dates, quarterly.

Wood also noted that the person running the fund, director of research Nick Grous, “is a tough sell,” leaving startups with the difficult task of getting him excited enough to advocate to lead a deal.

What’s even wilder is that ARK was particularly gun-shy about this sort of business because it got burned after investing in a somewhat similar company a few years ago.

Techcrunch event Meet your next investor or portfolio startup at Disrupt

Your next round. Your next hire. Your next breakout opportunity. Find it at TechCrunch Disrupt 2026, where 10,000+ founders, investors, and tech leaders gather for three days of 250+ tactical sessions, powerful introductions, and market-defining innovation. Register now to save up to $410. Meet your next investor or portfolio startup at Disrupt

Your next round. Your next hire. Your next breakout opportunity. Find it at TechCrunch Disrupt 2026, where 10,000+ founders, investors, and tech leaders gather for three days of 250+ tactical sessions, powerful introductions, and market-defining innovation. Register now to save up to $410. San Francisco, CA | REGISTER NOW

“We had actually owned a company called Skillz, which kind of operated in this space,” Grous said. “It didn’t work out well for us and many other investors.”

Skillz was a once-hot public company that later became mired in troubles and lawsuits. The big difference, the investor said, is that Lucra is a B2B platform, selling interactive esports as a loyalty program, rather than trying to license and run games directly to consumers.

“Overcoming our initial hurdle, especially given our experience with Skillz, overcoming our reticence, having Nick overcome it, that was our first screen,” Wood said of how this startup convinced her company to write a big check.

In this case, ARK Invest had participated in Lucra’s previous Series A round, and had grown familiar with its business model, its trajectory, and its founder and CEO Dylan Robbins, Grous told TechCrunch.

“We had been in constant communication,” Grous said, adding that his venture-esq fund attempts to have quarterly conference calls with the startups in the portfolio, similar to how public companies report to investors quarterly. ARK mostly works in the public market, offering a slate of publicly traded EFT funds.

Nick Grous Image Credits:ARK Invest

Despite already being in the portfolio, Lucra’s founder was grilled numerous times when it came time to buy more shares — first by Grous and then ARK’s investment committee, both he and Wood described.

During those calls, Robbins “had thought about all the things that went wrong” with similar companies like Skillz, as well as with Lucra, and had answers, Wood said. “No matter how many times we went at him, his conviction, there was just no let up,” she described.

It also helped that this company’s financials were promising, it was in an area that ARK knew well, and this was not AI, aka the most hyped, most expensive area these days.

“We’ve been underwriting the sports-betting space, understanding the gamification aspects of entertainment,” Grous said, meaning that the investment firm could “really understand the opportunity here.”

The ARK Invest Venture Fund holds shares of companies like Epic Games, Kalshi, and Discord, for instance. It also holds OpenAI, Anthropic, Replit, Grok, and Perplexity, so it knows the AI scene well.

“We are all over AI, just like everyone else, because it is a massive revolution,” Wood explained. “But in the process, a lot of companies are being neglected.” This means that spotting such potentially neglected companies is “our opportunity because we are doing research in many other areas than AI,” she said.

OpenAI has partnered with Infosys to integrate its artificial intelligence tools, including coding assistant Codex, into the Indian IT giant’s Topaz AI platform.

Infosys said the integration will be used to help its clients modernize software development, automate workflows and deploy AI systems at scale, initially focusing software engineering, legacy modernization, and DevOps.

India’s IT services firms face mounting pressure from a mix of slowing client spending and rapid advances in generative AI. Shares of Infosys have fallen over 22% this year amid a broader sell-off triggered by weak forecasts, investor concerns that AI tools could automate parts of traditional outsourcing work, and macroeconomic turmoil due to the U.S.-Iran war.

The move also reflects a broader trend of AI firms teaming up with global IT services providers to scale adoption in large enterprises. OpenAI has previously partnered with HCLTech, and Infosys has struck a similar deal with Anthropic.

OpenAI gains a distribution channel into large enterprises through Infosys’ global client base and delivery capabilities across more than 60 countries. The companies said the deal is aimed at helping enterprises move from experimentation to large-scale deployment.

Infosys has been ramping up its AI business. The company said earlier this year that AI-related services generated ₹25 billion (about $267 million) in revenue in the December quarter, or roughly 5.5% of its total.

The deal is part of a broader push by OpenAI to expand its enterprise footprint through initiatives such as Codex Labs, announced on Tuesday, which involves engineers working with clients to help deploy its tools. Initial partners include Accenture, Capgemini, CGI, Cognizant, Infosys, PwC and Tata Consultancy Services, as OpenAI aims to build a distribution network to scale adoption of Codex, which now has more than 4 million weekly active users.

Techcrunch event Meet your next investor or portfolio startup at Disrupt

Your next round. Your next hire. Your next breakout opportunity. Find it at TechCrunch Disrupt 2026, where 10,000+ founders, investors, and tech leaders gather for three days of 250+ tactical sessions, powerful introductions, and market-defining innovation. Register now to save up to $410. Meet your next investor or portfolio startup at Disrupt

Your next round. Your next hire. Your next breakout opportunity. Find it at TechCrunch Disrupt 2026, where 10,000+ founders, investors, and tech leaders gather for three days of 250+ tactical sessions, powerful introductions, and market-defining innovation. Register now to save up to $410. San Francisco, CA | REGISTER NOW

The companies did not disclose financial details of the deal.