Here are all the tech companies rolling back DEI or still committed to it — so far
Companies around America have started cutting DEI programs and eliminating DEI commitments from public documents in response to legal and political threats from the Trump administration.
Just a few weeks ago, Attorney General Pam Bondi instructed the Department of Justice to “investigate, eliminate, and penalize,” DEI programs deemed illegal in private sector companies that receive federal funds.
Trump has signed an executive order banning DEI in the federal government with the approval of his right-hand companion Elon Musk, an immigrant from South Africa and a long-time denouncer of DEI.
The reaction to the government’s anti-diversity push has taken many forms. One contractor said NASA has started to scrub mention of terms like “environmental justice” or “anything specifically targeting women.” PBS had to shutter its DEI office.
In corporate America, McDonald’s, Harley-Davidson, Booz Allen, John Deere, Tractor Supply Company, Polaris, Lowe’s, Ford, Molson Coors, Walmart, Nissan, Accenture, and Target, among others, have all announced some form of rollback to their DEI policies. Goldman Sachs said in February that it would ditch its requirement for corporate boards filing to go public to include women and people of color.
Other companies are trying to strike a balance. Some companies like Apple, Costco, Delta, McKinsey, and JPMorgan have indicated a commitment to sticking with their diversity efforts.
Last year, companies like Telsa, GameStop, Workday, and Salesforce started scrubbing mentions of DEI from their 10-K filings, a trend that has accelerated so far this year. Here’s a list of where DEI stands, so far, among tech companies in 2025.
We’ve reached out to each of the companies for comment.
Amazon
Amazon’s website “positions” page still showcases its diversity, equity, and inclusion commitment. However, it is winding down some DEI “outdated programs and materials,” according to a leaked memo reported by Bloomberg. It removed a post from 2021 that expressed a focus on DEI and removed some mentions of diversity from its latest annual 10-K report.
In a statement to TechCrunch, an Amazon spokesperson pointed to the “positions” page saying “as stated here, we’re committed to creating a diverse and inclusive company.”
Apple
Apple’s website still says, as it has for years, “We’re continuing to create a culture of inclusion, increasing representation across teams, and holding ourselves accountable at every level.”
Shareholders at Apple’s latest meeting rejected a proposal from a conservative group to strike DEI policies from the companies. Apple previously advised its shareholders to vote against the proposal.
Google is eliminating diversity hiring targets and evaluating releasing new diversity reports, which it has done since 2014, the Wall Street Journal reported. It also said it would review its DEI initiatives to ensure all comply with the latest executive orders aimed at curbing DEI.
Google also removed underrepresented language from its grant website and scrubbed mention of diversity and equity from its responsible AI team webpage. Furthermore, Google updated its 10-K filing, making no mention of diversity in its latest one, compared to the eight times it was mentioned in its 2023 filing. Late last year, it removed many cultural events from the default Google Calendar.
IBM
IBM still has its inclusion page up. Working with a shareholder, the conservative think tank Heritage Foundation requested the company drop its DEI pay incentive targets, though IBM has moved to block the proposal.
Intel
Intel cut back some of the language about DEI on its recent 10-K filing and removed some diversity targets. On its website, though, the company still states that “diversity and inclusion are among the most important forces driving the company’s evolution and reinvention.”
Medium
The publishing platform Medium said in a blog post that it would remain committed to DEI. The company’s CEO said most people value understanding and respect over hate and division.”
Meta is eliminating some DEI programs targeting hiring, suppliers, and training according to a leaked internal memo reported by Axios. Meta also said it eliminated its DEI team and ended representation goals to eliminate any impression “that decisions are being made based on race or gender,” the memo read. “While this has never been our practice, we want to eliminate any impression of it.” Meta declined to comment.
Microsoft
Microsoft released its 2024 inclusion report in October where CEO Satya Nadella reiterated the importance of a diverse workforce. However, in July of that year Microsoft laid off its internal DEI team because of “changing business needs,” according to a memo leaked to Business Insider.
Nvidia
Nvidia still has its “diversity, inclusion, and belonging” page up and released its 2024 Sustainability Report where it has a section still dedicated to “People, Diversity and Inclusion.”
OpenAI
OpenAI quietly changed its “Commitment to Diversity” website to now read “Building Dynamic Teams,” and removed all mention of diversity and inclusion from the web page.
Oracle
Oracle still has its “Culture and Inclusion” page up, which reads that “diverse perspectives make our teams stronger and empower collaboration.
Salesforce
Salesforce is getting rid of its diversity hiring goals. It also removed mention of “diversity” from its 10-K filing.
Tesla
Tesla scrubbed mention of diversity, equity, and inclusion from its 10-K last year. Elon Musk, the CEO of the company, is an outspoken critic of DEI. The company only released one diversity, equity, and inclusion report, back in 2020.
Workday
Although Workday erased mention of diversity targets in its 2023 10-K form after expressing a commitment to increase the representation of Blacks and Latinos at the company, its DEI page is still up on the website. On that page, the company wrote that “diversity isn’t just a business imperative. It’s core to everything we do.” It released a 2024 Global Impact Report and it still has a chief diversity officer.
Uber
Uber eliminated an incentive tied to determining executive compensation pay by measuring the diversity progress of the company; it got rid of the promise to consider women and people of color candidates when looking to fill empty board seats or positions; and it removed a section in its proxy statement that highlights diversity as a value. It removed the entire diversity and inclusion section from its latest 10-K filing and made no use of the word diversity at all, according to the SF Examiner.
UnitedHealth Group
UnitedHealth Group removed mention of DEI from much of its website. For example, several pages once dedicated to DEI no longer load, and the company also removed a 2022 blog post with a vice president about DEI.
Yahoo
Yahoo removed several pages and sections that relate to or mention diversity, equity, and inclusion. For example, its DEI page now redirects to the company’s executive leadership page, and the 2022 diversity report no longer loads. (Disclosure: Yahoo is the parent company of TechCrunch. On March 21, Regent announced it would acquire TechCrunch.)
Zoom
Zoom laid off its DEI team during a round of layoffs last year, telling staff it would work with external consultants who will engage “all of our employees,” while focusing on “inclusion,” according to a memo leaked to Bloomberg. It doesn’t appear to have released a diversity report since 2022, the first of which came out in 2020, according to its website.
Keep reading the article on Tech Crunch
OpenAI’s Stargate project sets its sights on international expansion
Stargate, a $500 billion project headed up by OpenAI, Oracle, and SoftBank to build AI data centers and other AI infrastructure in the U.S., is considering investments in the U.K. and elsewhere overseas, according to a Financial Times report.
While Stargate was initially launched as a way to boost U.S. AI infrastructure, the project is allegedly weighing international expansion. In addition to the U.K., Germany and France are on the table, per the Financial Times’ reporting.
Stargate remains focused on the U.S. at the moment, to be clear, as originally pitched — and it’s still in the process of raising its first $100 billion. SoftBank is expected to put forward tens of billions of dollars as a mixture of debt and equity.
When Stargate was announced in January, President Donald Trump praised the initiative as a “declaration of confidence in America.”
Keep reading the article on Tech Crunch
The latest viral ChatGPT trend is doing ‘reverse location search’ from photos
There’s a somewhat concerning new trend going viral: People are using ChatGPT to figure out the location shown in pictures.
This week, OpenAI released its newest AI models, o3 and o4-mini, both of which can uniquely “reason” through uploaded images. In practice, the models can crop, rotate, and zoom in on photos — even blurry and distorted ones — to thoroughly analyze them.
These image-analyzing capabilities, paired with the models’ ability to search the web, make for a potent location-finding tool. Users on X quickly discovered that o3, in particular, is quite good at deducing cities, landmarks, and even restaurants and bars from subtle visual clues.
Wow, nailed it and not even a tree in sight. pic.twitter.com/bVcoe1fQ0Z
— swax (@swax) April 17, 2025
In many cases, the models don’t appear to be drawing on “memories” of past ChatGPT conversations, or EXIF data, which is the metadata attached to photos that reveal details such as where the photo was taken.
X is filled with examples of users giving ChatGPT restaurant menus, neighborhood snaps, facades, and self-portraits, and instructing o3 to imagine it’s playing “GeoGuessr,” an online game that challenges players to guess locations from Google Street View images.
this is a fun ChatGPT o3 feature. geoguessr! pic.twitter.com/HrcMIxS8yD
— Jason Barnes (@vyrotek) April 17, 2025
It’s an obvious potential privacy issue. There’s nothing preventing a bad actor from screenshotting, say, a person’s Instagram Story and using ChatGPT to try to doxx them.
o3 is insane
I asked a friend of mine to give me a random photo
They gave me a random photo they took in a library
o3 knows it in 20 seconds and it’s right pic.twitter.com/0K8dXiFKOY— Yumi (@izyuuumi) April 17, 2025
Of course, this could be done even before the launch of o3 and o4-mini. TechCrunch ran a number of photos through o3 and an older model without image-reasoning capabilities, GPT-4o, to compare the models’ location-guessing skills. Surprisingly, GPT-4o arrived at the same, correct answer as o3 more often than not — and took less time.
There was at least one instance during our brief testing when o3 found a place GPT-4o couldn’t. Given a picture of a purple, mounted rhino head in a dimly-lit bar, o3 correctly answered that it was from a Williamsburg speakeasy — not, as GPT-4o guessed, a U.K. pub.
That’s not to suggest o3 is flawless in this regard. Several of our tests failed — o3 got stuck in a loop, unable to arrive at an answer it was reasonably confident about, or volunteered a wrong location. Users on X noted, too, that o3 can be pretty far off in its location deductions.
But the trend illustrates some of the emerging risks presented by more capable, so-called reasoning AI models. There appear to be few safeguards in place to prevent this sort of “reverse location lookup” in ChatGPT, and OpenAI, the company behind ChatGPT, doesn’t address the issue in its safety report for o3 and o4-mini.
We’ve reached out to OpenAI for comment. We’ll update our piece if they respond.
Keep reading the article on Tech Crunch
Judge rules Google illegally monopolized adtech, opening door to potential breakup
A federal judge has found that Google violated antitrust laws by “willfully acquiring and maintaining monopoly power” in the advertising technology market, rounding out a two-year saga after the U.S. and eight states filed its initial complaints against the Alphabet-owned company.
The court will set a briefing schedule and hearing date to determine appropriate remedies for the antitrust violations, per a Thursday filing.
The remedies could include forcing Google to break up its advertising business, like selling its Google Ad Manager, which includes the AdX ad exchange and DFP (DoubleClick for Publishers), the ad server used for publishers.
Or the courts could force behavioral remedies that would allow Google to keep its business intact, but would impose restrictions to ensure fair competition, like prohibiting Google from prioritizing its own exchange or demand in auctions.
In a separate antitrust case, another federal judge last year found that Google illegally monopolized the general internet search market. The judge has not yet issued remedies on that case, but is expected to do so in mid-2025.
In the adtech case, Judge Leonie M. Brinkema wrote in her memorandum opinion that the plaintiffs failed to prove that the “open-web display advertiser ad networks” are a relevant market where Google has monopoly power. These networks help advertisers buy display ads across the open web, so they’re outside of closed ecosystems like Facebook, Instagram, and Google Search.
The judge did agree that Google violated the Sherman Act by monopolizing and unlawfully tying two parts of the adtech stack together, specifically DFP and AdX, and that Google is guilty of abuse of monopoly power in the publisher-side adtech.
“We won half of this case and we will appeal the other half,” Lee-Anne Mulholland, Google’s vice president of regulatory affairs, said in an emailed statement. “The Court found that our advertiser tools and our acquisitions, such as DoubleClick, don’t harm competition. We disagree with the Court’s decision regarding our publisher tools. Publishers have many options and they choose Google because our ad tech tools are simple, affordable and effective.”
The Department of Justice – alongside California, Colorado, Connecticut, New Jersey, New York, Rhode Island, Tennessee, and Virginia – filed suit against Google’s alleged use of monopoly power over the digital ad market in January 2023.
The DOJ argued that Google achieved its monopoly through anti-competitive conduct when it purchased DoubleClick in 2008, which then became the backbone of its ad business. Google then bought AdMeld in 2011 to gain more control over the ad market’s supply side. The government says this allowed Google to hike up ad prices and harm publishers by taking larger cuts of each sale.
The trial for this case began in September 2024 and lasted for three weeks, with closing arguments presented in late November.
This article has been updated to include more background about the case.
Keep reading the article on Tech Crunch
Make a brand impact by hosting a Side Event during TechCrunch All Stage Week in Boston
Want to tap into the energy of 1,000+ startup founders, investors, and tech leaders descending on Boston for TechCrunch All Stage on July 15? Host your own Side Event during “TC All Stage Week,” happening July 13-19!
Whether it’s a networking mixer, workshop, morning run, fireside chat, or cocktail hour — you call the shots. Create a moment. Build your community. Increase brand visibility with a Side Event.
Why host a Side Event?
Along with extending the energy beyond the venue and the actual event, you’ll get:
- Visibility to the full TechCrunch audience
- Featured on the official agenda, Side Events page, and mobile app
- Highlighted in attendee emails and articles
- Custom discount code for you and your team
Must-know details about Side Events
Check out the full details on the TC All Stage Side Events page. As the host, you’ll handle the event details, and we’ll help spread the word. There’s no fee to be listed, and everything must be:
- In-person and Boston-based
- Held July 13-19 (only post-5 p.m. ET events allowed on July 15)
- 18+ only (21+ if there’s alcohol)
Apply now to make a brand impact at TC All Stage
Let’s make it happen! Hosting a Side Event is an easy way to grow your network and connect your brand to Boston’s startup crowd. Apply here to lock in your Side Event before the deadline. This is your moment to show up and stand out.
Keep reading the article on Tech Crunch
Mark Zuckerberg says TikTok slowed Meta’s growth
In Meta’s antitrust trial on Wednesday, Meta CEO Mark Zuckerberg testified that TikTok’s success was a risk to Meta’s business, saying the short-form video app was a “top priority” and a “highly urgent” competitive threat when it arrived in 2018, according to Bloomberg and other outlets.
The reports come from a packed first week of testimony in the U.S. Federal Trade Commission’s trial against the social networking giant. If the FTC’s lawyers prevail, the trial could potentially require Meta to spin out Instagram or WhatsApp as separate companies.
Notably, Zuckerberg admitted on Wednesday that TikTok’s arrival had directly impacted Meta, saying his company had observed that its “growth slowed down dramatically” when the TikTok became popular. He also said the ByteDance-owned app has continued to be a focus of Meta’s competitive efforts for several years.
ByteDance purchased Musical.ly in 2017, and merged it with TikTok the following year. Around the same time, Meta (then known as Facebook) stopped reporting Facebook’s user numbers in its quarterly reports, shifting to a new “family of apps” metric that encompassed Instagram and WhatsApp. The change was designed to hide the fact that Meta’s flagship app was seeing slowing growth.
Zuckerberg made another interesting comment during the trial in response to a question about social media platforms’ “network effects.” He said it’s become less important for social media apps to leverage connections between friends and family to grow.
“The apps now serve primarily as discovery engines,” Zuckerberg told the court. “People can take that content to messaging engines.”
At the same time, however, Facebook is trying to return to its “OG roots” by turning once again to those network effects. The company recently released new features designed to make it easier to connect with friends, including a revamped Friends tab that highlights friend requests and activity. In January, Zuckerberg told investors that a “return to OG Facebook” was a key goal for 2025.
Keep reading the article on Tech Crunch
Shein and Temu to raise prices for US shoppers in response to tariffs
Temu and Shein plan to raise prices for U.S. customers starting next week on April 25th, due to President Donald Trump’s tariffs on goods shipped from China, the Associated Press reports.
The 145% tariff on products made in China, along with Trump’s decision to end a customs exemption that had allowed goods under $800 to enter the U.S. duty-free, has disrupted the business models of both platforms. The report says nearly 4 million parcels, most of them from China, enter the U.S. every day under this soon-to-be-ended exemption.
Shein and Temu have gained significant popularity in the U.S. over the past few years due to the platforms’ discount pricing and influencer advertising.
The Wall Street Journal reported last year that Amazon now sees Shein and Temu as bigger threats to its business than retailers like Walmart and Target. In November, Amazon launched Amazon Haul, a storefront that offers mass-produced and discounted items to take on Shein and Temu.
Shein and Temu are encouraging customers to continue shopping. The companies say they will ensure orders arrive on time and are doing what they can to minimize the impact on shoppers.
Keep reading the article on Tech Crunch
Wasp’s platform is the glue that holds web dev apps together
When Matija Šošić started working in web development, he was surprised by how hard it was to build a full-stack production-ready web application.
One of the biggest hurdles Šošić faced was navigating the fragmented dev tooling landscape. Coding a web application required utilizing different tools for developing the front-end versus the back-end, and so on.
“The whole ecosystem is very modular and granulated,” Šošić told TechCrunch. “There’s a lot of these separate sub-systems, which you have to figure out how to paste together, and then also make sure they are both, you know, scalable and secure with everything you set up.”
In 2020, Šošić decided to pair up with his twin brother, Martin Šošić — who’d faced similar issues in his own developer career — to launch Wasp, a platform meant to connect these fragmented tools, in 2021.
Wasp is a full-stack web app dev tool that acts as the glue between the different platforms developers are already using, including React, Node.js, and Prisma, among others. Wasp helps compile the code from these different platforms together into one web application.
Wasp also spots and flags the gaps that are common when a developer mashes together different coding sources. Wasp will let a developer know if they are missing an API key, for example, or suggest potential code changes to prevent future issues.
That last piece is particularly important in today’s market as numerous new AI coding tools, like Windsurf (formerly Codeium) and Cursor, have made coding accessible to nontechnical folks. While that’s largely a good thing, Šošić said, “vibe coding doesn’t work for enterprise,” and Wasp can help budding developers build more secure, full-stack web applications.
Wasp’s platform is open source and can deploy to a public cloud or an enterprise server.
Šošić said that the founding team decided to build Wasp as a layer on top of existing tools as opposed to a completely new, out-of-the-box solution to eliminate developers having to learn a new programming language or process to use Wasp.
Wasp went through Y Combinator’s winter 2021 cohort and launched its product into beta in 2023. Since then, the company has racked up 26,000 GitHub stars, and it now works with numerous startups and Fortune 500 customers.
The company raised a previously-unannounced $3.7 million round led by HV Capital with participation from Fifth Quarter Ventures, Big Bets, and Metis Ventures, among other VCs, in late 2024. The round also included Ant Wilson, a co-founder and CTO of Supabase, and Søren Bramer Schmidt, the CEO of Prisma.
Wasp last raised funding in 2021 — a $1.5 million seed round. To date, the company has raised a total of $5.2 million.
“This gives us a very solid amount of, basically, freedom to work with,” Šošić said. “We are very excited to execute on the next level of the product, bring it to [version] 1.0, and further solidify the whole positioning in the AI spectrum.”
Šošić said the company is now focused on bringing the product to version 1.0, which will include features like support for different languages and server-side rendering.
“For us, you know, it is still focusing on the core product itself, which is the open source Wasp as well framework,” Šošić said. “With all the feedback that we have gotten from the last four years of building, I think now it’s become clear to us what we have to build and what we have to support to reach [version] 1.0.”
Keep reading the article on Tech Crunch
Archer Aviation wants to help New Yorkers skip airport traffic with electric air taxis
New York City — where you can get a dollar slice at 2am or a McDonald’s delivery in under 30 minutes, but still spend two hours crawling down the highway to catch a flight.
Archer Aviation wants to change that. The electric aircraft startup says its air taxis could one day fly passengers from Manhattan to nearby airports in five to 15 minutes.
The startup on Thursday unveiled its proposed air taxi network for the city in partnership with United Airlines, which would allow passengers to tack on an Archer ride to their traditional airline purchase.
“We’re starting with nine core nodes,” Adam Goldstein, co-founder and CEO of Archer, told me. We — along with Nikhil Goel, Archer’s chief commercial officer — sat comfortably on plush sofas at Casa Cipriani, a members only club at the bottom tip of the island of Manhattan, where we could watch helicopters land at the Downtown Skyport.
“So you have the three big international airports – JFK, La Guardia, Newark,” Goldstein continued. “You have the three big heliports, [including] the Downtown Skyport, and then the east and west side heliports. And then the three big regionals – Teterboro, Westchester, and Long Island Republic.”
Archer has shared a similar vision for air taxi networks in cities like Los Angeles. The company is still awaiting Federal Aviation Administration Approval for its aircraft – a five-seat eVTOL (electric vertical takeoff and landing vehicle) dubbed Midnight – before it can even begin testing out the routes it has planned. Archer also still needs to get a pilot into one of its aircraft to test-fly it – so far Archer has only flown the plane autonomously with no humans inside. Its competitors – Joby Aviation and Beta Technologies – have done piloted flights.
Goldstein said he’s hopeful that Archer will achieve the necessary certifications in 2026, and told TechCrunch that Archer will have an update about piloted flights at its next earnings call. The company went public in 2021 via a special purpose acquisition merger, and has raised $3.36 billion to date per PitchBook via public and private raises.
In the meantime, Archer is laying the groundwork of setting up infrastructure and operations. And for that, partnerships are essential.
In NYC, United will help with things like aircraft storage, maintenance, charging setup, and setting up vertiports (landing pads for eVTOLs) at airports. Archer has also teamed up with the fixed-base operators that run the Manhattan helipads – Atlantic Aviation, Signature, and Modern Aviation – which Goldstein says will give Archer passenger access and help set up the aircraft charging infrastructure.
“What makes New York very compelling is it’s the number one helicopter commuting city in the world, outside of San Paolo,” Goel said. “And so you don’t have to squint very hard to imagine any of these helicopters that are flying by on Hudson simply replacing them with one of our aircraft. The routes are already there. The air traffic already knows how to work with them. The FBOs and the landing facilities are already there. And so there’s no systemic changes that are necessary.”
Archer’s plan is to start small, bringing up to five aircraft to NYC – and other cities – to practice running the routes before launching. Ten to 20 years down the line, the goal is to fly hundreds of aircraft across several cities. Archer last month began building Midnights at its production facility in Georgia, built in partnership with strategic investor Stellantis, and hopes to build 650 aircraft annually by 2030.
Aside from NYC and LA, Archer also plans to launch in San Francisco and Miami, but the timeline is still dependent on FAA certification, and the company hasn’t picked its first U.S. launch city.
Archer plans to launch an air taxi service in Abu Dhabi, where regulations are less strict, later this year. Goldstein said the NYC proposed network gives people a vision they can understand.
“We’re hoping people look at [Abu Dhabi] and say, ‘Oh, it’s real. How is New York gonna work?’”
Keep reading the article on Tech Crunch
Smashing, the reading curation app by Goodreads’ founder, shuts down
Building an app dedicated to helping people aggregate and read news, articles and social media posts in one place is apparently a tough proposition these days. Smashing, an AI-powered app that lets users curate the news and posts they’d like to read, is shutting down, seemingly because of its inability to scale rapidly enough.
“We simply didn’t grow fast enough to keep going. We weren’t able to scale it into a sustainable product,” the company said in an email to customers announcing its closure.
GoodReads’ founder, Otis Chandler, launched Smashing last June, aiming to use AI and the community to curate news articles, blog posts, podcasts and social media posts from around the web. The app let users follow their interests, submit content, and vote on suggested content to indicate relevance. It also had AI-powered summaries and a bot that could answer questions.
The company said in the email that it had seven employees working on the product. Smashing had raised $3.4 million in funding from True Ventures, Blockchange, Offline Ventures, Advancit Capital, Power of N Ventures, and several angel investors.
Thankfully, there are plenty of startups working on the problem Smashing set out to address. We have AI-powered news readers like Bulletin and Particle, as well as feed aggregator apps like Feeeed, Tapestry and Reeder — to pick a few.
Smashing’s closure comes after Instagram’s co-founders shut down their AI-powered news app Artifact last year and eventually sold the tech to Yahoo.
Keep reading the article on Tech Crunch
xAI adds a ‘memory’ feature to Grok
Elon Musk’s AI company, xAI, is slowly bringing its Grok chatbot to parity with top rivals like ChatGPT and Google’s Gemini.
Wednesday night, xAI announced a “memory” feature for Grok that enables the bot to remember details from past conversations. Now, if you ask Grok for recommendations, it’ll give more personalized responses — assuming you’ve used it enough to allow it to “learn” your preferences.
ChatGPT has long had a similar memory feature, which was recently upgraded to reference a user’s entire chat history. Gemini, too, has persistent memory to tailor its replies to individual people.
“Memories are transparent,” reads a post from the official Grok account on X. “[Y]ou can see exactly what Grok knows and choose what to forget.”
Grok now remembers your conversations. When you ask for recommendations or advice, you’ll get personalized responses. pic.twitter.com/UXhX7BjS57
— Grok (@grok) April 17, 2025
Grok’s new memory feature is available in beta on Grok.com and the Grok iOS and Android apps — but not for users in the EU or U.K. It can be toggled off from the Data Controls page in the settings menu, and individual “memories” can be deleted by tapping the icon beneath the memory from the Grok chat interface on the web (and soon Android).
xAI says that it’s working on bringing the memory feature to the Grok experience on X.
Keep reading the article on Tech Crunch
Zoom restores service after an hours-long outage
Zoom says its platform is back online after suffering an outage for much of Wednesday afternoon.
“Service has now been restored after the earlier outage, and we sincerely appreciate your patience and understanding,” Zoom said in a post on X at around 2 p.m. Pacific.
The trouble began around 11:40 a.m., according to The Verge, citing data from Cisco’s ThousandEyes platform. Users on social media reported having trouble logging in, and getting an “Unable to Connect” error message when entering meetings. Zoom’s site also went down — as did its inbox for media inquiries.
More than 59,000 users had reported issues on outage-tracking site DownDetector.com by noon.
It’s not yet clear what the cause was. The Verge speculates that it might’ve been DNS-related, pointing to a Reddit thread. When reached for comment, a Zoom spokesperson referred TechCrunch to the above X post.
Keep reading the article on Tech Crunch
Chapter, a Medicare startup with links to Vance, Thiel, and Ramaswamy, just raised a round at $1.5B valuation
Chapter, a Medicare advisory startup co-founded by former Republican presidential hopeful Vivek Ramaswamy, has closed a $75 million funding round at a valuation of $1.5 billion led by private equity and venture firm Stripes.
The startup helps seniors choose Medicare health plans analyzing doctors, hospitals, and prescription drug coverage. Unlike many other Medicare insurance brokers, Chapter claims to prioritize client needs over insurer profits.
Naraya, the VC firm founded by Vice President J.D. Vance, led Chapter’s Series A round in 2020. Peter Thiel, who also invested in the company, assumed Vance’s board seat when Vance resigned to run for Senate in 2021.
While Thiel has since resigned from Chapter’s board, Democrat Donna Shalala, who served as Secretary of Health and Human Services during the Clinton administration and was later a U.S. Congresswoman, has filled the vacancy. Shalala told Bloomberg that Ramaswamy’s role and investments from Republicans Vance and Thiel were not an issue. “I don’t think of the company as political.”
Keep reading the article on Tech Crunch
Trump administration reportedly considers a US DeepSeek ban
The Trump administration is considering new restrictions on the Chinese AI lab DeepSeek that would limit it from buying Nvidia’s AI chips, and potentially bar Americans from accessing its AI services, The New York Times reported on Wednesday.
The restrictions are part of the Trump administration’s effort to compete with China on AI. Months after DeepSeek jolted both Silicon Valley and Wall Street, U.S. officials seem to be weighing several options to limit China’s access to American technologies and consumers.
On Tuesday, the White House moved to restrict more of Nvidia’s AI chip sales to China, strengthening rules created by the Biden administration.
DeepSeek’s popularity among U.S. AI developers has soared in recent months, and the startup’s competitive pricing has forced Silicon Valley to offer frontier AI models at lower costs.
There are lingering questions, however, around whether DeepSeek engaged in IP theft to create some of its more competitive models. OpenAI has alleged that the Chinese lab distilled its models, violating OpenAI’s terms of use.
Keep reading the article on Tech Crunch
Rippling wants Revolut to reveal who paid off Deel’s alleged ‘spy’
Another highly valued startup has just been added to the mix in the ongoing legal drama between Rippling and Deel: U.K. fintech giant Revolut.
In his blockbuster affidavit that reads like it’s straight out of a movie, Irish Rippling employee Keith O’Brien claimed that Deel paid him about $6,000 a month to spy on Rippling. He alleged the first $6,000 payment was sent to his Revolut account by Alba Basha Westgarth, the wife of Deel’s COO Dan Westgarth, back in November 2024.
At the time, Alba Basha Westgarth was working as the crypto compliance lead at Robinhood, according to a LinkedIn profile reviewed by TechCrunch that has since been deleted. Robinhood told TechCrunch that she left Robinhood earlier this year, but did not explain why.
The deleted LinkedIn profile listed her location as Dubai — that’s the same place where Deel’s CEO and legal director are now located as Rippling tries to serve them papers. The UAE has a reputation as an extradition haven.
Rippling is now suing Revolut to get the full name and address of the “Alba Basha” listed as the sender on that $6,000 receipt, court records show. Rippling is also seeking all account opening documentation — things like IDs and utility bills — to definitively confirm her identity.
In response, Revolut has hired a top lawyer in Ireland. It also sent a letter responding to Rippling last week, which a lawyer for Rippling described as “helpful but complicated” in a court hearing (without elaborating.)
To be clear, there’s no indication or accusation that Revolut has done anything wrong here. The E.U. has strict privacy laws, so supplying that kind of information isn’t simple, especially in a civil case like this one. A key sticking point is that there hasn’t been an order from the Irish court itself on Revolut to provide the information.
Revolut declined to comment specifically on the case but told TechCrunch that, in general, it will “always comply with any court order request for disclosure.”
Whether Revolut ends up providing more details about the “Alba Basha” behind the transfer remains to be seen. The only other option under European privacy laws would be customer consent, which seems unlikely in this case.
Deel, which has denied all wrongdoing to the media, didn’t respond to a request for comment. Deel COO Dan Westgarth and Alba Basha Westgarth didn’t respond either. Rippling declined to comment.
Although the evidence isn’t yet definitive, the Irish High Court gave a clue that it may honor Rippling’s wishes to verify Alba Basha’s identity in some way or form.
After a lawyer for Rippling rattled off evidence connecting Alba Basha to Dan Westgarth, including a Facebook photo of the two together, during an April 2 court hearing, Judge Mark Sanfey remarked that “it could be a coincidence, but it’s unlikely!”
Keep reading the article on Tech Crunch
Former CISA director Chris Krebs vows to fight back against Trump-ordered federal investigation
Former top Trump cybersecurity official Chris Krebs told The Wall Street Journal in an interview on Wednesday that he vowed to fight back against a federal investigation ordered into him by President Trump.
Krebs said in the interview that he will resign from his position at cybersecurity firm SentinelOne in order to challenge the federal investigation, which claims Krebs “falsely and baselessly denied that the 2020 election was rigged and stolen.” The order also stripped Krebs of his security clearance.
“It’s about the government pulling its levers to punish dissent, to go after corporate interests and corporate relationships,” Krebs told the Journal.
Trump hired Krebs in 2018 to lead the newly created U.S. Cybersecurity and Infrastructure Security Agency (CISA), which oversees federal cybersecurity and election security. Trump fired Krebs in November 2020 soon after publicly debunking Trump’s false claims of voting irregularities.
Per the Journal, Krebs joins a growing list of former government officials, law firms, and universities pushing back on the Trump administration’s use of the executive branch to target critics and dissenters.
Keep reading the article on Tech Crunch
OpenAI’s latest AI models have a new safeguard to prevent biorisks
OpenAI says that it deployed a new system to monitor its latest AI reasoning models, o3 and o4-mini, for prompts related to biological and chemical threats. The system aims to prevent the models from offering advice that could instruct someone on carrying out potentially harmful attacks, according to OpenAI’s safety report.
O3 and o4-mini represent a meaningful capability increase over OpenAI’s previous models, the company says, and thus pose new risks in the hands of bad actors. According to OpenAI’s internal benchmarks, o3 is more skilled at answering questions around creating certain types of biological threats in particular. For this reason — and to mitigate other risks — OpenAI created the new monitoring system, which the company describes as a “safety-focused reasoning monitor.”
The monitor, custom-trained to reason about OpenAI’s content policies, runs on top of o3 and o4-mini. It’s designed to identify prompts related to biological and chemical risk and instruct the models to refuse to offer advice on those topics.
To establish a baseline, OpenAI had red teamers spend around 1,000 hours flagging “unsafe” biorisk-related conversations from o3 and o4-mini. During a test in which OpenAI simulated the “blocking logic” of its safety monitor, the models declined to respond to risky prompts 98.7% of the time, according to OpenAI.
OpenAI acknowledges that its test didn’t account for people who might try new prompts after getting blocked by the monitor, which is why the company says it’ll continue to rely in part on human monitoring.
O3 and o4-mini don’t cross OpenAI’s “high risk” threshold for biorisks, according to the company. However, compared to o1 and GPT-4, OpenAI says that early versions of o3 and o4-mini proved more helpful at answering questions around developing biological weapons.
The company is actively tracking how its models could make it easier for malicious users to develop chemical and biological threats, according to OpenAI’s recently updated Preparedness Framework.
OpenAI is increasingly relying on automated systems to mitigate the risks from its models. For example, to prevent GPT-4o’s native image generator from creating child sexual abuse material (CSAM), OpenAI says it uses on a reasoning monitor similar to the one the company deployed for o3 and o4-mini.
Yet several researchers have raised concerns OpenAI isn’t prioritizing safety as much as it should. One of the company’s red-teaming partners, Metr, said it had relatively little time to test o3 on a benchmark for deceptive behavior. Meanwhile, OpenAI decided not to release a safety report for its GPT-4.1 model, which launched earlier this week.
Keep reading the article on Tech Crunch
Zuckerberg: Snapchat would have grown faster if it accepted $6B buyout offer
During Meta’s antitrust trial this week, Meta CEO Mark Zuckerberg said that Snapchat would have grown faster if it accepted his company’s offer to buy the social network back in 2013, Business Insider reports.
Court documents revealed that Meta, then called Facebook, offered to buy Snapchat for $6 billion (reports at the time said that the dollar amount was $3 billion). In response to questioning from an FTC attorney, Zuckerberg said he thought Snapchat “wasn’t growing at the potential that it could” and that his company would have improved the app.
“For what it’s worth, I think if we would have bought them, we would have accelerated their growth, but that’s just speculation,” Zuckerberg testified.
The government brought up the failed acquisition attempt to support its argument that Meta aims to preserve its dominance in the social media landscape by acquiring rivals, rather than competing with them directly.
The FTC is looking to force Meta to restructure or sell Instagram and WhatsApp, arguing that the company spent billions to acquire the apps to fend off Facebook competitors, creating an illegal monopoly.
Keep reading the article on Tech Crunch
Trump-appointed judge orders Trump admin to ‘turn the funding spigots back on’
The Trump administration has made no secret that it dislikes disbursing money authorized by Congress under the Inflation Reduction Act and the Bipartisan Infrastructure Law. But on Tuesday, a federal judge issued an order “requiring the agencies to turn the funding spigots back on.”
Under President Donald Trump, federal agencies have used his executive orders to justify withholding congressionally authorized grants and contracts, many of which had already been awarded. But U.S. District Judge Mary McElroy, who Trump appointed during his first term, said the administration’s actions were “neither reasonable nor reasonably explained.”
“The broad powers that [Office of Management and Budget], the [National Economic Council] Director, and the five Agencies assert are nowhere to be found in federal law,” McElroy wrote.
In addition to Office of Management and Budget and the National Economic Council, five federal agencies are being sued by just as many plaintiffs. The EPA, for example, is being sued by the Childhood Lead Action Project, which received $500,000 to fight childhood lead poisoning in Rhode Island. The other agencies include Agriculture, Energy, Housing and Urban Development, and Interior.
This case is separate from another one, in which the Trump administration told Citibank to freeze hundreds of millions of dollars in funds already held in nonprofits’ bank accounts. In that case, a federal judge said the Trump administration — and specifically the EPA — acted in an “arbitrary and capricious” manner when terminating contracts with three nonprofits. The judge issued a temporary restraining order that required the EPA and Citibank to give the nonprofits access to funds in their accounts.
McElroy acknowledged the Trump administration is within its rights to steer the country in a certain direction, though there are limits.
“The Court wants to be crystal clear: elections have consequences and the President is entitled to enact his agenda. The judiciary does not and cannot decide whether his policies are sound,” the judge wrote.
“But where the federal courts are constitutionally required to weigh in — meaning we, by law, have no choice but to do so — are cases ‘about the procedure’ (or lack thereof) that the Government follows in trying to enact those policies.”
Many companies and nonprofits have objected through court filings to the Trump administration’s control over executive branch departments and agencies to undo the effects of legislation that was passed by Congress and signed into law under the previous administration.
Here, McElroy agrees with the plaintiffs. “Agencies do not have unlimited authority to further a President’s agenda, nor do they have unfettered power to hamstring in perpetuity two statutes passed by Congress during the previous administration.”
Keep reading the article on Tech Crunch
OpenAI is reportedly in talks to buy Windsurf for $3B, with news expected later this week
Windsurf, the maker of a popular AI coding assistant, is in talks to be acquired by OpenAI for about $3 billion, Bloomberg reported.
If the deal happens, it would put OpenAI in direct competition with a number of other AI coding assistant providers, including Anysphere, the maker of Cursor, which OpenAI backed from its OpenAI Startup Fund.
The acquisition could jeopardize the credibility of the OpenAI Startup Fund, given that it is one of Cursor’s biggest investors, said a person familiar with Cursor’s cap table. It’s not clear whether OpenAI approached Cursor about an acquisition.
In addition to what sources told Bloomberg, there are a few more clues that something is going on between the two companies. A couple of days ago, Windsurf users received an email that said that because of an announcement later this week, they have the option to lock in access to the coding editor at $10 a month.
And OpenAI chief product officer Kevin Weil also released a video yesterday praising Windsurf’s capabilities.
Windsurf, the company formerly known as Codeium, has been in talks to raise fresh funds at a $2.85 billion valuation led by Kleiner Perkins, TechCrunch reported in February. The company has reached about $40 million in annualized recurring revenue (ARR), according to our reporting. That revenue run rate is much lower than Cursor’s, which reportedly makes $200 million on an ARR basis. Cursor has been in talks to raise capital at about $10 billion valuation, Bloomberg reported last month.
Since its founding in 2021 by Varun Mohan and his childhood friend and fellow MIT grad, Douglas Chen, Windsuf has previously raised $243 million from investors including Greenoaks Capital and General Catalyst, according to PitchBook data.
Additional reporting by Sarah Perez
Keep reading the article on Tech Crunch