The social media platform X is launching its private messaging service, known as X Chat, as a standalone application. The company announced on Monday that the initial beta version of the X Chat app was available to 1,000 users via Apple’s TestFlight platform.

The beta program reached its maximum capacity just two hours post-announcement. Michael Boswell, a product designer at xAI, noted in an X post that the beta would soon be accessible to more users beyond the first 1,000, although he did not specify a timeline.

“For the past few months, we’ve been secretly developing an independent X Chat app for iOS,” Boswell posted on X. “Try it out. Break it. We are eager for your feedback,” he stated.

X Chat is an enhanced version of X’s current direct messaging feature, which the company asserts is end-to-end encrypted. Nevertheless, security professionals have cautioned potential users that the X Chat service is less secure compared to other encrypted messaging applications like Signal, and may not be reliable. It remains unclear if the company has resolved the security concerns highlighted by the security community in this standalone version.

The launch of this app signifies a notable departure from previous intentions by X owner Elon Musk to transform X into an “everything app,” overseeing messaging, payments, creator content, and more, integrating all service components.

A few early testers have started to release screenshots of the app’s interface. One user mentioned that the beta version stylized the name as xChat rather than X Chat, indicating a possible rebranding. Others have been sharing images from the app, including the login interface, which features a starry backdrop.

Initial users have described the app as offering a more streamlined and simple interface for using X Chat compared to the X application.

xAI’s Grok responded to inquiries regarding the new app on X, confirming additional information such as the app’s age rating, which will match that of X (17+ on the iOS App Store), and that chats will be synchronized with both the X app and the chat.x.com web application that launched in December 2025.

The AI chatbot also mentioned that an Android variant of the X Chat app is anticipated “very soon.”

I received access to the XChat beta!

It’s time to rigorously test it and check for bugs.

I can already say that a fantastic addition to X chat would be an XSpaces feature.

Transform it into a Discord-like app with chat, images, video, and also live video/voice chat! pic.twitter.com/CQ6xfQ4JcB

— Overly Trev (@OverlyTrev) March 3, 2026

Apple revealed its latest line of laptops on Tuesday morning, featuring new MacBook Air and MacBook Pro variants that incorporate Apple’s M5 chips. The Pro models were introduced alongside the cutting-edge M5 Pro and M5 Max chips, which Apple claims to be its most sophisticated CPU cores to date.

The company mentioned that these enhanced M5 chips were specifically tailored to improve the MacBook Air and MacBook Pro laptops’ performance for demanding AI operations, which are increasingly becoming a priority for new Apple devices. As per Apple, both the new Air and Pro laptops are capable of executing AI tasks up to 4 times quicker than their M4 predecessors.

These AI-focused enhancements may not be instantly apparent to casual users who aren’t managing complex AI networks or producing rapid 3D visualizations. However, these innovations affect various other elements of the laptops as well.

Users of the MacBook Air benefit from features such as 18 hours of battery life (a six-hour increase compared to the previous Intel-based Apple laptops from 2020), along with a 12MP Center Stage camera for video conferencing, a three-microphone setup, and a sound system compatible with Spatial Audio and Dolby Atmos. The MacBook Air includes two Thunderbolt 4 ports, a MagSafe charging connection, and a traditional 3.5mm headphone jack.

The refreshed MacBook Air lineup consists of a 13-inch model (starting at $1099) and a 15-inch model (starting at $1299), available in color options of sky blue, midnight, starlight, and silver. The Air also now offers a base storage of 512 GB, which doubles the storage capacity of the previous model.

As is customary, the MacBook Pro is tailored for more advanced users, particularly developers engaged with AI. The M5 Pro and M5 Max chips deliver up to 4 times the speed in LLM prompt processing compared to the M4 Pro and M4 Max, and are up to 8 times quicker at AI image creation compared to the M1 Pro and M1 Max.

Apple indicates that this empowers AI researchers and developers to train custom models directly on their devices, and creative individuals may reap the rewards of accelerated 3D rendering, video production, and music composition tasks.

The MacBook Pro also boasts read/write speeds that are up to 2 times faster than the previous generation, featuring a starting storage of 1TB for the MacBook Pro with M5 Pro, and 2TB for the MacBook Pro with M5 Max. Apple claims these laptops can achieve up to 24 hours of battery life, and with a 96W or higher USB-C charger, users can reach 50% battery power in just thirty minutes. These laptops come with Thunderbolt 5 support and are equipped with a six-speaker audio system.

The 14-inch and 16-inch MacBook Pro models equipped with M5 Pro chips begin at $2,199 and $2,699, respectively, while the M5 Max chip models start at $3,599 and $3,899, offered in either black or silver finishes.

All of these new laptops will be up for preorder on Tuesday, March 4, and will be available for purchase starting Wednesday, March 11.

On Tuesday, Apple revealed a new MacBook Pro and MacBook Air, each equipped with the latest chips, alongside a fresh line of displays. This lineup features a $1,599 Studio Display and a $3,299 Studio Display XDR.

The updated 27-inch displays boast enhanced cameras and superior connectivity.

Each of the displays includes a 12MP Center Stage camera, which the company claims delivers better image clarity. It also offers Desk View, a functionality that simultaneously displays your face and an overhead perspective of your workspace.

The Studio Display features a 5K Retina screen with over 14 million pixels, 600 nits brightness, and adheres to the P3 wide-gamut color standard, which encompasses a wider array of visible colors compared to standards like sRGB.

These displays are equipped with Thunderbolt 5 ports, enabling connections to peripherals and daisy-chaining of up to four displays (a Thunderbolt 5 Pro cable is included). They also feature a three-microphone array and a six-speaker audio system with Spatial Audio. Apple claims the upgraded sound system provides 30% deeper bass compared to the previous generation.

The higher-tier Studio Display XDR offers a 5K Retina XDR screen (5120×2880 resolution) with a mini-LED backlight and over 2,000 local dimming zones. It can reach up to 1000 nits of SDR brightness and 2000 nits of peak HDR brightness, boasting a contrast ratio of 1,000,000:1. Additionally, it supports the Adobe RGB color standard.

The display features a 120Hz refresh rate, Adaptive Sync support, and includes DICOM medical imaging presets alongside a Medical Imaging Calibrator, making it suitable for medical applications. Apple has indicated that the Medical Imaging Calibrator for macOS is awaiting FDA approval and is anticipated to be available shortly in the U.S.

While the stand for the Studio Display can be adjusted for tilt, the Display XDR also offers height adjustments.

Both options are available in standard or nano-texture glass, as noted by Apple.

Pre-orders will commence on March 4, with availability starting on March 11.

On Tuesday, Apple introduced the newest members of the M-series chips, unveiling the M5 Pro and M5 Max, which are driving the latest MacBook Pro.

According to the tech giant, these chips are designed around its innovative Fusion Architecture, a sophisticated configuration that integrates two dies into a single, high-capacity system on a chip (SoC), encompassing a robust CPU, adaptable GPU, Media Engine, unified memory controller, Neural Engine, and Thunderbolt 5 support.

Each chip incorporates an 18-core CPU, representing an enhancement from the 14-core setup found in the M4 Pro and the 16-core in the M4 Max.

The CPU now includes six “super cores,” which is Apple’s designation for its highest-performing cores, as well as 12 newly developed performance cores. Together, this configuration increases performance by up to 30% for professional tasks.

“The GPU elevates the next-generation architecture introduced in M5 to a maximum of a 40-core GPU,” Apple stated in a press announcement. “With a Neural Accelerator embedded in each GPU core and enhanced unified memory bandwidth, M5 Pro and M5 Max achieve more than 4 times the peak GPU computing power for AI over the previous generation.”

Overall graphics performance has improved by as much as 20%, with ray-tracing tasks seeing enhancements of up to 35%.

M5 Pro offers support for up to 64GB of unified memory, an increase from 48GB on the M4 Pro, and features a bandwidth of 307GB/s. M5 Max continues to support up to 128GB of unified memory, with bandwidth raised to 614GB/s.

Apple indicates that the M5 Pro targets professional users including data modelers, sound designers in post-production, and STEM students who require robust CPU and GPU capabilities, along with substantial amounts of unified memory for intricate projects and tasks.

M5 Max is aimed at professional users such as 3D animators, application developers, and AI researchers who execute workloads necessitating peak GPU performance and the highest unified memory bandwidth, according to the tech giant.

The latest MacBook Pro models will be open for pre-orders tomorrow, with general availability starting on March 11.

For businesses, analyzing data offers much more than just novel income avenues. The contemporary enterprise technology stack is astonishingly intricate — it utilizes a multitude of tools that interact in various ways, often breaking systems in unexpected manners, which is why the ability to scrutinize data streams enables firms to pinpoint when, where, and why failures occur.

However, security teams cannot afford to wait for failures to initiate repairs. To illustrate: an alarm that hasn’t sounded in a long time cannot be relied on to function properly. Moreover, the modern security stack is so packed with various tools that even a minor modification in one can lead to unpredictable consequences that might endanger detection and response abilities.

Fig Security, a startup established by alumni of Israel’s cyber and data intelligence units 8200 and Mamram, asserts it aids security teams in managing this challenge by overseeing the security stack to determine if their rules, mitigation resources, and detection and response functions are operational or misaligned due to changes. The startup has recently emerged from stealth mode with $38 million in seed and Series A financing, as exclusively reported by TechCrunch.

In essence, the startup’s technology tracks data movement in the security stack, from the initial sources through data pipelines and lakes to security orchestration and automation response platforms, alerting security teams when alterations at any stage impact detection or response functionality. The platform also enables organizations to test how new fixes, patches, or adjustments might influence their systems prior to implementation.

“Rather than tracing data forward to see its eventual outcome, we examine your detections, as that is what you need to operate,” explained Fig’s CEO and co-founder, Gal Shafir (depicted above, center), to TechCrunch. “Detection or response serves as the single source of truth, and then we backtrack to assess the health and necessary adjustments to the data to trigger detection when an event occurs. Then we notify [the security team] if any inconsistency arises in real-time.”

Shafir indicated that Fig achieves this by sampling a company’s data as it flows through various tools within the infrastructure, analyzing how it transforms through the pipeline. This process allows the company to establish a “data lineage” that can reveal how any upstream modifications could disrupt security tools downstream in real time.

The startup claims to integrate with data links and Security Information and Event Management (SIEM) systems to accomplish all of this, allowing its technology to be compatible with a variety of security tools and settings.

Fig’s debut aligns with enterprises adapting in real-time, particularly as C-suites face pressure to explore how AI-driven tools can aid in cost savings, minimize human errors, and enhance efficiency. Yet, the overwhelming number of tools has complicated the lives of modern security teams. Which defenses should a CISO emphasize? What security stance is appropriate as hackers utilize AI to execute increasingly sophisticated attacks?

Shafir, who previously led Google Cloud Security’s global architecture group before founding Fig, notes that he witnessed this uncertainty directly when engaging with clients while promoting Google’s AI offerings.

“All CISOs, irrespective of team size, security budget, or company scale, were expressing, ‘I appreciate that AI is impressive, but how can I trust my detections at present, and subsequently trust the AI that assures me that everything will be alright tomorrow if my data foundation is untrustworthy?’” 

This prompted Shafir and co-founders Nir Loya Dahan (CPO) and Roy Haimof (CTO) to realize they could address security teams’ challenges in comprehending the current landscape.

“That was the pivotal moment when we recognized the substantial problem that was universally acknowledged but lacked a solution due to the multitude of vendors and the inherent complexity of infrastructure. That was the turning point for us to discontinue our prior endeavors and focus on establishing Fig,” Shafir remarked.

Shafir indicates that since its inception eight months ago, Fig has acquired significant enterprise clients numbering in the “low double-digits,” and he anticipates that this figure will grow to between 50 and 100 by year-end.

The startup intends to utilize the new funding to expand operations in North America and triple its workforce across engineering and market outreach.

Backers in the funding rounds comprise Team8 and Ten Eleven Ventures, alongside security experts like Doug Merritt (former CEO of Splunk), Rene Bonvanie (former CMO of Palo Alto Networks), and the founders of Demisto and Siemplify.

For months, speculation has surrounded Silicon Valley’s wealthy elite about finding a challenger for Representative Ro Khanna. Early on Tuesday, that challenger made his intentions clear.

Ethan Agarwal (shown above), a 40-year-old technology entrepreneur without prior political experience, informed TechCrunch on Monday night of his candidacy for California’s 17th congressional district. This development could lead to one of the most lavishly financed primary battles of the 2026 election cycle.

The competition highlights Khanna, a 49-year-old Democrat who is often viewed as a potential contender for the 2028 presidential election and has publicly endorsed a temporary wealth tax in California. His support has angered some of the state’s wealthiest founders and investors, yet Khanna has remained steadfast, introducing national legislation on Monday with Senator Bernie Sanders to implement a 5% annual wealth tax on all Americans with a net worth of $1 billion or more — a measure their offices project could generate $4.4 trillion over a ten-year period.

There’s a certain irony in this scenario. Agarwal is a Wharton graduate and spent three years at McKinsey before launching audio fitness company Aaptiv, which he sold in 2021. Recently, he co-founded financial services firm Coterie, which is supported by Andreessen Horowitz.

When Khanna first campaigned for this same seat in 2014, he was the tech-supported outsider, buoyed by tech figures such as Marc Andreessen, Sheryl Sandberg, and Eric Schmidt. He faced the popular Democratic incumbent Mike Honda, lost that bid, but returned in 2016 to secure victory.

Critics at that time labeled Khanna a puppet. A decade later, the same accusation will undoubtedly be aimed at Agarwal, who seeks to unseat him.

What follows is a revised excerpt from our discussion with Agarwal.

TechCrunch: Last summer, you revealed your intent to run for governor of California. Now you are opting for a congressional campaign instead. What prompted this change?

Agarwal: I made the decision to run for governor back in July when the candidate pool was quite sparse. Without a political background — coming from a tech background — I saw a few strong entrants like Matt Mahan, who I consider substantial. I’ve followed Ro since his first congressional race in 2012 — I was a significant supporter. However, in recent years, he has gradually leaned left, and when he expressed support for the wealth tax in late December, it was the last straw for me. I recognized I could have a greater impact by running in the 17th district and challenging Ro.

TC: Who is financing your campaign?

Agarwal: We’re filing the necessary paperwork tomorrow, so we don’t have a bank account established yet, and fundraising will commence afterward. That being said — [Y Combinator CEO] Garry Tan is supporting me, along with [DoorDash co-founder] Stanley Tang, and numerous others from the tech sector whose names will emerge in the upcoming days and weeks.

[Editor’s note: The involvement of Tan, Tang, and others will likely ignite a familiar line of attack: suggesting that Agarwal is more a puppet candidate than an independent voice for billionaire concerns. It’s important to note that Khanna faced very similar criticisms when he initially campaigned.]

TC: Can you provide more details about your plans? Besides closing loopholes, is there an alternative to the billionaire tax?

Agarwal: One approach is taxing loans taken against assets. Extremely wealthy individuals typically secure loans against their holdings and pay minimal interest. Since it’s classified as a loan, it escapes taxation. I believe taxing those loans is entirely reasonable.

Another area is capital gains — California currently has a rate of 13.4%, and I think considering an increase is fair. Thirdly, many homes in California are owned by private equity firms or individuals who hold them as investments. I contend that significantly higher property taxes should apply to a home used as an investment compared to one serving as a primary residence. This would both increase revenue and alleviate pressure on families residing in their own homes.

[The loan-tax concept has been circulating in affluent circles for a while — notably promoted by VC Chamath Palihapitiya, though it may have originated with hedge fund titan Bill Ackman. The proposal seeks to treat loans secured by stock assets as taxable occurrences, thus closing a long-standing loophole that allows investors to access their portfolio’s worth without selling, and consequently without incurring capital gains taxes.]

TC: If you make it to Washington, what will your primary three focuses be?

Agarwal: First, prohibiting stock trading for members of Congress and their families. Second, banning corporate PAC contributions. Third, establishing term limits.

[Earlier in our conversation, Agarwal elaborated on the 5,000 children living below the poverty line in the 17th district — which is the wealthiest congressional district nationally — and described his goal of making it “the first congressional district in history to eliminate childhood poverty” as one of his initiatives. That point did not rank among his top three priorities.]

TC: You have accused Ro Khanna of being an active stock trader. Can you clarify?

Agarwal: He has traded more stocks than any Democratic congressman in U.S. history — in sectors like tobacco, oil and gas, Big Pharma, and big tech. He publicly proposed a ban on congressional stock trading, then proceeded to execute 4,000 trades last year. Even if the bill didn’t pass, he could still self-impose it. As for me, I plan to divest my entire portfolio on my first day in office, ensuring no one questions whether my votes reflect my personal financial interests or my true beliefs.

[Both statements warrant examination. Khanna has co-sponsored the TRUST in Congress Act and proposed reforms advocating a ban, but has yet to author standalone legislation. Regarding the trading figures, Khanna has consistently stated he doesn’t personally own or trade individual stocks, and that the trades in question belong to his wife, whose assets are held in an independently managed trust — thereby avoiding any conflict under the Office of Government Ethics regulations. Whether this clarification will satisfy voters remains to be seen.]

TC: Should social media platforms be held accountable for negatively impacting teens? Currently, Section 230 of the Communications Decency Act absolves them of responsibility for user-generated content. What are your views on modifying this?

Agarwal: I believe Section 230, as it was initially drafted [in 1996], served a valuable purpose. The objective was for platforms to function mainly as hosts. However, as their roles have shifted, they now influence what we see through the algorithms they implement. I don’t think it’s appropriate to hold social media companies fully liable for user content — the volume is too staggering, and having a third party make subjective judgments about harmful content leads to potentially perilous outcomes.

That being said, I do think it’s essential to reassess the impact on teenagers’ mental health in the long term. If you consult Meta, or X, or any of them, they will all affirm they don’t gain from harming teenagers. We share the common goal of wanting a better outcome.

TC: What about supervising AI firms, many of which operate right in your vicinity?

Agarwal: I view this through the lens of national security. It’s crucial for America to possess the most advanced models, and if we fail to develop them, China will outpace us.

Some regulations are sensible — AI shouldn’t endanger individuals or others. Nevertheless, I don’t believe we should constrain companies’ ability to innovate and enhance these models. Allowing them to succeed is vital, primarily for national security reasons.

TC: Do you think we need something akin to an FDA for AI?

Agarwal: I’ve encountered that notion. The FDA has largely excelled in safeguarding the health and well-being of Americans — I place trust in the individuals working there, which I cannot affirm for most government entities. If we can establish an independent, nonpartisan authority with rotating terms, that seems reasonable to me. However, it should be designed to bolster America’s national security, not politicized motives.

TC: What about prediction markets like Polymarket and Kalshi? Do they require more oversight?

Agarwal: To clarify, Kalshi and Polymarket are regulated by the CFTC already. Part of the confusion stems from the surge in sports betting applications, creating ambiguity about what is permissible in various states, leading to the emergence of alternatives like Polymarket and Kalshi. However, their existing regulatory framework is generally effective.

TC: How do you intend to conduct your campaign? Are you dedicating yourself to this full time?

Agarwal: This is my utmost priority. I attended [Harker, a private school in San Jose, CA] located within the district. I’ve spent my life nearby. I know countless residents. My campaign will be grassroots-driven — I’m heading to Chinese and Hindi educational institutions, cultural festivities. Holi is approaching; Chinese New Year, Purim is on Tuesday. I’m attending all these events, meeting with people, and visiting small businesses.

I believe this underscores the core difference between Ro and myself: he is cultivating a national presence, which is fine if that’s his ambition. However, he’s doing this while neglecting the needs of his constituents. I’m committed to California. I’m not viewing this as a stepping stone. He’s focused on the national stage; I’m concentrated on local matters. I think the constituents in the 17th district recognize they need a representative devoted solely to them.

TC: What motivated you to enter politics initially?

Agarwal: This might sound cliché, but — my father arrived here with nothing, earning $14,000 a year upon arrival. He established a company, took it public, and sold it. I am benefiting from the fruits of his labor. I’ve created two companies and sold both.

Now, I observe individuals in my vicinity no longer reaping the rewards of the same system that enabled my success. The residents here are industrious, have great potential — but the current environment no longer supports them. I’ve spent enough time expressing my concerns, and I felt it was time to take action.

TC: Is this the beginning of a political journey for you?

Agarwal: This isn’t a shift in career direction. I recognize a distinct problem in the 17th district that I am eager to address. I will enforce term limits on myself — I won’t serve more than five terms — and likely return to the private sector afterward. Public service should be a vocation, not merely a job. I genuinely believe constituents deserve better representation when it doesn’t morph into a profession. Even if a term limits measure doesn’t succeed, I intend to enforce it upon myself. That reflects my true convictions.