The company said it expects total expenses for 2026 to come to between $165bn and $169bn.
Tech giant Meta’s income and operating margin for the second quarter of 2026 were down in comparison to the same time last year amid significant outgoings for the period.
Although revenues were up 28pc to $60.8bn for the period ending 30 June, the company saw a big jump in costs and expenses from around $27bn in Q2 2025 to more than $42bn this year.
Total income fell from $20.4bn to $18.8bn, with operating margin dropping from 43pc to 31pc. Net income fell year-on-year from $18.3bn to $15.8bn.
Advertisement
The 55pc year-on-year jump in expenses includes a $2.4bn outlay on charges around various legal proceedings and $1.8bn in severance costs invoked from recent layoffs by the company. It said it expects total expenses for 2026 to come to between $165bn and $169bn.
Meta’s free cash flow for the most recent quarter was $784m, compared to more than $8bn a year earlier, following declared spending of more than $31bn on leases, property and equipment.
It said it expects capital expenditures for 2026 to amount to between $130bn and $145bn.
“As AI usage in our products and businesses continues to ramp, we continue to invest aggressively in infrastructure to meet the demand,” said Meta CEO Mark Zuckerberg on the company’s quarterly earnings call yesterday.
Advertisement
“Overall, we expect that a significant portion of our compute is going to go towards training our models, growing our core business, and delivering personal agents and new products. But we also expect to grow a large business serving large customers as well.”
Meta shares were down last night following an announced projected revenue for Q3 of around $62.5bn – lower than analysts’ expectations of more than $63bn, according to media reports.
Commenting on the financial results, analyst Mike Proulx of Forrester said that “Meta believes AI infrastructure is now a strategic asset, but its bill is arriving faster than the payoff,” noting that “what it generated in cash this quarter almost all got eaten by AI infrastructure spending”.
He added: “Meta’s AI spend was easier to celebrate when margins were expanding. It’s harder to celebrate now that the costs are showing up in the numbers.
Advertisement
“There’s a bit of similarity to Meta’s metaverse missteps in that Meta is once again spending ahead of proven product demand. The difference is that AI adoption and value are real.
“Investors now have to decide whether Meta’s growing list of AI initiatives represents company diversification or distraction. What makes that question more complicated is that Meta’s legal and regulatory challenges are getting more expensive, too.”
The company’s reported headcount as of 30 June was put at 75,472, which still includes approximately 8,000 employees who have recently been or will soon be laid off.
CFO Susan Li told the earnings call that the company would “continue to monitor active legal and regulatory matters that could significantly impact our business and financial results”.
Advertisement
She added: “For example, we continue to see scrutiny on youth-related issues in several markets and have a number of youth-related trials scheduled for this year in the US, which may ultimately result in a material loss.”
Proulx noted: “Meta’s biggest regulatory battles typically centred on privacy and competition. Now the pressure is mounting around youth wellbeing, addiction and platform safety.
“That’s a different kind of risk because it impacts the future audience growth that powers Meta’s ad business that’s literally underwriting the company’s exorbitant AI costs.”
Don’t miss out on the knowledge you need to succeed. Sign up for the Daily Brief, Silicon Republic’s digest of need-to-know sci-tech news.
Artificial intelligence came at just about the right time, speeding up app and software development as the world started to contend with skills shortages, but it changed the pace so much that security teams have not been able to keep up.
Recently, we’ve seen AI being applied across multiple other domains with role-specific agents and tools, but that’s introduced its own challenges. While tools like Claude Code have proven a hit for generating, reviewing and editing code in seconds, security-focused tools like Anthropic’s Claude Mythos family of models are having broader impacts on the industry.
Anthropic itself has even admitted that Mythos is so powerful that the worry it could be abused by malicious criminals is extremely real – the Preview model is currently only available to a select number of pre-approved partners.
Advertisement
Latest Videos FromTechRadar
So with AI now capable of inspecting code, discovering vulnerabilities and suggesting fixes, do organizations even need as many human workers on the case, or can they get by with significantly fewer humans in the loop serving as AI reviewers? Recent layoffs have certainly implied as much.
The evolving role of security workers in an AI-first world
But with the entire lifecycle of development now amplified by AI, experts are warning that companies could actually be creating more work for themselves, and more than they could ever handle, leaving them facing strains from angles they weren’t previously exposed to.
Advertisement
For example, fewer than one in 10 companies now fix 90% of identified vulnerabilities within 90 days – implying that the volume of vulnerabilities is indeed increasing, rather than that fix efficiency is slipping.
Anthropic even revealed that its around 50 early Mythos Preview partners discovered more than 10,000 high- or critical-severity vulnerabilities – and thousands more of lesser significance.
Sign up to the TechRadar Pro newsletter to get all the top news, opinion, features and guidance your business needs to succeed!
Advertisement
Checkmarx CEO Sandeep Johri predicts we could soon find a balance, where vulnerabilities volume matters less and we revert our focus back toward exploitable risks. I spoke with Johri about the evolution of AppSec, where AI is and isn’t useful, and how organizations can balance speed and control.
With the rise of AI coding tools and AI-generated software, some are questioning whether traditional application security practices are becoming outdated. Is AppSec actually becoming obsolete, or is it evolving?
Traditional application security is not obsolete. It is evolving to meet the reality of how software is being built today.
For years, the process was fairly linear: developers wrote code, security teams scanned it, and vulnerabilities were addressed later. That approach becomes much harder when software is being created at a much faster pace with the help of AI.
AI accelerates development and risk simultaneously: 70% of developers say AI-generated code created more vulnerabilities in 2025, according to our research. As code volume and complexity compound, security needs to move earlier into the development process, giving developers the tools and guidance they need while they are building.
Advertisement
Security teams will continue to play a critical role to help organizations develop software and maintain confidence in their enterprise applications. But their focus needs to shift from finding vulnerabilities to remediating them at scale, because we are tracking an enormous gap in most companies. Our data finds that fewer than 10% of organizations fix 90% of identified vulnerabilities in 90 days.
AI coding tools are helping developers create software faster than ever before. What new security challenges does this introduce for organizations adopting these technologies at scale?
The biggest challenge is that development speed is increasing faster than many security processes can keep up with. AI coding tools allow teams to create and deploy software quickly, but the code generated by AI still needs to be reviewed, tested, and secured.
Companies that ship 81-100% of their code with AI are nearly three times more likely to ship vulnerable code than those who use AI 1-20% of the time. This volume can overwhelm security teams with thousands of findings, many of which don’t represent meaningful risk. The priority needs to be identifying the vulnerabilities that actually create exposure and helping teams fix those issues faster.
Many organizations are looking to AI to help identify and fix security vulnerabilities. Why shouldn’t companies rely solely on AI models to secure the code that AI is helping create?
AI is a valuable tool for security teams, but organizations still need accuracy, context, and human oversight. AI can help identify patterns, analyze code, and accelerate remediation, but security decisions require confidence in what risks actually matter.
Frontier models can uncover hidden exploit paths, but they can also deliver inconsistent findings and false positives. Their results may change depending on the prompt, and they can still miss known critical vulnerabilities.
The challenge with relying only on AI is that organizations may create a false sense of security, or “automation bias.” AI models can generate code and help analyze vulnerabilities, but they need to be paired with security expertise and proven security practices.
Advertisement
The most effective approach combines AI-driven capabilities with strong security foundations, so teams can move faster while maintaining control over risk.
As companies adopt more AI tools throughout the development process, what are the biggest security risks they need to consider beyond just AI-generated code?
Organizations need to think beyond the code itself and look at the entire AI ecosystem being introduced into software development. Many companies are adopting AI tools, models, agents, libraries, and other components faster than they can establish governance around them. This creates visibility challenges because security teams may not know what AI technologies are being used, where they exist in applications, or whether they meet security requirements.
Another concern is shadow AI, where employees use AI tools without formal approval or oversight. Organizations need visibility, clear policies, and a way to manage these technologies as part of their overall software supply chain.
Perhaps the most urgent problem is the expansion of the attack surface itself. With LLMs, it has never been faster, cheaper, or easier for bad actors to exploit software. Issues that sat undetected for years are now being surfaced and weaponized at machine speed. Of the vulnerabilities Mythos has found so far, 99% haven’t been patched, according to Gartner.
How does the rise of AI change the role of security teams? Does the traditional approach to finding vulnerabilities need to shift toward a model focused more on prioritization, remediation, and continuous protection?
Identifying vulnerabilities is no longer enough when organizations already have more findings than they can realistically address. Security has to become continuous, embedded in development workflows, working in lockstep with developers, to build securely from the start while maintaining visibility and control.
We are shifting the focus to understand which issues create the greatest risk to give developers the context to address them fast, where code is written in the IDE.
Advertisement
Fidelity now matters more than volume. One verified true positive is worth more than a hundred low-confidence findings. If developers can’t trust what they’re shown, they’ll start ignoring it. That’s why organizations are increasingly looking at metrics like F1 score, which measure precision and recall together, rather than raw finding counts.
What does the future of application security look like in an AI-driven software development world? Will organizations need a different approach to balancing speed, innovation, and security?
The future of application security will require a more integrated approach. Organizations are going to continue adopting AI because the productivity benefits are significant, but security needs to evolve alongside that innovation.
Security will become more agentic, more intelligent, and more closely connected to the development process. Part of that evolution is combining deterministic, rules-based scanning with AI-driven reasoning in a single process, rather than running them as separate, disconnected tools. Deterministic methods catch what’s already proven; AI reasoning catches what’s novel. Together they’re more complete than either alone.
In addition, deterministic models have real cost advantages. Asking a frontier model to reason its way to security (i.e. extra review passes, self-generated threat models) burns tokens fast. That cost compounds the longer a vulnerability survives: cheap to fix in the IDE, more expensive in CI/CD, most expensive once it’s live in runtime. And every time a developer has to stop and pull a vulnerability out of code that’s already shipped, that’s velocity lost to rework instead of innovation.
Teams will need technology that can help identify real risks, support faster remediation, and provide visibility across the entire software lifecycle. The organizations that succeed will be those that make security part of how they build software, allowing developers to move quickly while reducing unnecessary risk.
Advertisement
For organizations that are embracing AI coding tools today, what steps should they take to make sure they can innovate quickly without introducing unnecessary security risks?
The first step is visibility. Organizations need to understand where AI is being used, what tools are being introduced, and what impact those tools have on their applications.
Remediation is far cheaper the earlier it happens — catching an issue in the IDE costs a fraction of catching it further down the pipeline. But there’s another unsettling gap in our research: nearly all developers have access to in-IDE security tools, but fewer than one in five actually secure code as they write it. The cost of fixing that issue compounds as it passes through later stages of development.
In addition, organizations need clear governance around AI adoption, because only 22% currently have formal AI governance policies in place. That means defining policies, monitoring usage, and making sure teams have the right security controls as they continue to innovate.
AI will continue to change software development. The companies that benefit most will be the ones that embrace the technology while building security into the process from the beginning.
The Federal Trade Commission is suing Hims & Hers, accusing the telehealth company of sharing customers’ sensitive health information with advertising platforms including Meta and Snap.
The complaint, filed on 29 July and joined by Utah and California, also alleges the company charged people without proper consent and made subscriptions deliberately hard to cancel.
The privacy claim is the most serious. The FTC says Hims & Hers passed customers’ health details to third-party advertisers, through uploaded customer lists and automatic tracking that fired off user actions to the platforms, echoing how hospital websites have leaked patient data despite promises to protect it.
The data at issue is not trivial. Hims & Hers sells treatments for conditions people rarely discuss in public, from hair loss to erectile dysfunction to mental health, which makes an alleged leak to ad platforms unusually sensitive.
Advertisement
Hims & Hers has grown into a telehealth giant on exactly these categories. It expanded into weight-loss drugs and built a subscription model that turned stigmatised prescriptions into a mass-market online business, which is why the data it holds is so revealing.
The billing allegations run alongside. The FTC says customers were charged the moment they submitted an intake form, even though the company implied they could speak to a provider first, and were enrolled in recurring subscriptions without a clear chance to review their options.
Then came the hard part: leaving. Before 2023, cancelling required contacting customer service by phone, email, or chat, and even after Hims added an online option, the FTC says it buried the cancel button behind multiple steps, a textbook dark pattern of the sort the agency has been chasing across the web.
The agency’s language was unsparing. Consumers were “unknowingly locked into recurring subscriptions” while their “most private health information” was disclosed to third parties, said Christopher Mufarrige, the FTC’s consumer-protection director.
The legal basis spans several laws. The complaint leans on the FTC Act, the Restore Online Shoppers’ Confidence Act, and consumer-protection and false-advertising laws in Utah and California, a multi-front case rather than a single charge.
Advertisement
The market reacted at once. Hims & Hers shares fell about 10% on the news, a sign investors read the suit as a real threat to a company whose growth has been built on frictionless online sign-ups.
The case is part of a wider reckoning. Regulators have spent the past few years pursuing health and wellness services that quietly fed sensitive data to ad platforms, and Hims & Hers, with its scale and its intimate categories, is a high-profile target.
Meta and Snap are not the defendants, but they hover over the case. The tracking tools at issue are the advertising pixels and data pipelines that power much of the online ad economy, and health data flowing into them has become a recurring legal flashpoint.
The legal gap is part of the problem. Federal health-privacy law was written for hospitals and insurers, not for ad-funded apps, which has let sensitive data flow to platforms in ways patients rarely understand.
Advertisement
The dark-pattern allegations may resonate more widely. Hard-to-cancel subscriptions are a familiar consumer grievance, and the FTC has made them a priority, so a case pairing privacy breaches with a buried cancel button is one it will want to win in public.
Hims & Hers did not comment in the FTC’s announcement. The company has grown fast by making telehealth feel as easy as ordering anything else online, and the suit argues that some of that ease came at the customer’s expense.
The commission voted 2-0 to file. The case now heads to federal court in northern California, where the questions will be whether the data sharing broke the law and whether the sign-up and cancellation flows crossed from aggressive into deceptive.
For a sector built on convenience, the message is pointed. Telehealth promised to strip the friction out of getting care, and the FTC is now testing how much of that friction was removed from the company’s side and quietly added to the customer’s.
We covered Amazon’s $99 AirPods 4 deal yesterday, and today AirPods Max 2 are $100 off, bringing the over-ear headphones down to $449.
AirPods Max 2, which were released in 2026, are $100 off at Amazon today, with all five color options eligible for the triple-digit markdown at press time.
This AirPods deal reflects the lowest price seen this month on the over-ear headphones. In our hands-on AirPods Max 2 review, we found the 2026 release delivers better active noise cancellation (ANC) and great call quality.
Advertisement
On the earbuds side, AirPods 4 and AirPods Pro 3 are on sale as well, with prices as low as $99.
Today’s best AirPods deals
AirPods Max 2 highlights
Powered by Apple’s H2 chip
Up to 1.5x stronger Active Noise Cancellation than first-gen AirPods Max
Transparency mode
Adaptive EQ
Lossless Audio and ultra-low latency audio via a wired USB-C connection (requires a supported service)
Netscape still ruled the browser market, but Microsoft had Windows 95, WorldNet, and a very valuable foothold
It is 30 years since AT&T handed Microsoft a valuable foothold in the browser wars by making Internet Explorer 3 the default for its WorldNet service.
Windows 95 was barely a year old when AT&T set out to challenge CompuServe and America Online with its WorldNet service in 1996. The telecommunications giant needed a browser. Netscape Navigator dominated the market, but Microsoft was keen to make up lost ground.
Advertisement
A promotion and distribution agreement was announced on July 25, 1996, with a further announcement in October.
Under the deal, AT&T WorldNet software was included in versions of Windows 95 supplied to PC manufacturers, with Internet Explorer 3 designated as the service’s default browser. Netscape Navigator remained available, but Internet Explorer secured the valuable default slot.
Microsoft’s browser-bundling strategy would later draw antitrust action on both sides of the Atlantic, including the EU’s browser choice screen in 2010.
At the time, Tom Evslin, Vice President for the AT&T WorldNet Service, said: “Users now will find that everything they need to sign up for AT&T WorldNet Service is pre-loaded on their new computers, along with the Microsoft Internet Explorer 3.0 browser.”
Advertisement
Thirty years on, Evslin told The Register the deal came down to price. AT&T had to pay per copy activated, and he reckoned Microsoft pretty much gave its browser away (but wasn’t completely sure – it was, after all, a long time ago). Netscape Navigator was, however, better known at the time, and so was offered as an option.
By January 1997, Microsoft claimed Internet Explorer’s corporate usage had more than tripled since August while Netscape’s share declined. WorldNet was also attracting users by offering straightforward internet access rather than another AOL-style walled garden.
It rapidly became the world’s largest ISP by one contemporary measure, although it eventually wheezed its last in 2010. Internet Explorer proved more influential, dominating the browser market for more than a decade before its grip began to loosen in the late 2000s.
AT&T’s choice was only one factor in Internet Explorer’s rise, but it put the browser before WorldNet’s growing audience just as users were moving beyond the walled gardens of AOL and CompuServe. Microsoft may have been late to the browser game, but it already controlled the operating system on millions of PCs.
Advertisement
Three decades later, the deal remains an early demonstration of the value of being the default. ®
[Evgenij Spitsyn] spotted a KVM build on these very pages some time ago. That inspired their own build, leveraging the versatility of the ESP32-P4 microcontroller.
The concept is straightforward. Named the ESPKVM, the device is designed to hook up to a computer’s HDMI and USB ports. It captures the video output, while presenting itself as a standard keyboard and mouse device. In this way, it allows remote control of the machine over IP. It achieves this feat with the aid of the Toshiba TC358743 HDMI-to-CSI bridge, which is essentially the video capture hardware of the build.
The video output of the machine is streamed in MJPEG or H.264 format. The device is capable of serving up storage from a micro SD card or the onboard flash, as well as handling things like power/reset control and wake-on-LAN. All in all, it’s a very complete package, and full of useful features. Just don’t use it over the public internet yet — [Evgenij] notes it hasn’t been reviewed for potential security holes yet, even though it has some basic authentication features baked in.
Advertisement
If you’ve got an ESP32-P4 ready to go with a TC358743 HDMI bridge, you can actually head over to the ESPKVM website and flash the code right in your browser to get going. Meanwhile, if you found this build interesting, you might like to scope out the one that inspired it. If you’re cooking up similar utility hacks, be sure to notify the Hackaday tipsline.
Survey finds more workloads run off-site than at in-house corporate facilities for the first time
IT is going remote while sucking up more power.
Most corporate IT is now off-premises for the first time, according to Uptime Institute, while the average rack power density has crested above 11 kW for the first time as server fleets are gradually replaced with more powerful hardware.
Advertisement
Uptime’s Global Data Center Survey 2026 reveals how the industry is managing to adapt to challenging circumstances, with the usual evergreen concerns over rising costs plus staffing and skills shortages. The survey polls more than 800 datacenter owners and operators across multiple countries, with more than half (52 percent) in North America and Europe.
Off premise dominates: Every year, Uptime asks its enterprise respondents to estimate what percentage of their IT workloads are run in-house versus in third-party facilities.
For the first time, third-party sites have the larger share, accounting for 46 percent of IT workloads, compared with 44 percent residing in enterprise-owned corporate server farms.
Those figures don’t add up to 100 percent, as some respondents (10 percent) say they are using IT rooms and server cabinets rather than a dedicated facility.
Advertisement
Uptime’s experts estimate that, by 2028, the proportion of workloads in self-owned server halls will remain the same, while those running in third-party sites will expand to 48 percent, eating away at those currently based in IT rooms and server cabinets.
More power: While the headlines have featured AI infrastructure pushing IT infrastructure power density to 120 kW per rack or even higher, the reality is that most datacenters and servers operate at a much lower level than that.
This year, the average of the most typical rack densities now surpasses 11 kW, as a gradual ongoing shift toward higher-powered hardware was compounded by a small number of new high-density facilities with racks above 30 kW.
Without those few high-density facilities skewing the average, it sits at 7.8 kW, just slightly up from 7.5 kW in 2025.
Advertisement
The majority of facilities still do not have any racks of 30 kW or above, Uptime finds, but more respondents (24 percent) now say they have some of these, compared with 19 percent last year. The increase was mostly in the 50-plus kW ultra-high-density range, including some respondents deploying AI and GPU servers into racks configured for above 100 kW.
The trend for rising rack density will continue as organizations upgrade their infrastructure with newer hardware. Updated servers boost both workload capacity and energy performance, but maximizing these benefits means a corresponding rise in overall system power, Uptime states.
Refresh shortened: Some operators are also pursuing more aggressive technology refresh timelines of less than 4 years, the report claims. If true, this would be the reverse of what some hyperscalers such as Microsoft, Google and Meta have been doing in recent years, extending lifecycles out to 6 or 7 years to save on depreciation expenses.
Outages: When it comes to outages, this year’s report shows improvement for the sixth year in a row, with the number of respondents who experienced an outage in the past three years down by three percentage points.
Advertisement
But Uptime warns against complacency, noting that many of the factors behind outages, such as reduced or unstable power availability, local grid reliability, supply chain constraints, and extreme weather, are on the increase.
The flip side is that the costs of any outages that do occur continue to rise. This is because organizations have become more dependent on digital infrastructure, the report says, and so outages may have more financial impact than in the past.
Overall, 71 percent of survey respondents reported that their most damaging outage cost at least $100,000, compared with 57 percent a year ago.
Staffing shortage: Staffing has long been an issue for datacenter operators, and this year the greatest skills gaps reported were in electrical (38 percent of respondents), junior level operations (38 percent), operations management (35 percent) and mechanical roles (34 percent).
Advertisement
However, more than half (53 percent) of operators report difficulties finding qualified candidates for vacant roles, up from 46 percent a year ago.
Finally, Uptime found that financial pressure and resource constraints continue to grow among operators. In fact, the high prices of power, staff, and equipment, particularly for AI-related infrastructure, is the primary issue. Alongside that are escalating concerns over capacity forecasting, power availability and supply chain disruptions. ®
The decision comes just months after the company announced a $75m Galway expansion plan.
Global medical device manufacturer Boston Scientific is planning layoffs at its Irish operations, which employs more than 7,000 workers across facilities in Galway, Cork and Tipperary.
The Massachusetts-headquartered company informed the Department of Enterprise, Tourism and Employment of the upcoming job cuts via a collective redundancy notification on 22 July, RTÉ reported on Wednesday (29 July).
Businesses in Ireland that employ more than 300 people must notify the Government if they plan to cut 30 or more jobs. Globally, the company employs more than 59,000.
Advertisement
The exact extent of the layoffs, however, is unclear. SiliconRepublic.com has requested details from Boston Scientific. The company is still hiring across all three of its locations in Ireland.
Last week, the Boston Scientific approved a global restructuring plan, which it said would lead to “some” headcount reductions. The plan is expected to cost the company between $700m and $800m.
It expects a “substantial portion” of the savings to be reinvested in strategic growth initiatives, according to a regulatory filing on 21 July.
Advertisement
In a statement to RTÉ, a Boston Scientific spokesperson said that the restructuring programme includes “includes enhancing operational and functional capabilities as well as global supply chain network optimisation.
“Ireland has been an important part of Boston Scientific’s global organisation for more than 30 years. We are committed to our long-term future in the country and recognise the significant contributions our employees in Ireland make across our global business.
“In general, we do not break out programme details by business, region or function.”
Net sales at the global medtech grew 7.5pc year-on-year to nearly $5.5bn during the second quarter of 2026.
Advertisement
Don’t miss out on the knowledge you need to succeed. Sign up for the Daily Brief, Silicon Republic’s digest of need-to-know sci-tech news.
Dave, a New AI-Powered Tool for Sellers, Helps Resolve 54% of Their Cases and Efficiently Manages Demand Spikes During AAA Game Launches. Now available in 180 countries, it has reduced the number of tickets submitted to Seller’s Support teams by 30%.
G2A.COM, one of the world’s largest digital marketplaces, today announced the global launch of ‘Dave’, its proprietary autonomous customer support agent, which will be available to sellers operating on the platform, following a highly successful two-month trial period.
Although sellers ensure the high quality of their offerings, and remain autonomous in all their decisions, large-scale surges in customer support demand may occur during blockbuster video game releases such as GTA 6.
Despite their best efforts, sellers may not be able to handle all these support requests within a short period of time. Dave was built specifically to help sellers manage such spikes and now operates 24/7 across 180 countries.
Advertisement
The 💜 of EU tech
The latest rumblings from the EU tech scene, a story from our wise ol’ founder Boris, and some questionable AI art. It’s free, every week, in your inbox. Sign up now!
Dave has streamlined seller’s operations and processes, managing roughly 10,000 conversations globally per week. Unlike traditional chatbots, Dave connects directly to the seller’s transaction history and enables instant verification of payment status and the seller’s policies, allowing issues to be resolved immediately. By relying on verified data rather than solely on generative text, the system avoids the risk of AI hallucinations and errors and effectively manages complex cases involving vouchers, game keys, software licenses, and refunds processed by sellers.
In its first 63 days of full deployment, Dave achieved several key operational milestones:
Advertisement
Ticket Reduction: Dave streamlined the handling of approximately 15,000 tickets, helping to resolve them quickly on behalf of sellers and reducing the overall volume of tickets requiring human support by 27–30%.
High-Precision Routing: Achieved 93.8% routing accuracy, instantly directing cases to the relevant third-party seller or, where they related specifically to the operation of the platform, to G2A.COM Support, in line with marketplace policies.
Strong User Acceptance: Dave AI received a satisfaction score of 4.16 out of 5. Additionally, 74% of users rated the automated process as easy or very easy.
Quality and Trust Metrics: Dave’s responses achieve an average quality score of 91.2%, with only 0.81% receiving negative in-chat feedback. Every response is automatically evaluated across 13 quality and safety dimensions.
“As G2A.COM continues to expand globally, ensuring that sellers can provide the highest possible standard of customer support remains a key priority. Activity across the gaming ecosystem, including major game launches, can trigger sudden spikes in user inquiries within hours, making it inefficient to scale support simply by hiring more agents. Dave was created to deliver immediate, high-quality, multilingual assistance at scale. If a case requires full manual intervention by the seller, the seller and their support team receive a comprehensive analysis and summary of the case so users never have to repeat themselves.
Dave is also another step in building an AI-native organization and environment. AI is no longer just a productivity tool or a competitive advantage. The real advantage comes from embedding AI into the core of how a company operates. Dave is a practical example of that approach, helping us and sellers using platform to scale globally while delivering faster and more trusted user experiences,” said Paweł Wróbel, CGO at G2A.COM.
About G2A.COM
G2A.COM is one of the world’s largest marketplaces for digital entertainment, serving over 35 million users from 180 countries with more than 200 million visits in 2025. As a true gateway to digital entertainment and a rapidly evolving digital ecosystem, G2A.COM gives access to over 125,000 digital offerings, including vouchers, games, DLCs, in-game items, and non-gaming items such as gift cards, subscriptions, software, or e-learning, sold by sellers from all over the world. A compliance-driven organization audited by Deloitte for over a decade, G2A.COM is a leader in online security, recognized by the prestigious American CNP Award alongside industry giants like Microsoft and PayPal. Today, the company is a key gateway for recognized global brands, leveraging its scalable infrastructure to drive the next phase of e-commerce through agentic AI and global M&A – continuously expanding its role as the Gate 2 Adventure in the digital world.
The startup, which sells everything from sneakers to sports cards over livestream, is raising again barely a year after an $11.5bn round, as live commerce catches on in the West.
Whatnot, the livestream-shopping platform where hosts sell sneakers, trading cards, and vinyl to a live audience, is in talks to raise money at a valuation of about $20 billion. The figure would nearly double the $11.5 billion the company was worth as recently as late 2024.
The pace of the markup is the story. A near-doubling in under a year puts Whatnot among the fastest-appreciating consumer startups around, at a time when venture money has flowed overwhelmingly toward artificial intelligence rather than shopping apps.
Whatnot runs live video auctions and sales across categories from fashion to collectables, taking a commission on each transaction, a model that turns online shopping into something closer to entertainment.
The mechanics are half the appeal. A host holds up an item, buyers bid or tap to purchase in real time, and the urgency of a live sale, the countdown, the banter, the scarcity, does work that a static product page never could.
Advertisement
The numbers behind it are real. The company says it handled about $8 billion in livestream sales over the past year across North America and Europe, the kind of volume that justifies, at least to its backers, a valuation usually reserved for software firms.
The investor roster reads like a who’s who. Andreessen Horowitz, Sequoia, Lightspeed, and Google’s CapitalG have all backed Whatnot, the same names that chase the biggest rounds in technology, now betting that live commerce is more than a novelty.
The idea is not new, just newly working in the West. Livestream shopping has been enormous in China for years, where platforms like Taobao turned hosts into salespeople for millions of viewers, and Western investors have long waited for the format to cross over.
Whatnot’s bet is that it finally has. The company has grown by leaning into niche communities, the collectors and resellers for whom a live auction is both a marketplace and a hangout, rather than trying to be a general store.
Advertisement
It has also started buying capability. This month Whatnot acquired Shaped, a startup that builds real-time recommendation systems, a purchase aimed at pointing viewers toward the streams and items most likely to make them spend.
That acquisition hints at where the money would go. A larger raise would fund the recommendation engine, expansion in Europe, and the fight for hosts and buyers against social platforms circling the same behaviour.
Because the competition is coming. TikTok, Instagram, and Amazon have all pushed into live and social shopping, and Whatnot’s independence is both its advantage, a platform built for this alone, and its vulnerability against far larger rivals.
The valuation, for now, is a talk rather than a term sheet. Business Insider frames it as a round under discussion, and startup valuations at this stage can move before they are signed, especially in a market as selective as this one.
Advertisement
Still, the direction is telling. That investors would price a shopping app at $20 billion, in a year when nearly every large cheque has gone to AI, suggests live commerce has graduated from experiment to category.
It also stands out for what it is not. In a funding market fixated on AI models and infrastructure, Whatnot is a reminder that a consumer business with real transactions can still command a software-sized valuation, provided the growth is there.
The risk is the one every marketplace faces. Whatnot’s value rests on keeping hosts and buyers on the same platform, and that loyalty can erode quickly if a bigger rival offers better terms or a larger audience.
For now, the momentum is with the company. A near-$20 billion valuation would confirm that livestream shopping has arrived in the West, and that Whatnot, for the moment, is the name investors are willing to pay up for.
Apple wants to use China-sourced chips for products sold in China during the AI-driven memory shortage, but a letter from US senators hopes to persuade Apple otherwise.
The hype cycle around what is basically a natural evolution of an advanced autocomplete engine has destabilized the global economy and supply chain. The ever-increasing demand for more components to build data centers has led to a global chip and component shortage.
Apple has looked to blacklisted Chinese companies to bolster its supply chain, but, according to a letter viewed byBloomberg, US senators want to urge Apple away from this solution. They worry about various potential calamities like other companies following suit, potential national security issues, and yet another industrial exodus from the United States.
Apple has officially requested to buy from ChangXin Memory Technologies Inc. and has purchased chips, but not used them in products, from Yangtze Memory Technologies Co. The US senators have asked Apple to reconsider using chips from either company and commit to not using them by August 21.
Advertisement
The letter arrived only four days after Micron asked the White House to reject Apple’s plans.
Apple’s latest China problem
While Apple isn’t banned from buying chips from either company, it could face a few repercussions. First, the US Department of Defense could stop buying Apple products, even if the chips are only used in China-sold products, due to the Chinese Military Company Blacklist.
That’s only a drop in the bucket for Apple, but an important relationship that could be burned.
Apple could also face regulatory issues and increased pressure from the Trump administration. The company has already had to walk on eggshells to avoid the worst of the random and illegal global tariffs, and such a move could undo all of that overnight.
Advertisement
A possible Trump card
Alternatively, Apple could be attempting to play right into a developing situation where President Trump hopes to ease trade tensions with China. He is expected to meet with Xi Jinping in September, and offering Apple the blessing to build and sell using Chinese components could be quite the bargaining chip.
Apple has worked hard to keep Trump out of its business
However, the senators aren’t sure this would be a good idea:
“The limitation that CXMT memory would, for the moment, be confined to devices sold in China cannot be expected to hold, because once a part clears qualification for Apple production, extending it worldwide is a single procurement decision away,” they wrote in the letter to Apple. “The precedent Apple would set by addressing the memory shortage this way would shape the memory market and this country’s security for years after supply has returned to normal, and we urge you to weigh that carefully.”
Advertisement
Perhaps the senators could regulate the industry and attempt to stem the flood created by the far-fetched promises made by grifting billionaires. The memory shortage is tied to a bubble that is sure to deflate or pop at any time, which could be disastrous for the global economy.
When the AI boom ends, Apple sourcing chips for devices sold in China will be the least of our worries. We can only hope Senators Banks, Schumer, and the others will stop chasing distractions and target the root of the problem.
You must be logged in to post a comment Login