Within the narrative of the American defense contractor who disclosed hacking instruments to Russia

Within the narrative of the American defense contractor who disclosed hacking instruments to Russia

A seasoned cybersecurity leader who prosecutors claimed “betrayed” the United States is set to serve a minimum of seven years in prison after admitting guilt to the theft and sale of hacking and surveillance tools to a Russian company.

Peter Williams, a former executive at U.S. defense contractor L3Harris, received a sentence of 87 months on Tuesday for disclosing trade secrets from his former employer in exchange for $1.3 million in cryptocurrency between 2022 and 2025. He sold these exploits to Operation Zero, described by the U.S. government as “one of the most notorious exploit brokers globally.” 

The conviction of Williams is part of one of the most notable breaches of sensitive, Western-developed hacking tools in recent times. Even with the conclusion of the case, many questions remain unanswered.

Williams, a 39-year-old Australian national living in Washington, D.C., was the general manager of Trenchant, the segment of L3Harris responsible for creating hacking and surveillance instruments for the U.S. government and its key intelligence partners globally. Prosecutors suggest Williams exploited his “full access” to the company’s secure networks to transfer hacking tools to a portable hard drive, and subsequently to his personal computer. He communicated with Operation Zero using an alias, making it uncertain whether they knew his true identity.

Trenchant comprises a team of hackers and bug hunters who deeply analyze popular software produced by companies like Google and Apple, pinpointing vulnerabilities in vast lines of code, then inventing methods to convert these vulnerabilities into usable exploits that can reliably compromise those products. These tools are commonly referred to as zero-day exploits due to their exploitation of software defects unknown to the developer, often valued in the millions.

The U.S. Department of Justice claimed that the hacking tools sold by Williams could have enabled their users to “potentially access millions of computers and devices globally.”

For several months, I had been in conversations with sources and covering Williams’ situation before news of his arrest circulated. However, the information I gathered was fragmented and at times inconsistent. I had heard of an arrest, but due to the clandestine nature involved in exploit development, establishing the facts would prove difficult.

Contact Us

If you possess further insights into this case and the supposed leak of Trenchant hacking tools, you can reach Lorenzo Franceschi-Bicchierai securely on Signal at +1 917 257 1382 from a non-work device, or through Telegram, Keybase, and Wire @lorenzofb, or via email.

Initially, when I learned of Williams, I wasn’t certain I had even accurately identified him. At that time, his story circulated as a rumor within the confidential networks of zero-day exploit developers, sellers, and individuals linked to the intelligence sector. 

I heard varying names, perhaps John, or maybe Duggan? Or all the various ways to spell that in English. 

Some of the early rumors were conflicting. It seemed he had taken zero-days from Trenchant and sold them to Russia, or possibly to another adversary of the United States and its allies, like North Korea or China? 

It took several weeks just to verify that indeed there was an individual who matched that description. (It became apparent that Williams’ middle name is John, and Doogie is his hacker alias.)

Over time, as the reporting proceeded, clarity began to emerge. 

The Russian connection

As I initially disclosed in October, Trenchant terminated an employee after Williams, who was still leading Trenchant at that point, accused him of stealing and leaking Chrome zero-days. This situation grew even more intriguing when the employee informed me that after his dismissal, Apple alerted him that someone had targeted his personal iPhone.

What I uncovered was merely the beginning. I had gathered more from my sources, yet we were still piecing the narrative together. 

Shortly thereafter, prosecutors formally accused a man named Peter Williams of stealing trade secrets, which first appeared in the U.S. public court documentation. The initial court filing confirmed that the purchaser of these secrets was a buyer based in Russia.

However, there was no direct mention of L3Harris or Trenchant, nor was it established that the trade secrets that Williams pilfered were zero-days. Importantly, we still could not definitively verify whether it was the same Peter Williams thought to possess access to highly sensitive exploits as Trenchant’s head, or if this was just an unfortunate case of mistaken identity.

We still weren’t there.

On a hunch, with nothing to lose, we reached out to the Department of Justice to inquire if they would verify that the individual named in the document was indeed Peter Williams, the former head of L3Harris Trenchant. A spokesperson confirmed.

At last, the story became public. A week later, Williams entered a guilty plea. 

When I first learned about his situation, I trusted my sources, yet I remained doubtful. Why would someone like Williams engage in the actions described by the rumors? Yet he did, allegedly for financial gain, which he utilized to purchase a home, jewelry, and luxury watches. 

This marked a significant fall from grace for Williams, once considered a talented and brilliant hacker, and particularly for someone who had previously served at Australia’s foremost foreign intelligence agency and in the nation’s military. 

the L3Harris building in Burlington, CanadaImage Credits:JHVEPhoto / Getty Images

What happened to the stolen exploits?

We still lack specific details regarding which exploits and hacking tools Williams appropriated and distributed. Trenchant projected a loss of $35 million, according to court records. However, Williams’ legal team indicated that the pilfered tools were not designated as classified government secrets.

We can deduce certain insights based on the case’s circumstances. 

Since the Justice Department stated that the stolen tools could be utilized to infiltrate “millions of computers and devices,” it is plausible that they pertain to zero-days in widely-used consumer software such as Android devices, Apple’s iPhones and iPads, and various web browsers.

There is some evidence suggesting this direction. At a hearing last year, prosecutors recited a post from Operation Zero on X, as per independent cybersecurity journalist Kim Zetter, who attended the hearing. 

“Due to high demand on the market, we’re increasing payouts for top-tier mobile exploits,” stated the post, explicitly mentioning Android and iOS. “As always, the end user is a non-NATO country.”

Operation Zero offers significant sums for insights into security vulnerabilities in Android devices and iPhones, messaging applications like Telegram, as well as various software, such as Microsoft Windows, and hardware vendors, including multiple brands of servers and routers. 

Operation Zero alleges cooperation with the Russian government. When Williams sold the exploits to the Russian broker, Putin’s extensive invasion of Ukraine was already active.

On the day Williams was sentenced, the U.S. Treasury announced sanctions against Operation Zero and its founder Sergey Zelenyuk, labeling the organization a national security hazard. This was the first government confirmation that Williams had sold the exploits to Operation Zero. 

In its statement, the Treasury noted that the broker “sold those stolen tools to at least one unauthorized user.” At this juncture, we remain unaware of the identity of this user, who could be a foreign intelligence agency or possibly a ransomware group, as the Treasury also sanctioned Oleg Vyacheslavovich Kucherov, an alleged member of the Trickbot gang, who reportedly worked with Operation Zero.

In a court filing, prosecutors indicated that L3Harris managed to determine that “an unauthorized vendor was selling a component” of one of the stolen trade secrets “by matching company-specific vendor data found on a stolen component.” 

Prosecutors also claimed that Williams “recognized code he authored and sold” to Operation Zero “being used by a South Korean broker,” further implying that both L3Harris and prosecutors are aware of which tools were pilfered and sold to Operation Zero. 

Another question that remains is: Did anyone, whether the U.S. government or L3Harris, notify Apple, Google, or whichever technology company’s products were compromised by the zero-day vulnerabilities, now that the exploits had leaked?

Any organization or developer would want to ascertain that someone could have employed (or could still utilize) a zero-day against their users and clients to patch the flaws as quickly as possible. At this point, the zero-days are of no benefit to L3Harris and its government clients.  

Inquiries to Apple and Google yielded no responses. L3Harris also did not provide a reply. 

Who hacked the scapegoat, and why?

Then there is the enigma of the scapegoat, who was dismissed following Williams’ allegations of theft and leaking of code.

During sentencing, Justice Department prosecutors affirmed that the employee was terminated, stating that Williams “stood idly by while another employee of the company was essentially blamed for [his] own actions.” In defense, Williams’ attorney countered prosecutors by asserting that the former employee “was dismissed for misconduct,” referencing claims of dual-employment and improper management of the company’s intellectual property.

According to a court document filed by Williams’ legal representatives, in the course of the L3Harris internal investigation, the company placed the employee on leave, confiscated his devices, sent them to the U.S., and “offered them to the FBI.” 

In response to requests for comments, an unnamed FBI spokesperson stated that the bureau had nothing further to contribute aside from the Justice Department’s press release. 

After his termination, the employee, who we’re identifying as Jay Gibson, received a warning from Apple stating that his personal iPhone was targeted “with a mercenary spyware attack.”

Apple issues these alerts to users it suspects are targeted by attacks utilizing tools produced by firms like NSO Group or Intellexa.

Who attempted to hack Gibson? He received the alert on March 5, 2025, more than six months after the FBI investigation had commenced. The FBI “regularly interacted with [Williams] in late 2024 through the summer of 2025,” according to a court document.

Considering the nature of the leaked tools, it is reasonable to speculate that the FBI, or possibly even a U.S. intelligence agency, might have targeted Gibson as part of the inquiry into Williams’ leaks. However, we cannot say for certain, and it’s possible that neither the public nor Gibson will ever learn the truth.

Updated to clarify the 22nd paragraph attributing the tools’ lack of classification to Williams’ attorneys.

Samsung unveils innovative display technology that incorporates a privacy screen for applications and notifications.

Samsung unveils innovative display technology that incorporates a privacy screen for applications and notifications.

During Samsung’s Galaxy S26 live presentation on Wednesday, the firm unveiled an innovative display technology that enables users to activate a privacy mode, thereby preventing so-called “shoulder surfing” on an app-by-app basis. This display technology will provide a more intelligent and adaptable alternative to privacy films, which are placed over a smartphone’s screen to restrict viewing unless you’re gazing directly at the device.

As highlighted by Samsung, these conventional privacy films have drawbacks — such as dimmer displays or increased challenges when attempting to show something to another person on your device.

Image Credits:Samsung

Their answer is a novel display technology for mobile devices employing two distinct types of pixels: narrow and wide pixels. This arrangement, termed Black Matrix by Samsung, will constrict the path of light emitted from each pixel to accurately manage the light visible to you when the privacy mode is activated. When deactivated, a widened standard pixel collaborates with the narrow pixels to distribute light in every direction.

samsungImage Credits:Samsung

The Privacy Display function is also adjustable, permitting you to configure certain apps or alerts to initiate privacy mode, while others are displayed in a regular manner. For example, you might set your messaging or banking application to always appear in private mode, while disabling it for other apps where privacy is less critical. This adaptability likewise applies to notifications, effectively obscuring specific alerts when seen from a side perspective.

Samsung claims there’s an additional setting for “maximum privacy protection,” which enhances the privacy effect by reducing the brightness of lighter areas and elevating the darker ones.

Image Credits:Samsung

The latest display will first be featured on the premium Galaxy S26 Ultra smartphone, which additionally supports enhanced and quicker AI alongside a specialized chipset.

The privacy screen was demonstrated live on stage during the event on Wednesday, with creator Miles Franklin (from MilesAboveTech) showcasing the technology in action. During the demo, the feature appeared to function as expected, including for smaller elements on the display such as specific notifications.

Techcrunch event

Boston, MA
|
June 9, 2026

The company had hinted at the technology prior to Wednesday’s event, but had not yet detailed how it would function. It mentioned that the feature would work with certain apps, ensuring user protection when entering PINs, patterns, or passcodes.

The Final Enigma of Antarctica's 'Blood Falls' Has Ultimately Been Resolved

The Final Enigma of Antarctica’s ‘Blood Falls’ Has Ultimately Been Resolved

There exists a section of Antarctica that brings to mind a scene from a David Cronenberg movie. Situated within the arid valleys of McMurdo, this expansive frozen wasteland sporadically expels a crimson plume from the glaringly white Taylor Glacier. Referred to as the Blood Falls, since their identification by geologist Thomas Griffith Taylor in 1911, they have generated scientific fascination for over a century.

Recent studies from 2018 onward have shed light on numerous enigmas, including the origin of their red coloration and the reason they remain liquid at nearly –20 degrees Celsius. Newly released research in the journal Antarctic Science resolves the mystery, clarifying the processes that compel the falls to surface from below ground.

The Science of the Blood Falls

Taylor initially attributed the hue to red microalgae. More than a hundred years later, it is now understood that the redness arises from iron particles encapsulated with other elements such as silicon, calcium, aluminum, and sodium in nanospheres. These are likely derived from ancient bacteria imprisoned underground: When the iron comes into contact with air, it oxidizes, leading to the characteristic rust color.

The fluid is a hypersaline brine that formed around two million years ago as the Antarctic Ocean receded. Its elevated salinity prevents it from freezing, enabling it to flow out intermittently.

The Recent Revelation

The ultimate question was what physically triggered the fluid to burst forth. The answer surfaced from examining GPS data, thermal sensors, and high-resolution photographs from 2018. The Blood Falls are a result of pressure fluctuations in the brine beneath the glacier.

As Taylor Glacier shifts, its ice mass compresses the subglacial channels, generating substantial pressure. When the stress reaches a critical point, the ice fractures: Pressurized brine seeps through fissures and erupts. This expulsion functions as a hydraulic brake, temporarily hindering the glacier’s movement. This finding seems to clarify the enigmas surrounding the Blood Falls, although the effects of global warming on the system remain ambiguous.

This narrative originally appeared on WIRED Italia and has been translated from Italian.

YouTube enhances its $7.99/month Lite subscription by adding offline downloads and background playback.

YouTube enhances its $7.99/month Lite subscription by adding offline downloads and background playback.

YouTube is broadening its budget-friendly Premium Lite subscription service at $7.99 a month, introducing new features such as downloading videos for offline viewing and allowing videos to play in the background while the screen is off or when using other applications. These features were previously exclusive to users on the full subscription plan priced at $13.99 a month.

The company indicated that these enhancements stemmed from user feedback, as participants in its pilot initiative expressed a desire for these functionalities to enhance the appeal of the subscription.

Launched in March, YouTube Premium Lite provided a cost-effective subscription option aimed at eliminating ads from “most” videos on the platform, encompassing popular categories such as gaming, fashion, beauty, cooking, news, and others. However, advertisements will still appear on music content and music videos. Moreover, Lite plan subscribers will not have access to the ad-free YouTube Music application.

With these new features, ad-free music content will now be the sole incentive to upgrade to the full Premium subscription. It is also expected to enhance the Lite tier’s attractiveness to customers who were reluctant to pay solely for ad-free content but were seeking additional benefits.

The Lite subscription option was initially rolled out in Thailand, Germany, and Australia, before its arrival in the U.S. last year. It is currently available in various other global markets, including Canada, Brazil, the U.K., India, Mexico, and other regions in Europe and Asia.

YouTube’s subscription model has been consistently expanding. When combined with advertising, YouTube’s total revenue hit $60 billion in 2025, based on information provided by its parent company Alphabet earlier this month during its Q4 earnings announcement.

The company also disclosed that YouTube’s advertising revenue grew by 9%, reaching $11.38 billion in the fourth quarter. At the same time, the “subscriptions, platforms and devices” segment saw a revenue growth of 17% to $13.6 billion in Q4, which the company attributed to robust expansion in YouTube subscriptions, particularly for YouTube Music and YouTube Premium.

Techcrunch event

Boston, MA
|
June 9, 2026

Alphabet reported over 125 million YouTube Music and YouTube Premium users globally as of March 2025. While the company did not provide an updated statistic in its recent Q4 earnings report, it noted that the total now surpasses 325 million paid subscriptions across consumer services, including YouTube Premium and others like Google One.

Music creator ProducerAI partners with Google Labs

Music creator ProducerAI partners with Google Labs

On Tuesday, Google revealed that the generative AI music platform ProducerAI will join Google Labs.

Supported by The Chainsmokers, the ProducerAI service enables users to compose music using natural language prompts — such as “create a lofi beat” – to generate tracks. It employs Google DeepMind’s Lyria 3 music-generation model, capable of transforming text and even images into audio outputs.

Last week, Google announced plans to integrate Lyria 3 features into the main Gemini app, but ProducerAI allows users to interact with the AI model more as if it were a “collaboration partner,” according to Elias Roman, Senior Director of Product Management at Google Labs.

“ProducerAI has opened up new ways for me to create,” Roman expressed in a blog entry. “I’ve played with genre fusions, crafted personal birthday songs for friends and family, and produced custom workout playlists for myself and others.”

[embedded content]

Google also disclosed that three-time Grammy winner Wyclef Jean utilized the Lyria 3 model and Google’s Music AI Sandbox for his latest track “Back From Abu Dhabi.”

“This isn’t merely a machine where you push a button a hundred times and you’re finished. It involves a thoughtful curation process where you sift through options and say, ‘I believe that’s something we can work with,’” stated Jeff Chang, Director of Product Management at Google DeepMind, in a video released by the company.

Jean reflected on his curiosity about how a flute might sound in an already recorded track, and his ability to swiftly incorporate a flute sound using Google’s tools.

Techcrunch event

Boston, MA
|
June 9, 2026

“What I want everyone to grasp […] is that we now live in an era where human creativity must prevail,” Jean mentioned in the video. “One advantage you have over AI: a soul. One advantage AI has over you: infinite information.”

AI in the music industry

Some musicians have vigorously opposed the integration of AI tools in music creation, as it’s almost certain that generative AI tools were trained on copyrighted materials from artists without permission. Numerous musicians, including big names like Billie Eilish, Katy Perry, and Jon Bon Jovi, endorsed an open letter in 2024 urging tech companies not to compromise human creativity through AI music generation tools.

A group of music publishers has also recently filed a lawsuit against the AI firm Anthropic for $3 billion, alleging that the company illegally downloaded over 20,000 copyrighted songs, inclusive of sheet music, lyrics, and compositions. (Anthropic had previously been ordered by the court to offer a $1.5 billion settlement to authors whose books were used without authorization for AI training.)

Conversely, certain artists have welcomed this technology’s potential to enhance audio quality rather than seek creative support.

Paul McCartney utilized AI-based noise reduction technology — similar to what Zoom or FaceTime use to eliminate unwanted background sounds during video calls — to restore an old, low-quality demo by John Lennon. The resultant “new” Beatles song, “Now and Then,” garnered a Grammy in 2025.

In the meantime, AI music generation tools like Suno have produced synthetic music that is sufficiently realistic to dominate charts on Spotify and Billboard. Telisha Jones, a 31-year-old from Mississippi, transformed her (allegedly organic) poetry into the viral R&B track “How Was I Supposed To Know” using Suno, landing a record deal with Hallwood Media worth an estimated $3 million.

Legal clarity surrounding the use of copyrighted works as training data remains ambiguous — one federal judge, William Alsup, determined last year that training on copyrighted data is permissible, while piracy is not.

Marquis files a lawsuit against firewall provider SonicWall, claiming that deficiencies in its firewall backup contributed to a ransomware attack.

Marquis files a lawsuit against firewall provider SonicWall, claiming that deficiencies in its firewall backup contributed to a ransomware attack.

The fintech powerhouse Marquis has initiated legal proceedings against its firewall service provider SonicWall, asserting that a previous breach facilitated hackers in stealing sensitive data related to customer firewalls, culminating in a ransomware incident affecting Marquis’ infrastructure.

The legal action, lodged on Monday in the U.S. District Court for the Eastern District of Texas, calls for a jury trial. It asserts that the 2025 compromise at SonicWall “revealed vital security information for Marquis and all clients utilizing SonicWall’s firewall cloud backup service.”

Satin Mirchandani, Marquis’ CEO, informed TechCrunch in a statement that SonicWall purportedly neglected to safeguard its backup service, leading to “significant damage to reputation, operations, and finances” for the company.

The announcement of the lawsuit follows weeks after TechCrunch disclosed that Marquis intended to pursue restitution from SonicWall. The fintech giant, based in Plano, Texas, had communicated to its clientele that it held SonicWall accountable for the breach that allowed hackers to exfiltrate sensitive data about customer firewall configuration files, including its own.

“SonicWall permitted a threat actor to acquire the means to circumvent that protective barrier and gain access to Marquis’s internal network, precisely what SonicWall’s firewall was designed to prevent,” the complaint states.

Firewalls are designed to block unauthorized access to corporate networks, but Marquis contends that the ransomware incident was caused by hackers utilizing information stolen from SonicWall regarding customer firewall configurations, including emergency access codes (referred to as scratch codes) that granted entry to Marquis’ internal network.

Marquis, which enables numerous banks and credit unions to analyze their customers’ data, indicated that the hackers acquired “personally identifiable information related to customers of several of Marquis’s financial institution clients” during the cyberattack.

The compromised information encompasses customer names, birthdates, residential addresses, and financial data, including bank account, debit, and credit card numbers, as well as Social Security numbers of the customers.

A representative from SonicWall has not yet responded to inquiries regarding the lawsuit.

SonicWall first acknowledged a breach in its systems in mid-September, claiming that less than 5% of its customer firewall configuration backup files had been exfiltrated from its storage servers, which are hosted on Amazon’s cloud and managed by SonicWall. In October, the firewall manufacturer admitted that in fact every customer had their firewall backup files compromised in the breach.

In December 2025, Marquis began alerting those affected that its networks had been infiltrated the previous August. SonicWall has not disclosed when hackers first gained entry to its systems.

The exact cause of the breach at SonicWall remains uncertain. Marquis asserts in its complaint that SonicWall altered the code in one of its APIs months prior, in February 2025, which “introduced a vulnerability that could be exploited by threat actors.” Marquis claims this flaw enabled hackers to access customer firewall configuration backup files “without appropriate authentication” by predicting firewall serial numbers.

“Although we swiftly managed to secure our network and client data, our investigation indicated that our exposure to threat actors resulted from SonicWall’s network breach and its failure to inform us that our firewall defenses were possibly compromised,” stated Mirchandani, the CEO of Marquis, in a statement provided to TechCrunch.

Mirchandani mentioned to TechCrunch that SonicWall has yet to supply any non-public insights regarding the root cause of its breach. 

“We anticipate learning more through the legal proceedings,” remarked Mirchandani.

Marquis has not disclosed the number of individuals impacted by its data breach. A record with Texas’ attorney general shows that at least 400,000 individuals across the U.S. are confirmed to be affected by the fintech giant’s breach. 

The total number of affected individuals is expected to grow as additional data breach notifications are submitted to various U.S. attorneys general.

Ukrainian startups continue to innovate

Ukrainian startups continue to innovate

When the edtech firm Preply reached unicorn status earlier this year, its Kyiv team marked the occasion with cake — as is customary, even amid a conflict.

Over the four years following Russia’s large-scale invasion of Ukraine, Ukrainian startups have not only managed to endure: they continue to innovate and expand. Preply, for example, plans to use funds from its latest investment round to recruit around 100 engineers globally — including in Ukraine, where one-third of its engineering workforce is located.

Preply stands as one example among numerous cases. The defense technology sector has drawn significant attention, particularly for the rapid pace at which new innovations are deployed in combat. Yet, the same engineering expertise and creativity are proving to be vital across various fields also aiding in fortifying Ukraine.

Facing a significantly larger aggressor, Ukraine adopts a serious stance towards attempts to evade conscription. Nonetheless, startups may receive special status that protects essential personnel from being drafted if they are recognized as bolstering the nation — and Aspichi is among them, as its founder, Victor Samoilenko, explained to TechCrunch.

Initially established in the U.S. in 2021, the company underwent a total transformation with the onset of the war. It is now primarily known for Luminify, a mixed-reality platform dedicated to mental health care that assists Ukrainians in dealing with trauma from the conflict. The startup works closely with military units as well as several clinics that offer mental health services to the community.

This includes soldiers and veterans, grieving families, and millions who have relocated to western Ukraine or overseas. For those who have always resided in Kyiv, the psychological impact is equally severe.

“Everyone is enduring hardship,” Samoilenko remarked. “My daughter has celebrated several New Year’s and Christmas underground; thus, the emotional fallout is immense.”

Techcrunch event

Boston, MA
|
June 9, 2026

During a winter characterized by almost daily strikes on the power grid, Ukraine’s capital has not been spared. Natali Trubnikova, CMO at the Kyiv-based IT consulting firm Gart Solutions, noted that residents have adapted to outages with high-capacity power banks, gasoline cookers, and diesel generators. However, increasing costs necessitate that these resources are used judiciously, often resulting in indoor temperatures being kept at a bare minimum.

These severe conditions have transformed offices into shelters from the cold.

“Our office is equipped with various generators to ensure we have electricity, internet, and warmth, and it remains open 24/7 for any Ukrainian team member to utilize at any time,” Preply CEO Kirill Bigai shared with TechCrunch last month.

For smaller enterprises, coworking spaces have also evolved into sanctuaries. LIFT99 Kyiv Hub, a six-year-old facility that was damaged by a Russian attack last August, has experienced a surge in membership since reopening two months ago, according to sales and partnership manager Lada Samarska’s LinkedIn update.

Despite missile assaults, Kyiv remains distant enough from the frontlines to retain its status as Ukraine’s primary startup hub, though it is not alone in this regard. Lviv, the largest city in western Ukraine, has attracted a considerable number of displaced Ukrainians, including tech professionals. LEM Station, a revamped tram depot transformed into a creative hub, symbolizes its expanding ecosystem, which has also gained from its closeness to the Polish border.

Although the number of international visitors has decreased, Lviv is still accessible by train, and the IT Arena 2025 tech conference welcomed 6,450 participants from over 40 countries. While a segment focused on defense tech, the event also highlighted support for a broader spectrum of startups. Despite ongoing hostilities, VC firms remain active in the nation, including 1991, Flyer One Ventures, and SMRK.

Regardless of the challenges, Ukrainians continue to attend tech conferences abroad. At Techarena in Stockholm earlier this month, members of a Ukrainian delegation expressed fatigue and anticipation for the spring. Yet before parting ways, they began to enumerate Ukraine’s unicorns — a fitting response, even in the face of war.

Meta secures up to $100B AMD chip agreement as it pursues ‘personal superintelligence’

Meta secures up to $100B AMD chip agreement as it pursues ‘personal superintelligence’

Meta intends to acquire potentially as much as $100 billion in AMD chips, sufficient to power nearly six gigawatts of data center energy demand, the companies revealed on Tuesday.

As part of the long-term agreement, AMD has granted Meta a performance-linked warrant for up to 160 million shares of AMD common stock — equivalent to around 10% of the company — at a price of $0.01 each, designed to vest in accordance with specified milestones. The complete stock award is contingent upon AMD’s share value, which must reach $600 for Meta to obtain its final portion, according to The Wall Street Journal. AMD’s stock price closed at $196.60 on Monday.

In line with the agreement, Meta will acquire AMD’s MI540 series GPUs and its most recent generation of CPUs. CPUs are increasingly recognized as a key component of the AI inference compute stack due to their efficiency, scalability, and the fact that they do not solely bind companies to Nvidia.

“The CPU market is truly booming,” AMD CEO Lisa Su remarked on Tuesday morning during a briefing for investors. “Demand is substantial. It continues to expand, driven by the deployments of AI infrastructure as inferencing intensifies and agentic AI develops, and our portfolio is exceptionally well-positioned.”

AMD has been progressively making headway as AI companies seek to lessen their dependency on Nvidia, the long-time frontrunner in AI chips that has commanded a premium for its position. Last October, AMD and OpenAI forged a similar agreement that involved equity in exchange for a chip purchase contract.

Meta CEO Mark Zuckerberg stated that the collaboration with AMD is “a crucial advancement” as it diversifies its computing capabilities and progresses toward “personal superintelligence.” Zuckerberg has characterized personal superintelligence as AI systems intended to profoundly understand and enhance individuals’ daily lives.

Meta has committed to investing a minimum of $600 billion in U.S. data centers and AI infrastructure over the upcoming years, including an estimated capital expenditure of $135 billion in 2026. Meta recently announced plans for a $10 billion gas-powered data center campus in Indiana aimed at providing 1 gigawatt of compute power.

Techcrunch event

Boston, MA
|
June 9, 2026

The AMD collaboration follows a recent agreement in which Meta expanded its data centers with millions of Nvidia’s latest CPUs and GPUs. The company formerly known as Facebook is also developing its in-house chips but has reportedly encountered delays.

This article has been revised with additional information from AMD CEO Lisa Su.

Oura introduces a unique AI model centered on women's health.

Oura introduces a unique AI model centered on women’s health.

On Tuesday, Oura declared the debut of its inaugural proprietary AI model, aimed at equipping its AI chatbot, Oura Advisor, with the capability to provide tailored insights related to women’s health. The firm claims the model accommodates inquiries across the entire reproductive health spectrum, from initial menstrual cycles to menopause. 

This innovative model is being introduced in Oura Labs, the company’s voluntary, experimental feature hub located within the Oura app.

Oura states that the new model is based on established medical guidelines, research, and knowledge resources evaluated by its internal team of board-certified clinicians and experts in women’s health. It also incorporates biometric signals and long-term trends to offer customized guidance.

As individuals increasingly rely on AI chatbots for health advice, addressing issues from cycle changes to perimenopause symptoms, Oura emphasizes the necessity for models tailored specifically for women.

Image Credits:Oura

Ricky Bloomfield, MD, Oura’s chief medical officer, stated in a press release, “This customized model marks a significant shift in how we judiciously utilize AI in healthcare to fulfill the needs of our members. Women’s health is intricate—and frequently neglected—so relying on generic systems is insufficient. By creating a model specifically for women and anchoring it in trusted clinical research and real-world biometric data, we are establishing the benchmark for responsible intelligence to be constructed and expanded into more areas of healthcare, merging rigorous science with the lived, longitudinal data that renders Oura extraordinarily effective.”

The rollout of the new women’s health AI model occurs as Oura’s chief commercial officer, Dorothy Kilroy, informed TechCrunch last October that the company’s rapidly expanding user demographic consists not of gym enthusiasts, but women in their early twenties.

When a user inquires about women’s health via Oura Advisor, the chatbot activates the new model to draw on its research and knowledge resources while simultaneously assessing the user’s pertinent biometric signals, which encompass sleep, activity, cycle and pregnancy data, stress, and more.

Techcrunch event

Boston, MA
|
June 9, 2026

The newly developed model is deliberately created to be non-dismissive, comforting, and emotionally supportive, as noted by the company. However, it is not intended to replace a physician, as users should refrain from relying on the chatbot for diagnoses or treatment plans.

Oura asserts that the model is entirely hosted on infrastructure controlled by Oura, ensuring that conversations are never disclosed or sold.

Users interested in utilizing the new model can opt into Oura Labs by accessing the dropdown menu in the upper left corner of the Oura app.

Last 4 days to save as much as $680 on your TechCrunch Disrupt 2026 ticket

Last 4 days to save as much as $680 on your TechCrunch Disrupt 2026 ticket

You don’t come to TechCrunch Disrupt to simply observe. You attend to enhance your influence. If 2026 is for building, fundraising, hiring, or scaling, this is the moment where energy multiplies.

With only 4 days remaining until Super Early Bird pricing concludes on February 27 at 11:59 p.m. PT, this is your chance to save as much as $680 on your ticket and reserve your position at the heart of the tech community. Sign up here to secure your savings.

TechCrunch Disrupt 2026 4 days left

Benefits of Attending TechCrunch Disrupt

From October 13 to 15 at Moscone West, over 10,000 founders, operators, and VCs will gather for three days of impactful discussions and opportunity-making. Disrupt is far more than just content. It represents access.

You will receive:

  • Practical advice from professionals currently navigating today’s landscape.
  • Personal interactions with investors ready to make investments.
  • Insight into nascent startups ahead of the broader market.
  • Tailored networking opportunities geared towards tangible results.

Last year, over 20,000 specifically arranged meetings occurred. In 2026, enhanced networking tools will further refine those connections. A single conversation can alter your path. At Disrupt, that’s the objective.

Image Credits:TechCrunch

Gain insights from the innovators and VCs who influenced the landscape

Disrupt has consistently been a platform for founders and investors who shape the future. These speakers are candid, strategic, and often straightforward. Past presenters include leaders from ground-breaking startups and esteemed venture firms, such as:

Image Credits:Kimberly White / Getty Images

In 2025, Disrupt featured over 200 onstage discussions with more than 250 leaders impacting AI, venture capital, hardware, growth strategies, and beyond. Anticipate that same level of knowledge in 2026. Stay updated by checking the event page as the agenda becomes available.

Techcrunch event

Boston, MA
|
June 9, 2026

Where startups ascend to the legendary pitch stage

Startup Battlefield is back with 200 pre-Series A companies vying for $100,000 in equity-free funding, worldwide visibility, and direct access to investors. Alumni feature brands like Discord, Cloudflare, and Trello.

TechCrunch Disrupt 2025 Startup Battlefield
Image Credits:TechCrunch

If you aim to discover who and what is on the horizon, and listen directly to leading VCs about scaling a viable startup, the Disrupt stage is where it occurs.

Uncover the next significant innovation in the tech landscape

Over 300 startup exhibitors will display entrepreneurial innovations throughout the venue, particularly in the Expo Hall, where deal flow meets discovery. You are not merely tracking trends; you are witnessing them before they escalate.

“Disrupt Week” broadens the chances to network

From October 11 to 17, TechCrunch Disrupt Side Events will be held across the Bay Area, featuring breakfasts, cocktail gatherings, panel discussions, and founder meetups that extend connections beyond the primary event. The main event is impactful, but the surrounding ecosystem enhances its potency.

Last 4 days to grab your pass at the best rates

Super Early Bird pricing concludes on February 27 at 11:59 p.m. PT. If you intend to be in the spaces where capital flows, companies expand, and ideas evolve into industries, now is the critical moment to secure your pass.

Sign up now:

TechCrunch Disrupt 2026 exhibitor
Image Credits:Slava Blazer Photography