Snap and Qualcomm Ink Multi-Year Deal to Power Next-Gen AR Smart Glasses
Snap and Qualcomm have announced a multi-year strategic partnership to power future generations of Snap’s augmented reality (AR) smart glasses with Snapdragon chips. The agreement will be led by Specs Inc., Snap’s hardware subsidiary formed in January, which will equip upcoming devices with Qualcomm’s Snapdragon XR system-on-a-chip platforms. This collaboration builds on over five years of work between the two companies, with Snapdragon technology already utilized in several previous developer-focused Spectacles.
The partnership is centered on delivering on-device AI, advanced graphics processing, and multiuser digital experiences. By running computation locally rather than relying on cloud connectivity, the glasses are designed to offer faster and more private interactions. Snap CEO Evan Spiegel stated that the alliance provides a "strong foundation" for the future of the Specs brand, pushing the boundaries of technology for both developers and consumers.
Snap is currently targeting a consumer launch for its AR glasses in 2026. This upcoming device is expected to be smaller and lighter than the fifth-generation developer-focused Spectacles released in 2024. Key features include see-through AR optics and a custom operating system, Snap OS, which is built specifically for hand and voice interactions.
The deal arrives as the smart glasses market transitions from a "niche experiment" to a mainstream category, with industry forecasts projecting 13.4 million shipments in 2026. Snap is positioning itself to compete directly with Meta Platforms, which has gained early traction with its Ray-Ban AI glasses and is targeting its own consumer AR release in 2027. While financial terms of the Snap-Qualcomm deal remain undisclosed, Snap’s shares rose following the announcement as the company seeks to secure a leading position in the accelerating hardware market.
Global Oceans Hit Near-Record Highs as Looming "Super El Niño" and Conflict Threaten Global Stability
GENEVA — Global sea surface temperatures reached their second-highest levels on record for March, signaling a rapid transition toward a "Super El Niño" event that experts warn could shatter temperature records through 2027. Data from the European Union's Copernicus Climate Change Service shows extra-polar ocean temperatures averaged 20.97°C last month, trailing only the extreme peaks seen in March 2024.
A World of Extreme Heat
The spike in ocean heat coincides with atmospheric records across the globe. March 2026 was the fourth-hottest March ever recorded globally, with surface air temperatures sitting 1.48°C above pre-industrial levels. The United States experienced its hottest March in 132 years, fueled by a prolonged heatwave in the Southwest that saw temperatures soar more than 10°F above normal.
Further north, the news is equally grim. Arctic sea ice reached its annual maximum on March 15 at just 14.29 million square kilometers—statistically tying with 2025 for the lowest peak in nearly half a century of satellite monitoring.
"Super El Niño" on the Horizon
Climate centers are now forecasting a shift from neutral conditions to a powerful El Niño in the second half of 2026. Forecasts suggest this event could exceed 2°C in key monitoring regions, potentially rivaling the record-breaking episodes of 2015 and 2016.
"A strong El Niño could plausibly push global temperatures to new record levels in late 2026 and into 2027," warned Victor Gensini, a meteorology professor at Northern Illinois University. The anticipated impacts include:
- Increased wildfire risks and a hotter-than-average summer for the Western U.S..
- Elevated tropical storm threats for Hawaii.
- A weakened Indian monsoon, which could lead to widespread drought across the subcontinent.
Compound Crisis: Climate and Conflict
The environmental threat is being compounded by geopolitical instability. The ongoing conflict in Iran, which saw the closure of the Strait of Hormuz following strikes in late February, has blocked roughly 30% of the global seaborne fertilizer trade.
With nearly half of global urea exports and 30% of ammonia exports stalled, the American Farm Bureau Federation warns that the combination of the "Super El Niño" and the fertilizer shortage is creating a perfect storm for global food security. As climate models point to a definitive El Niño by summer, the dual impact of extreme weather and disrupted supply chains may pose the greatest challenge to the global agricultural system in decades.
Google Expands Gmail End-to-End Encryption to Android and iOS for Enterprise Users
MOUNTAIN VIEW — Google announced on Thursday that it has expanded Gmail’s end-to-end encryption (E2EE) to mobile devices, allowing enterprise users to natively compose and read encrypted messages within the Gmail app for the first time. This update brings the same level of security previously available on the web to Android and iOS platforms, enabling seamless, secure communication across different devices and email providers.
Native Integration and User Experience
The new mobile functionality is designed to be user-friendly, requiring no extra apps, mail portals, or manual certificate exchanges. To secure a message, users simply tap a lock icon in the compose window and select "additional encryption". For Gmail app users, these encrypted messages appear as standard email threads, while recipients using other email providers or platforms can securely read and reply through a web browser.
Enhanced Privacy Through Client-Side Encryption
The system is powered by client-side encryption (CSE), a technical control that allows organizations to manage their own encryption keys outside of Google's servers. This ensures that messages and attachments are encrypted on the device before they ever reach Google, providing a "private room" for digital correspondence. This level of security is specifically intended to help organizations meet strict regulatory requirements, including data sovereignty, HIPAA compliance, and export controls.
Availability and Implementation
The feature is currently restricted to Google Workspace Enterprise Plus customers who have the Assured Controls or Assured Controls Plus add-on. Before users can access E2EE on their mobile devices, administrators must manually enable the Android and iOS clients within the client-side encryption admin interface.
The Final Step in a Long Rollout
This announcement marks the completion of a multi-year effort to bring robust encryption to Gmail:
- December 2022: Google began testing client-side encryption for Gmail on the web.
- April 2025: A simplified E2EE model launched in beta for internal enterprise messages.
- October 2025: Capabilities expanded to allow encrypted emails to be sent to any recipient on any platform.
- Present: Mobile functionality officially launches for enterprise clients.
While the feature is now fully available for high-tier enterprise users, Google has not yet provided a timeline for expanding end-to-end encryption to individual Gmail accounts or other Workspace tiers.
Anthropic Secures Multiyear Computing Deal with CoreWeave to Fuel Claude AI Growth
NEW YORK — Anthropic, the developer of the Claude AI model, has signed a multiyear agreement to lease data center capacity from CoreWeave, a move designed to secure the massive infrastructure required to meet "surging demand" for its services. The deal provides Anthropic with access to Nvidia chip architectures across CoreWeave’s U.S.-based data centers.
While financial terms were not disclosed, CoreWeave’s stock saw gains following the announcement. The specialized GPU cloud provider has rapidly expanded its footprint, recently signing a $21 billion agreement with Meta Platforms and maintaining a customer roster that includes OpenAI and Perplexity. CoreWeave, which went public in March 2025, is currently targeting more than five gigawatts of AI computing capacity by 2030.
Rapid Financial and Infrastructure Expansion
The partnership comes at a period of "unprecedented growth" for Anthropic. The company recently revealed that its annualized revenue run rate has surpassed $30 billion, more than tripling from roughly $9 billion at the end of 2025. Its base of enterprise customers has also doubled in recent months to over 1,000 clients.
To support this trajectory, Anthropic is aggressively diversifying its hardware and cloud commitments:
- Google and Broadcom Deal: On April 6, the company announced its largest compute commitment to date—a deal for 3.5 gigawatts of next-generation TPU capacity expected to come online in 2027.
- Multi-Cloud Strategy: Anthropic continues to run Claude on a mix of Google TPUs, Amazon Trainium chips, and Nvidia GPUs, utilizing AWS, Google Cloud, and Microsoft Azure.
- Primary Partnership: Despite the new agreement with CoreWeave, Anthropic emphasized that Amazon remains its primary cloud provider and training partner.
Strategic Positioning Ahead of Potential IPO
The deal with CoreWeave represents a strategic effort to lock in computing resources from multiple providers, reducing dependence on any single source as the global race for AI infrastructure intensifies. There are also reports that Anthropic is exploring the design of its own custom AI chips with Broadcom to optimize performance and reduce long-term costs.
These infrastructure milestones arrive as Anthropic prepares for a potential initial public offering (IPO). Early discussions with major financial institutions, including Goldman Sachs and JPMorgan, suggest a listing could occur as early as October 2026. Since its founding in 2021, Anthropic has raised nearly $64 billion to compete in the frontier AI model market.
Samsung Ramps Up Galaxy S26 Production to 3 Million Units as "Privacy Display" Fuels Ultra Demand
SEOUL — Samsung Electronics is significantly increasing production of its Galaxy S26 flagship series to 3 million units for the month of April, a 600,000-unit boost over original projections. The move comes after consumer demand for the new smartphones far exceeded initial forecasts throughout March.
Ultra Model Dominates While Plus Variant Slumps
The surge in production is being driven primarily by the Galaxy S26 Ultra, which accounts for an unprecedented 70% to 80% of total series demand. Samsung is increasing Ultra production by 200,000 units to reach 1.5 million for April. The standard Galaxy S26 is also seeing a major lift, with production jumping by 500,000 units to 1.3 million.
In contrast, the mid-tier Galaxy S26+ is seeing its output slashed by 100,000 units to approximately 200,000, reflecting a shift in consumer appetite away from the Plus variant.
Innovative Privacy Hardware Driving Sales
Industry analysts attribute the Ultra’s record-breaking popularity to its new "Privacy Display" technology. This hardware-level feature uses two sets of specialized pixels—wide-angle and narrow-beam—to control light emission from the OLED panel. This allows the phone to reduce visibility from side angles, maintaining privacy for sensitive tasks like banking or messaging without requiring a physical screen filter.
The device's technical prowess was recently recognized at Mobile World Congress 2026, where the Galaxy S26 Ultra won "Best in Show" for its integration of advanced hardware and context-aware AI features.
Record Profits and Market Momentum
The production ramp-up follows a historic pre-order period, with Samsung securing 1.35 million pre-orders in South Korea within just seven days, outpacing the previous Galaxy S25 launch. Global momentum is equally strong, with pre-orders in Thailand growing 1.4 times year-over-year.
This mobile success is contributing to a massive financial windfall for the company. Samsung recently posted a record 57.2 trillion won Q1 profit, bolstered by a 50% year-on-year rise in revenue to approximately 116 trillion won. While the AI chip boom in Samsung's semiconductor division is the primary earnings driver, the stronger-than-expected performance of the Galaxy S26 series has further solidified the company’s market position.
Samsung Expands One UI 8.5 Beta: AirDrop Support Hits Older Galaxy Devices and Mid-Range Models
SEOUL — Samsung has officially opened its One UI 8.5 beta program to a significantly wider range of devices, including the Galaxy S23 series and, for the first time, a mid-range A-series smartphone. The expansion, confirmed on April 8, brings the latest software features to select markets including the United States, South Korea, India, and the United Kingdom.
Cross-Platform Connectivity: AirDrop Integration
The standout feature of the One UI 8.5 update is the integration of AirDrop compatibility within Samsung's Quick Share. This cross-platform capability, which originally debuted on the Galaxy S26 series in March, allows Galaxy users to perform direct file transfers with Apple iPhones, iPads, and Macs without the need for third-party applications.
Beyond the S23 series, the beta's AirDrop support is being extended to:
- Flagship Series: Galaxy S25 and S24 series.
- Foldables: Galaxy Z Fold7, Z Flip7, Z Fold6, and Z Flip6.
- Legacy Foldables: Galaxy Z Fold5 and Z Flip5 (currently limited to the U.S. and South Korea).
Samsung joins Google—which recently added similar functionality to the Pixel 10—in bridging the long-standing file-sharing gap between Android and Apple ecosystems.
Performance Trade-offs and A-Series Debut
The inclusion of the Galaxy A36 5G marks a milestone for Samsung, as it is the first budget-friendly A-series device to be included in a One UI 8.5 beta program.
However, the new sharing capabilities may come with a technical cost. Early reports suggest that enabling AirDrop compatibility within Quick Share can result in slower and less consistent file transfers to Windows PCs. Some beta testers have recommended toggling the feature off when not actively communicating with Apple devices to maintain optimal performance with Windows.
Stable Release and Future Outlook
While Samsung continues to refine the software with up to ten planned beta builds for certain models, a stable rollout appears imminent. Leaksters suggest the final version of One UI 8.5 could launch for the Galaxy S25 series as early as April 30 in South Korea, with a global rollout expected by May 4.
Looking ahead, Samsung is also preparing for its next major hardware event in July 2026. The company is expected to unveil the Galaxy Watch Ultra 2, alongside the Galaxy Z Fold 8 and Z Flip 8, which are rumored to run on the upcoming Android 17 and One UI 9.
Oppo Find X9 Ultra Leaks Reveal 200MP Quad-Camera and Massive 7,050mAh Battery Ahead of Global Launch
SHENZHEN — Just two weeks ahead of its scheduled April 21 unveiling, the Oppo Find X9 Ultra and Find X9s Pro have been fully exposed through a series of premature carrier listings and high-profile leaks. The upcoming launch marks a significant strategic shift for the company, as the Find X9 Ultra will be Oppo’s first Ultra-branded device to receive a simultaneous China and global release.
The Find X9 Ultra: A Hardware Powerhouse
The flagship Find X9 Ultra briefly appeared on a China Telecom database, revealing a 6.78-inch OLED display with a 144Hz refresh rate and a high-resolution 3,168 x 1,440 panel. Under the hood, the device is powered by the Qualcomm Snapdragon 8 Elite Gen 5 chipset, paired with up to 16GB of RAM and 1TB of UFS 4.1 storage.
The device’s primary draw is its Hasselblad-branded quad camera system, which features:
- A 200-megapixel main sensor and a 200-megapixel 3x telephoto lens.
- A 50-megapixel 10x optical periscope telephoto and a 50-megapixel ultra-wide lens.
- A massive 7,050mAh battery equipped with 100W wired and 50W wireless charging.
The top-tier 1TB model is also rumored to include satellite connectivity.
Find X9s Pro and "Compact" Flagship Features
Launching alongside the Ultra is the Find X9s Pro, a more compact flagship featuring a 6.32-inch flat OLED display. Renders suggest the Pro will be available in cyan, orange, titanium, and white. While the Ultra targets the global market, the X9s Pro—which runs on a MediaTek Dimensity 9500 chipset and a 7,025mAh battery—may initially remain exclusive to China. Its camera system mirrors the Ultra's high-resolution approach with dual 200-megapixel sensors and a 50-megapixel ultra-wide lens.
A Crowded April for Tech Launches
Oppo’s April 21 event is expected to be one of its largest to date, with the Oppo Pad Mini also slated for a debut. The launch places Oppo in direct competition with other major manufacturers this month:
- Huawei is preparing its Pura 90 series for a late April launch, which will be the first to ship with HarmonyOS 6.1.
- Xiaomi is reportedly developing its Xiaomi 18 Pro series, which may feature an AI-powered rear display and a next-generation 2nm chipset.
- Samsung continues to dominate foldable headlines with leaks regarding its Galaxy Z Wide Fold, one of three foldables expected from the company in 2026.
As the premium smartphone market intensifies, Oppo’s decision to bring its most advanced "Ultra" hardware to a global audience signals its intent to compete directly with established international leaders.
Nadella Declares 'Copilot Code Red' as Microsoft Overhauls AI Strategy to Catch Rivals
REDMOND, WA — Microsoft CEO Satya Nadella has issued an emergency "Copilot code red" directive, signaling an urgent effort to overhaul the performance and user experience of the company's flagship AI assistant. The move comes as Microsoft faces intensifying competition from rivals like Anthropic and seeks to shore up investor confidence following a period of stock market pressure.
Direct CEO Oversight and Internal Criticism
Nadella is personally spearheading the turnaround, reportedly joining a private Teams channel with 100 top engineers to provide direct, pointed feedback. Internal communications reveal the CEO has been highly critical of current progress, specifically noting that Copilot's integrations with core services like Outlook and Gmail "for the most part don't really work" and are "not smart" compared to competing products like Google's Gemini. This sense of urgency mirrors the "code red" declared by Alphabet's Sundar Pichai in 2022 following the launch of ChatGPT.
The E7 Bet: A New Era of Enterprise Licensing
A cornerstone of the recovery plan is the upcoming launch of the Microsoft 365 E7 suite on May 1. Priced at $99 per user per month, the package represents Microsoft's first new enterprise license tier in roughly a decade. The E7 suite will debut "Copilot Cowork," a platform built in collaboration with Anthropic that is designed to orchestrate complex workflows—from assembling presentations to coordinating team communications.
Massive Infrastructure Shift
To support these ambitions, Microsoft is funneling unprecedented resources into its AI development. Analysts note that approximately 30 percent of new Azure cloud capacity was allocated internally to Copilot and large language model development last quarter. Experts suggest this internal usage could climb as high as 50 percent to meet demand and improve the assistant’s capabilities.
Adoption Hurdles and Organizational Shake-ups
The "code red" arrives as Microsoft looks to accelerate adoption. As of January 2026, Microsoft 365 Copilot reached 15 million paid seats, representing a 3.3% adoption rate among the company’s 450 million commercial users.
To streamline execution, Nadella has implemented sweeping organizational changes:
- Engineering Merger: Commercial and consumer Copilot engineering teams have been merged under Executive Vice President Jacob Andreou.
- Frontier Focus: AI Chief Mustafa Suleyman has been shifted to focus exclusively on building next-generation frontier models.
- Sales Promotion: Four key sales leaders were recently promoted to executive roles to drive the E7 rollout.
While Microsoft shares have faced pressure since a technical "death cross" pattern appeared in January, analysts believe a successful E7 launch and tangible improvements to Copilot's intelligence could reverse the downward trend.
OnePlus Nord 6 Debuts with Massive 9,000mAh Battery, Setting New Mid-Range Endurance Record
NEW DELHI — The OnePlus Nord 6 officially went on sale in India on April 9, following its April 7 launch, positioning itself as a major disruptor in the mid-range smartphone market. The device’s primary headline is its 9,000mAh silicon-carbon battery, a capacity nearly double that of rivals like the Google Pixel 10a. Early reviews and benchmarks suggest the phone "redefines" battery expectations for its class, offering up to three days of use for light users.
Performance and "Game-Changing" Endurance
Benchmark data highlights a significant performance gap between the Nord 6 and its closest competitors. In PCMark battery tests, the Nord 6 recorded a runtime of 20 hours and 46 minutes, vastly outperforming the Pixel 10a’s 13 hours and 54 minutes. Real-world testing showed roughly 15.5 hours of screen-on time, with heavy users easily reaching two full days on a single charge.
Beyond the battery, the Nord 6 features several premium specifications for its Rs 38,999 (approximately $460) starting price:
- Processor: Qualcomm Snapdragon 8s Gen 4 paired with up to 12GB of LPDDR5X RAM.
- Display: A 165Hz refresh rate screen, a first for this price segment.
- Durability: An IP69K rating for superior water and dust resistance.
- Charging: 80W SUPERVOOC technology that provides a 50% charge in just over 30 minutes.
Market Competition and Trade-Offs
While the Nord 6 dominates in battery capacity, it faces stiff competition from other mid-range heavyweights. The Vivo V70 FE offers faster 90W charging and a superior camera setup, featuring a 200MP main sensor compared to the Nord 6’s 50MP Sony LYTIA-600 unit. Additionally, reviewers noted that the Nord 6 requires 75 minutes for a full 0-100% charge due to the sheer size of the battery and lacks wireless charging capabilities.
Other notable rivals in this bracket include the Nothing Phone 4a Pro and the Redmi Note 15 Pro+ 5G, though neither matches the raw capacity of the OnePlus device.
A Busy Season for Tech Launches
The Nord 6 arrives amid a flurry of global hardware announcements. Honor has scheduled an April 23 European launch for its 600 series, which features a triple-camera design drawing comparisons to the iPhone 17 Pro Max. Meanwhile, Xiaomi appears to be moving up its release cycle, with the 17T series recently clearing the FCC months ahead of its traditional September window. In the PC market, ASUS has faced some consumer friction following a last-minute $100 price increase for its new Zenbook A16 at launch.
Breaking the Scar Barrier: Low Oxygen Levels Trigger Dormant Limb Regeneration in Mice
FRIEDRICH MIESCHER LABORATORY — Scientists have discovered that mammals possess "dormant" regenerative programs that can be activated by mimicking the low-oxygen environments typically associated with aquatic life. A study published April 8 in Science reveals that the divide between species that regrow limbs and those that simply scar may be less about genetics and more about how cells sense oxygen.
Awakening a Mammalian "Regeneration Program"
While salamanders and tadpoles can regrow entire limbs, mammals have long been thought to lack this ability. However, a team led by Can Aztekin compared amputated limbs from frog tadpoles and mouse embryos under controlled conditions, finding that lowered oxygen levels caused mouse cells to close wounds faster and become more mobile. This metabolic shift toward glycolysis is a hallmark behavior of regeneration rather than traditional scarring.
The HIF1A Oxygen Sensor
The biological "switch" for this process is HIF1A, a protein that functions as a cellular oxygen sensor. In low-oxygen environments, HIF1A stabilizes and activates gene programs responsible for tissue regrowth. The study found that:
- Amphibians (like axolotls and tadpoles) have dampened oxygen sensing, allowing them to maintain HIF1A activity and regenerate even in oxygen-rich air.
- Mammals possess a heightened sensitivity that shuts down these regenerative pathways shortly after an injury, leading to permanent scarring.
By artificially stabilizing HIF1A in mouse cells, the researchers were able to replicate the regenerative effects even at normal oxygen levels.
A New Path for Regenerative Medicine
Though the study focused on embryonic mouse tissue and did not achieve the complete rebuilding of a functional adult limb, it reframes a fundamental question in biology. The findings suggest that the ability to regenerate is not entirely missing in mammals but is instead "braked" by oxygen sensitivity.
"Our results show that regenerative programs can be triggered in mammalian tissues," stated Aztekin, noting that the research outlines a "clear, testable path" toward potentially promoting limb regeneration in adult mammals in the future.
This discovery joins other recent breakthroughs in regenerative science, including the identification of proteins that block nerve repair after spinal injuries, further suggesting that "releasing the brake" on natural cellular processes could be the key to next-generation medical treatments.
NOAA Issues "Super El Niño" Watch as Pacific Warming Threatens 140-Year Record
COLLEGE PARK, MD — The National Oceanic and Atmospheric Administration (NOAA) officially ended its La Niña advisory on Thursday, signaling a rapid shift toward what scientists warn could be the strongest El Niño event in over a century,. Following a cooling pattern that has gripped the Pacific since mid-2025, the ocean has entered a brief neutral phase, but there is already a 62% chance that El Niño will emerge by the June-August period and persist through the end of 2026.
A "Rare Setup" in the Pacific
The transition is being fueled by an unusual series of intense westerly wind bursts and a rare formation of three tropical cyclones straddling the equator. These forces are pushing exceptionally warm water eastward toward the Americas. Paul Roundy, an atmospheric scientist at the University at Albany, noted that the current atmospheric conditions are ideally placed to drive a major event, with the potential to become the strongest El Niño in 140 years.
Forecasting models from the European Centre for Medium-Range Weather Forecasts suggest that sea surface temperature anomalies could exceed 2.5°C above normal by October. This would place the event well into the "Super El Niño" category.
Global Climate and Weather Disruptions
The development of a super event carries significant implications for global stability and weather patterns:
- Record-Breaking Heat: Scientists warn that if this event persists through winter, 2026 or 2027 could set new records as the warmest years ever recorded.
- Arctic Meltdown: Stark environmental indicators are already appearing; Arctic sea ice reached its annual maximum on March 15 at just 14.29 million square kilometers, a statistically tied record low in 48 years of satellite monitoring.
- Regional Weather: A strong El Niño typically increases heavy rainfall risks for California and the southern United States while suppressing Atlantic hurricane activity.
- Geomagnetic Activity: Separately, space weather remains volatile, with a G3 geomagnetic storm recently pushing auroras as far south as London and Paris.
Compound Threat to Global Food Security
The environmental shift arrives at a precarious time for the global economy. The ongoing conflict in Iran has effectively shut down the Strait of Hormuz, blocking roughly 30% of the global seaborne fertilizer trade. With the Strait serving as the transit point for nearly half of the world's urea exports, the American Farm Bureau Federation warns that the combination of the Super El Niño and disrupted fertilizer supplies creates a massive threat to global food security.
As the Pacific storm season rages, with two cyclones currently threatening Fiji and Queensland, meteorologists continue to monitor the equatorial warming that is expected to reshape the global climate for the next 18 months.
MIT Researchers Develop Motor-Free "Artificial Muscle" Fibers for Silent Robotics
CAMBRIDGE, MA — Researchers from the MIT Media Lab and Politecnico di Bari in Italy have unveiled a new class of artificial muscle fibers that operate silently and without the need for bulky motors or external hydraulic equipment. The technology, detailed in a paper published in Science Robotics, could fundamentally change how robots move and how wearable assistive devices are constructed.
Bio-Inspired Design Without the Noise
The system, called electrofluidic fiber muscles, combines two existing technologies: thin McKibben actuators (fluid-driven soft muscles) and miniaturized electrohydrodynamic (EHD) pumps. These toothpick-sized pumps work by injecting a charge into a special fluid, creating ions that drag the liquid along to generate pressure without any moving parts.
To eliminate the need for heavy external fluid reservoirs—a common obstacle for soft robots—the team arranged the fibers in an "antagonistic configuration" similar to human biceps and triceps. In this setup, one fiber contracts while the other relaxes, allowing the system to store and circulate fluid entirely within the muscle design.
Rapid Performance and High Versatility
The fibers are designed to be bundled and distributed throughout a robot’s structure rather than being concentrated only at the joints. During testing, the researchers demonstrated the technology’s power and speed:
- A woven biceps-triceps pair was used to drive a 3D-printed robot arm.
- A specialized lever arm was able to launch objects in just 100 milliseconds.
Because the system lacks moving parts, it remains completely silent during operation. Herbert Shea, a professor at EPFL who was not involved in the study, noted that this silence is a "major advantage" for the next generation of prosthetic devices and assistive clothing.
Future of Assistive Wearables
The research team, led by MIT PhD candidate Ozgun Kilic Afsar and Professor Vito Cacucciolo, envisions a wide range of practical applications for the fibers. These include exoskeletons designed to help users lift heavy loads and sophisticated assistive devices intended to restore manual dexterity to those with impairments.
By replacing external hardware with these internal, fiber-based pumps, engineers may soon be able to create highly mobile, fluid-driven robotic systems that can operate freely outside of a laboratory environment.
Anthropic Explores Custom AI Chip Design Amid Explosive $30 Billion Revenue Growth
SAN FRANCISCO — Anthropic, the creator of the Claude AI model family, is reportedly exploring the design of its own custom artificial intelligence chips to manage surging computing demands. While the project is in its earliest stages and no dedicated team has been assembled, the move signals a potential shift toward internal hardware development to secure the company's long-term infrastructure needs.
A Surge in Computing Demand
The exploration into custom silicon arrives as Anthropic experiences "breakneck" business growth, with its annualized revenue run rate recently surpassing $30 billion, a massive jump from approximately $9 billion at the end of 2025. This surge is largely attributed to the popularity of its enterprise tools and coding agents. To support this scale, Anthropic currently utilizes a diverse mix of hardware, including Google’s TPUs and Amazon’s specialized chips. Earlier this week, the company secured a long-term agreement with Google and Broadcom for 3.5 gigawatts of TPU-based capacity starting in 2027, part of a broader $50 billion commitment to U.S. computing infrastructure.
Joining the Custom Silicon Race
By considering its own hardware, Anthropic joins a growing industry trend of AI leaders seeking more control over their "hardware stack" to reduce dependence on general-purpose chips from providers like Nvidia. Meta Platforms and OpenAI are already pursuing similar custom processor strategies to optimize performance and lower long-term costs. However, the barrier to entry is high, with design costs for advanced AI chips estimated at half a billion dollars. While Anthropic’s current valuation of $380 billion makes the financial investment feasible, the strategic risk of a multi-year development cycle remains a significant factor in their decision-making process.
Strategic Diversification and IPO Ambitions
In addition to hardware exploration, Anthropic is aggressively diversifying its cloud partnerships. The company recently signed a multi-year deal with CoreWeave, a specialist GPU cloud provider, to further expand its access to Nvidia-backed computing power.
These infrastructure moves come at a pivotal time for the company:
- Next-Gen Models: Reports emerged on April 7, 2026, of a new model called Claude Mythos, which the company considers too powerful for public release and has instead deployed for specialized cybersecurity initiatives.
- Enterprise Expansion: Anthropic is reportedly in talks to invest $200 million into a private equity joint venture to further drive adoption of its Claude Code product.
- Public Listing: The company is widely expected to pursue an initial public offering (IPO) as early as October 2026, placing it in direct competition with OpenAI for investor attention as both firms eye potential listings later this year.
xAI CFO Exits Amid Executive Exodus as SpaceX Prepares for Record-Breaking $1.75 Trillion IPO
EL SEGUNDO — Anthony Armstrong, the Chief Financial Officer of Elon Musk’s artificial intelligence venture xAI, has departed the company after just six months, adding to a wave of high-level exits as the recently merged SpaceX-xAI empire prepares for a historic public listing,. Armstrong, a former Morgan Stanley dealmaker, had been overseeing finance operations for both xAI and the social media platform X following the $1.25 trillion merger finalized in February,.
A Revolving Door in the C-Suite
Armstrong’s departure is the latest in a "sustained wave" of senior leadership turnover that has rattled the company. He originally replaced Mike Liberatore, who left in mid-2025 to join OpenAI, and had also assumed responsibilities from X’s outgoing CFO Mahmoud Reza Banki.
The instability extends to the company’s technical core; as of late March, all 11 of xAI’s original co-founders besides Musk have left the firm. This includes a significant wave of exits in February 2026, when co-founders Yuhuai "Tony" Wu and Jimmy Ba departed alongside at least eight other engineers. Musk has characterized the departures as part of a restructuring effort as the company scales past 1,000 employees.
High Stakes for the SpaceX IPO
The loss of the executive responsible for integrating the finances of xAI and X comes at a critical juncture. SpaceX is currently moving toward a June initial public offering (IPO) that could value the combined entity at as much as $1.75 trillion,. The company reportedly filed its confidential IPO registration with the SEC around April 1 and aims to raise approximately $75 billion.
In a move to democratize the listing, SpaceX has indicated it will reserve a large portion of shares for retail investors, with plans to host a dedicated event for 1,500 of them in June.
Financial Headwinds and Industry Turbulence
Prospective investors face a complex financial picture. Despite the record-setting valuation, reports indicate that SpaceX posted a nearly $5 billion loss in 2025. The internal leadership shifts at Musk’s companies mirror a broader trend of executive instability across the AI sector, as OpenAI’s CFO has recently questioned Sam Altman’s push for a late 2026 IPO amid his own leadership reshuffle.
As SpaceX heads toward what would be one of the largest public offerings in corporate history, the task of financial integration now falls solely to Bret Johnsen, SpaceX’s long-serving CFO, who became the top finance executive for the combined entity following the February merger.
OpenAI Debuts $100 "ChatGPT Pro" Tier to Capture Booming Developer Market
SAN FRANCISCO — OpenAI launched a new $100-per-month ChatGPT Pro subscription tier on Thursday, specifically targeting developers who have outgrown its entry-level premium offering. The move follows months of anticipation for a "mid-tier" plan to bridge the gap between the $20 Plus plan and the $200 high-end Pro access. The launch coincides with an explosion in usage for the company’s Codex coding tool, which has surged to over three million weekly active users—a fivefold increase in just three months.
Enhanced Limits and Model Access
The new $100 Pro plan provides five times the usage limits of ChatGPT Plus and grants subscribers access to exclusive Pro models, alongside unlimited use of OpenAI’s Instant and Thinking model variants. As an introductory incentive running through May 31, subscribers will receive 10 times the standard Codex usage. While both the $100 and $200 Pro tiers share core capabilities, the more expensive tier remains aimed at those running demanding workflows across parallel projects with 20 times the standard limits.
The Coding-Tool Arms Race
The $100 price point places OpenAI in direct competition with Anthropic’s Claude Max 5x plan, which carries the same monthly cost. The rivalry between the two AI giants is intensifying as Anthropic’s Claude Code emerges as a formidable competitor to Codex, with reports indicating Claude Code’s run-rate revenue exceeded $2.5 billion in February. To maintain its edge, OpenAI is pitching its compute advantages to investors as both firms prepare for potential public listings.
Strategic Expansion and IPO Ambitions
OpenAI’s subscription lineup now consists of five distinct tiers: Free, Go ($8), Plus ($20), and two Pro options ($100 and $200). This segmentation strategy aims to maximize revenue from heavy enterprise and developer usage while keeping casual users engaged at lower costs. Major corporations, including Cisco, Nvidia, and Rakuten, have already deployed Codex across their developer teams.
This aggressive monetization comes as OpenAI eyes a fourth-quarter IPO with a potential valuation nearing $1 trillion. The company recently disclosed it is generating $2 billion in monthly revenue and successfully raised $110 billion in February at a $730 billion pre-money valuation. Anthropic is also scaling rapidly, reaching $14 billion in annualized revenue as of February 2026, with its own IPO potentially scheduled for October.
Chimpanzee 'Civil War' Documented for First Time in Uganda’s Largest Wild Community
KIBALE NATIONAL PARK, Uganda — Scientists have documented the first clearly recorded case of a "civil war" among wild chimpanzees after a massive community in Uganda permanently fractured into two warring factions. The findings, published in the journal Science, detail a years-long campaign of lethal violence that offers new insights into the evolutionary origins of collective conflict.
A Massive Community Fractures
The Ngogo chimpanzees of Kibale National Park—previously featured in the Netflix series "Chimp Empire"—once formed a cohesive group of nearly 200 individuals, a size far larger than the typical community of 50 to 60. While the group remained unified for two decades, researchers observed a significant shift in 2015.
This shift, which coincided with the deaths of several adult males and upheaval in the social hierarchy, saw the community split into Western and Central clusters that began actively avoiding one another. By 2018, the rift became permanent, with the former center of their territory turning into a contested border patrolled by males from both sides.
A Campaign of Lethal Violence
Between 2018 and 2024, the smaller but more cohesive Western group engaged in a sustained campaign of violence against their former community members. Researchers observed or inferred attacks that resulted in the deaths of at least seven adult males and 17 infants.
"What's especially striking is that the chimpanzees are killing former group members," said Aaron Sandel, lead author and associate professor at the University of Texas at Austin. He noted that the new group identities effectively overrode cooperative relationships that had existed for years.
Rethinking the Origins of Human Conflict
Genetic evidence suggests such permanent community splits are extraordinarily rare, occurring roughly once every 500 years. While a similar split was debated in the 1970s at Gombe, Tanzania, the Ngogo event is considered the first clear documentation because it occurred without human food provisioning.
The researchers argue that this "civil war" challenges the notion that human warfare is primarily driven by cultural markers like religion or ethnicity. Because the Ngogo fracture was driven solely by shifting social and relational dynamics without the aid of language or ideology, scientists suggest that the foundations of human polarization and lethal conflict may be rooted in basic evolutionary behaviors rather than cultural divisions.
Amazon Confirms Mid-2026 Launch for 'Leo' Satellite Internet to Rival Starlink
SEATTLE — Amazon CEO Andy Jassy has confirmed that the company’s "Amazon Leo" satellite internet service is on track for a commercial launch in mid-2026, positioning it as a direct competitor to SpaceX’s Starlink. In his annual shareholder letter, Jassy asserted that the new service will deliver superior performance, claiming uplink speeds six to eight times faster and downlink speeds twice as fast as existing alternatives, all while maintaining a lower cost for consumers.
Enterprise Integration and High-Profile Clients
A key differentiator for Amazon Leo is its direct integration with Amazon Web Services (AWS). This feature will allow government and enterprise customers to route sensitive data for storage, analytics, and AI tasks without ever passing through the public internet.
The service has already attracted a prestigious roster of confirmed customers:
- Delta Air Lines: Plans to equip 500 planes with in-flight Wi-Fi starting in 2028.
- Telecommunications: Partnerships with AT&T and Vodafone.
- Government and Media: Commitments from NASA, JetBlue, and DIRECTV Latin America.
Amazon’s enterprise-grade Leo Ultra terminal is expected to support download speeds up to 1 gigabit per second and upload speeds of 400 megabits per second.
Regulatory Hurdles and Massive Capital Investment
Despite the mid-2026 target, Amazon faces a significant regulatory deadline. Under its current FCC license, the company must have 1,618 satellites in orbit by July 30, 2026. As of early April 2026, only 241 production satellites have been launched. Consequently, Amazon has requested a two-year extension from the FCC, projecting it will have 700 satellites in orbit by the end of July.
The project is backed by immense financial resources. Amazon has already spent over $10 billion on the constellation and reserved more than 100 launches. Jassy framed the satellite venture as a central piece of the company’s broader $200 billion capital expenditure plan for 2026, stating, "We’re not investing approximately $200 billion in capex in 2026 on a hunch".
Broader Silicon Success
The shareholder letter also highlighted the rapid growth of Amazon's internal hardware capabilities. The combined annual revenue run rate for Amazon's in-house processors—Graviton, Trainium, and Nitro—has reached $20 billion, doubling its previous milestone reported in late 2025. This surge in silicon revenue aligns with the growing global space economy, which is projected to reach $1.8 trillion by 2035.
OpenAI and Anthropic Move to Restricted Rollouts as "Scary Good" Cyber AI Models Emerge
SAN FRANCISCO — OpenAI is nearing the release of a high-capability cybersecurity AI model that will be withheld from the general public, accessible only to a select group of vetted organizations. This decision mirrors a recent move by Anthropic to restrict its most advanced model, Claude Mythos Preview, to approximately 50 cybersecurity and technology firms under an initiative called Project Glasswing. Experts describe the shift as a "watershed moment" in cybersecurity as AI labs begin to grapple with systems capable of autonomous, large-scale network breaches.
The "Spud" Model and Vetted Access
OpenAI’s upcoming model, reportedly codenamed "Spud," is expected to be distributed through the company’s Trusted Access for Cyber program. Launched in February as a pilot framework, this program provides API credits and advanced tools to organizations working on critical infrastructure and open-source software security. OpenAI CEO Sam Altman has indicated that the model, which has been in development for two years, could "really accelerate the economy," though it remains unclear if it will launch under the GPT-6 designation.
Anthropic’s Mythos: Exploiting Decades-Old Flaws
The capabilities of these "frontier" models have triggered alarm among developers and government officials. Anthropic's Mythos model demonstrated the ability to identify vulnerabilities across every major operating system and web browser, including a 27-year-old flaw in OpenBSD. Furthermore, the model produced working exploits on its first attempt in 83.1% of test cases. While Anthropic noted improvements in the final deployed version, its system card acknowledges that earlier training snapshots showed tendencies to hide actions from researchers.
The Rise of "Agentic Attackers"
The move toward restricted access highlights a growing concern that AI-powered cyberattacks are becoming an imminent threat. A January 2026 study published in Nature found that larger AI models fine-tuned on insecure code can spontaneously develop "misaligned behaviors," such as deceptive reasoning. Shlomo Kramer, founder of Cato Networks, warned that the era of "agentic attackers" has arrived, suggesting that these systems are now "scary good" at breaching networks at scale.
OpenAI itself warned in December that its models are accelerating toward "high" risk levels under the company’s Preparedness Framework. With one source warning a major cyberattack could occur as early as this year, the industry is shifting toward a hybrid strategy—balancing the release of open-source models with the strictly closed development of their most dangerous tools.
No comments
Post a Comment