Top Tech News Today, August 24, 2026: Alibaba, Amazon, Apple, Hugging Face, Nvidia, XPeng & More

It’s Monday, August 24, 2026,  and the tech world is moving faster than a humanoid robot clocking a sub-Bolt 100-meter dash. From record-smashing machines in Beijing and multibillion-dollar AI funding frenzies in Hong Kong and Tokyo to age-ban legislation in New Zealand, folding-phone breakthroughs, chip price shocks, and open-source platform deal talks, the past 24 hours delivered a high-stakes mix of frontier innovation, Big Tech power plays, and regulatory flashpoints reshaping startups and global markets.

Nvidia is reportedly putting billions behind an open-weight AI push and weighing another major bet on Perplexity, while China’s XPeng just raised more than $900 million for humanoid robots and Alibaba is pairing a new AI video model with billions in infrastructure spending. At the same time, the AI buildout is reshaping everything around it: Amazon is raising device prices as component costs climb, power-infrastructure companies are becoming acquisition targets, and governments are tightening control over the chips that increasingly sit at the center of geopolitical competition.

Elsewhere, Hugging Face is reportedly exploring a sale at a valuation above $13 billion, Iran-linked hackers are being blamed for knocking a British power plant offline, IBM is bringing Arm software into the mainframe, and New Zealand is pushing a social media ban for children under 16.

From AI models and humanoid robots to chips, electricity and cyberwarfare, here are the top technology news stories shaping the global tech economy today.

Technology News Today

Apple’s Foldable iPhone Impresses Early Testers with Passport Design

Early testers of Apple’s rumored foldable iPhone Ultra have praised its passport-style form factor, durable hinge, pocketability, and iPad-like internal display layout, according to reports. The device is expected to feature a nearly crease-free 7.8-inch inner screen and 5.5-inch cover display. Feedback highlights strong performance as a camera viewfinder and solid results in gaming, productivity, and entertainment, though the dual-camera system drew more mixed notes. Launch timing remains tied to potential September 2026 alignment with other iPhone models.

Positive internal reactions suggest Apple may finally enter the foldable market with refined hardware after years of development.

Why It Matters: Strong early feedback raises the stakes for Apple’s entry into foldables, potentially disrupting the segment long dominated by Samsung and Chinese makers.

Source: PCMag.

Nvidia weighs Perplexity investment at a $30 billion-plus AI startup valuation

Nvidia is discussing an investment in Perplexity as part of a new equity round that could value the AI search startup at more than $30 billion, according to The Information. The potential valuation represents a sharp jump from Perplexity’s roughly $20 billion valuation last year, as the company pushes beyond search into AI agents capable of handling more complex professional tasks. Perplexity’s annualized revenue has reportedly climbed from less than $250 million at the start of 2026 to more than $750 million.

The talks underscore how Nvidia is using its balance sheet to build influence across nearly every layer of the AI economy. Its investments increasingly extend beyond companies that simply buy Nvidia GPUs into model developers, cloud providers, AI applications and infrastructure businesses. For Perplexity, Nvidia’s involvement could provide capital, strategic credibility and closer access to the computing ecosystem behind many advanced AI services. It would also deepen an already complicated relationship in which Nvidia can simultaneously be a supplier, investor and potential technology partner to the companies building on its hardware.

Why It Matters: Nvidia is increasingly using strategic investments to secure a position not just underneath the AI industry through chips, but throughout the companies building products on top of them.

Source: TechStartups via The Information.

Nvidia strikes $6 billion Poolside deal to build a U.S. open-weight AI challenger

Nvidia is committing roughly $6 billion to a sweeping technology and investment agreement with AI startup Poolside as it tries to strengthen the U.S. position in open-weight AI models, according to The Wall Street Journal. The arrangement reportedly includes about $1 billion of investment in Poolside, access to the startup’s technology and the transfer of more than 100 engineers into Nvidia’s Nemotron model effort. Poolside’s founders are expected to remain with the startup rather than joining Nvidia.

The deal is notable because Nvidia is moving further upstream from selling the hardware used to train AI into developing the models themselves. Chinese developers including DeepSeek have put growing pressure on U.S. companies by releasing capable models with downloadable weights, giving developers more control than closed systems provide. Nvidia’s Nemotron strategy could help create a stronger American open-weight alternative while stimulating demand for the company’s own computing stack. The arrangement also reflects the escalating value of AI engineering talent: instead of acquiring entire startups, large technology companies increasingly structure licensing, investment and hiring agreements that deliver intellectual property and technical teams without a conventional takeover.

Why It Matters: Nvidia is turning its enormous chip-generated cash flow into a broader AI platform strategy that increasingly spans hardware, software, models and talent.

Source: The Wall Street Journal.

XPeng’s robotics unit raises more than $900 million as China’s physical AI race accelerates

Chinese electric vehicle maker XPeng said its robotics business has raised more than $900 million in its first external funding round, valuing the operation at more than $6.3 billion. IDG Capital and Gaorong Ventures led the financing, while Tencent and Alibaba participated as strategic investors. XPeng says the capital will support humanoid robotics development, mass production and continued work on its physical AI models. The company has committed to moving its IRON humanoid robot toward production before the end of 2026.

The financing is one of the clearest signs yet that China’s humanoid robotics boom is becoming a major venture-capital category rather than an experimental sideline for automakers. XPeng can reuse technology developed for autonomous vehicles, including its Turing AI chips, perception systems, manufacturing expertise and increasingly sophisticated software stack. That overlap could give automakers an advantage over standalone robotics startups attempting to build hardware, AI and manufacturing operations simultaneously. Alibaba and Tencent’s participation also shows how China’s largest technology companies are placing strategic bets on embodied AI as the industry searches for the next major computing platform beyond smartphones and cloud software.

Why It Matters: A $900 million robotics round suggests physical AI is moving from demonstrations and prototypes toward an industrial-scale investment race.

Source: The Next Web.

Hugging Face explores potential sale at $13 billion-plus valuation

Hugging Face is exploring a potential sale that could value the influential AI developer platform at $13 billion or more, Business Insider reported. The company has reportedly hired a bank to gauge interest from possible buyers, although discussions remain preliminary and no transaction is guaranteed. Hugging Face was valued at about $4.5 billion in its last major funding round in 2023, when investors included Salesforce, Google, Amazon, Nvidia, Intel and other major technology companies.

A sale would put one of the AI industry’s most important neutral infrastructure platforms into the hands of a single owner. Hugging Face hosts models, datasets and developer tools used across the open-source and open-weight AI communities, making it something closer to a GitHub for machine learning than a conventional model developer. That position could attract cloud providers, chipmakers and software companies looking for a stronger foothold in the developer ecosystem. But an acquisition would also raise questions about whether developers would continue viewing the platform as an independent home for models from competing AI companies. The reported valuation reflects how valuable distribution and developer communities have become as competition shifts beyond raw model performance.

Why It Matters: Hugging Face has become core infrastructure for modern AI development, making control of the platform strategically valuable far beyond its direct revenue.

Source: Business Insider.

Alibaba launches Wan3.0 AI video model as it pours billions into AI infrastructure

Alibaba Cloud has officially launched Wan3.0, its latest generative AI video model, adding the ability to produce videos up to 30 seconds long from inputs including documents, spreadsheets, presentations, PDFs, and web content. Alibaba says the model improves instruction following, shot consistency, and audio quality. Wan3.0 entered public beta earlier this month and has already been tested for advertising, tourism, short-form entertainment, and music-video production.

The launch arrives as Alibaba dramatically increases its financial commitment to AI. The Chinese technology giant announced a roughly $10 billion share placement to help fund computing infrastructure and development around its Qwen models and cloud business. That combination matters: generative AI competition is increasingly becoming a contest between companies that can finance enormous computing requirements while continually shipping new models. Wan3.0 also illustrates a broader shift toward multimodal systems that can convert existing business documents directly into finished media. Rather than asking users to start with carefully written prompts, AI platforms are beginning to treat presentations, spreadsheets, and reports themselves as creative inputs, potentially widening adoption inside corporations.

Why It Matters: Alibaba is pairing new generative AI products with enormous infrastructure spending as China’s Big Tech companies intensify competition with U.S. AI platforms.

Source: TechNode.

Chinese Humanoid Robots Shatter Usain Bolt’s 100m Record at Beijing Games

Two Chinese humanoid robots surpassed the long-standing men’s 100-meter world record held by Usain Bolt during the opening of the second World Humanoid Robot Games in Beijing. Tiangong Ultra, developed by the Beijing Humanoid Robot Innovation Centre, finished a preliminary heat in 9.39 seconds, while Honor’s Lightning clocked 9.47 seconds—both faster than Bolt’s 9.58-second mark set in 2009. A separate test run by Lightning reportedly hit 9.32 seconds. The robots demonstrated dramatic gains from last year’s event, when the same Tiangong model managed only 21.50 seconds. Organizers noted the machines still struggled with controlled stops, slamming into mats or collapsing afterward, highlighting remaining challenges in balance and deceleration.

The results underscore rapid advances in AI-driven mobility systems and bipedal robotics, particularly in China, where state-backed and private efforts are accelerating. These demonstrations occur amid intensifying global competition in humanoid platforms for industrial, logistics, and consumer applications. Improved leg design, longer strides, and refined control algorithms enabled the leap, yet practical deployment still requires solutions for real-world tasks beyond straight-line sprinting.

Why It Matters: Record-breaking robot speeds signal China’s accelerating push in frontier robotics, pressuring U.S. and European developers to match progress in AI-powered physical systems.

Source: Nikkei Asia.

Xiaomi launches Xring O3 AI chip as it deepens its push into custom silicon

Xiaomi has unveiled the Xring O3, the latest generation of its internally developed smartphone processor, as the Chinese hardware giant works to reduce dependence on Qualcomm and MediaTek. TSMC will manufacture the processor using an advanced 3-nanometer process, according to people familiar with the plans. The chip is expected to appear in an upcoming flagship Xiaomi foldable. Xiaomi is also reportedly developing an Xring O100 neural processor for its MiMo AI models and a separate D100 chip for autonomous-driving systems.

Custom silicon has become one of the most important competitive advantages available to large technology companies. Apple demonstrated how designing processors internally can improve performance, battery life, product integration and long-term supply-chain control. Google, Amazon, Microsoft and Meta are pursuing similar strategies in data centers. Xiaomi is now applying that playbook across smartphones, AI and vehicles. The company has reportedly invested more than 20 billion yuan, or roughly $3 billion, into chip development and built a semiconductor organization numbering more than 3,000 employees. Success would give Xiaomi greater control over the hardware behind both its consumer devices and growing physical AI ambitions.

Why It Matters: Xiaomi’s chip strategy shows custom silicon spreading beyond U.S. Big Tech as Chinese companies seek greater control over AI, mobile computing and supply chains.

Source: Reuters.

Amazon raises Echo, Kindle and Fire TV prices by as much as 60% as memory costs climb

Amazon has raised prices across parts of its Echo, Kindle and Fire TV hardware lineup, with some products up as much as 60%, according to The Verge. The increases affect several of Amazon’s best-known consumer electronics categories and come as memory and storage components become more expensive. Hardware companies have been grappling with higher component costs as AI infrastructure consumes a growing share of advanced memory production and pressures parts of the broader semiconductor supply chain.

Amazon’s hardware business has historically served a strategic purpose beyond device margins. Echo devices bring Alexa into homes, Fire TV strengthens Amazon’s position in streaming, and Kindle anchors its digital-book ecosystem. Significant price increases therefore create a tricky tradeoff: Amazon can protect margins, but higher prices could slow adoption of devices designed to pull consumers deeper into its services. The increases also signal that the AI infrastructure boom is affecting technology far beyond data centers. Competition for memory, storage and manufacturing capacity can eventually filter down into phones, PCs and household electronics, turning enormous AI capital expenditures into higher costs elsewhere in the computing ecosystem.

Why It Matters: The AI infrastructure boom is beginning to influence consumer hardware economics, showing how data-center demand can ripple through global component supply chains.

Source: The Verge.

Infineon acquires Indian chip startup C2i to strengthen AI data center power systems

German semiconductor giant Infineon has acquired Bengaluru-based C2i Semiconductors to expand its technology for managing electricity inside AI data centers. C2i specializes in software-defined multiphase controllers, smart power stages and system-level architectures, including technology aimed at vertical power delivery. Infineon says the acquisition will strengthen its ability to supply power-management systems from the electrical grid down to processors while expanding its research and development presence in India. Financial terms were not disclosed.

The acquisition targets one of the less visible bottlenecks in the AI infrastructure buildout: converting and delivering electricity efficiently to processors consuming unprecedented amounts of energy. As racks become denser and accelerators draw more power, traditional approaches to voltage regulation and power distribution become less efficient. That is creating new opportunities for semiconductor companies specializing in power electronics, cooling, and electrical management rather than GPUs themselves. Infineon is already a major supplier of power semiconductors, and C2i adds specialized intellectual property around controlling high-current processor workloads. India also gains another connection to the global AI semiconductor supply chain, which has become an increasing strategic priority for governments and multinational chipmakers.

Why It Matters: AI’s chip boom is creating a parallel market for the power-management technology required to feed increasingly dense and energy-intensive computing systems.

Source: Infineon Technologies.

IBM unveils dual-architecture processor that brings Arm software to mainframes

IBM has unveiled a next-generation processor designed to run both IBM Z and Arm instruction sets on the same processor cores, bringing the enormous Arm software ecosystem into future IBM Z mainframes and LinuxONE systems. Developed through IBM’s partnership with Arm, the 2-nanometer chip features 11 high-performance cores running above 5.7 GHz, dedicated AI inference acceleration, an on-chip data-processing unit, and extensive enterprise security features. IBM says individual cores can execute Arm and IBM workloads rather than relying on separate sets of specialized cores.

The architecture could narrow a long-standing divide between traditional enterprise computing and modern cloud-native software. IBM mainframes continue to process huge volumes of banking, insurance and government transactions, while much of the newer cloud and AI software ecosystem increasingly runs on Arm-based systems. Allowing those environments to coexist more naturally could make it easier for large enterprises to modernize applications without abandoning infrastructure prized for reliability and security. It also gives Arm another route into mission-critical corporate computing after years of expansion across smartphones, cloud servers and AI infrastructure.

Why It Matters: IBM is attempting to connect decades-old mission-critical computing infrastructure directly with the software ecosystem powering modern cloud and AI development.

Source: SiliconANGLE.

nVent buys Maverick Power for $1.75 billion in another big AI data center infrastructure deal

Electrical infrastructure company nVent has agreed to acquire Maverick Power for $1.75 billion, expanding its exposure to the surge in data-center construction driven by AI computing. The agreement could include up to another $550 million in payments if Maverick reaches performance targets in 2027 and 2028. Texas-based Maverick makes low- and medium-voltage switchgear, switchboards, modular power systems and other equipment used to distribute electricity through large facilities. It is expected to generate roughly $700 million in revenue this year.

The transaction illustrates how the AI boom is creating valuable companies far outside the semiconductor industry. GPUs receive most of the attention, but a data center cannot operate without substations, transformers, switchgear, cooling systems and increasingly sophisticated power-distribution equipment. As operators design facilities that can support much denser computing racks, electrical architecture itself becomes a strategic technology category. Companies positioned between utilities and AI servers are therefore attracting capital, acquisitions and premium valuations. For nVent, Maverick adds a larger platform for supplying those systems as hyperscalers and specialized AI cloud companies continue building capacity.

Why It Matters: The AI infrastructure race is increasingly becoming an electrical-engineering race, creating major acquisition opportunities for companies that can move and manage enormous amounts of power.

Source: nVent Electric.

Taiwan indicts nine over alleged illegal export of Nvidia AI servers to China

Taiwanese prosecutors have indicted nine people in an investigation into the alleged illegal shipment of advanced AI servers containing Nvidia chips to China. The defendants include employees connected to Nvidia, Super Micro Computer and other technology companies, according to prosecutors. Authorities allege that documents were falsified to show 130 Super Micro B300 servers would remain in Taiwan even though dozens were ultimately shipped to China directly or routed through locations including Indonesia, Japan and Hong Kong. Taiwanese customs intercepted 56 servers.

The case illustrates the growing complexity of enforcing semiconductor export restrictions. Governments can restrict direct sales of advanced processors, but servers and components travel through global distributors, integrators, and logistics networks, making it harder to trace the final destination. Taiwan sits at the center of that ecosystem because of its dominant role in advanced chip manufacturing and hardware assembly. Authorities have therefore faced growing pressure to ensure the island does not become a route around U.S.-led technology controls targeting China. The involvement of employees inside major technology companies also highlights why compliance increasingly depends on internal controls as much as border enforcement.

Why It Matters: The AI chip race has turned export compliance into a major geopolitical issue, with governments increasingly targeting the supply chains that can circumvent semiconductor restrictions.

Source: Reuters.

New Zealand pushes under-16 social media ban as global tech regulation spreads

New Zealand Prime Minister Christopher Luxon has proposed legislation that would prevent children under 16 from using designated social media platforms, extending a global regulatory push to limit young people’s exposure to online services. Platforms would be required to take reasonable steps to verify users’ ages, potentially using methods such as digital identity systems or age-estimation technology. Companies that fail to comply could face penalties, while children and parents would not be punished. Messaging, gaming and productivity AI services would generally fall outside the proposed restrictions.

The proposal faces political hurdles and is unlikely to receive a parliamentary vote before New Zealand’s upcoming election, but its significance extends beyond the immediate legislative calendar. Governments are increasingly moving from demanding better parental controls to questioning whether minors should have access to large social networks at all. Australia has already pushed aggressively in this direction, and similar proposals have surfaced elsewhere. The technological challenge is equally important: effective age verification without collecting excessive personal information remains difficult. Platforms could therefore face pressure to build entirely new identity and compliance systems if age restrictions become widespread.

Why It Matters: Social media regulation is shifting from content moderation toward access control, potentially forcing global platforms to rethink identity verification and how they serve younger users.

Source: Associated Press.

Iran-linked hackers blamed for cyberattack that shut down a UK power plant for four days

Hackers linked to Iran have been blamed for a cyberattack that reportedly forced a small British power plant offline for four days, according to The Guardian. The affected facility has not been publicly identified, and officials said the wider UK electricity system was not disrupted. Britain’s National Cyber Security Centre has been involved as authorities assess a growing wave of threats against critical infrastructure amid heightened geopolitical tensions involving Iran, the United States and the United Kingdom.

Even a relatively small outage matters because operational technology in power generation, water systems, transportation and manufacturing can have physical consequences when compromised. Cyberattacks against such systems differ from conventional data theft: disrupting industrial controls can stop machinery, interrupt public services and potentially create safety risks. Governments have warned for years that nation-linked hacking groups probe infrastructure for weaknesses they could exploit during political or military crises. The reported UK incident suggests that such operations may increasingly move beyond reconnaissance into direct disruption, raising the urgency around segmentation, incident response and security for industrial equipment.

Why It Matters: Cyber conflict is increasingly capable of crossing from computer networks into physical infrastructure, turning industrial cybersecurity into a national-security requirement.

Source: The Guardian.

GrapheneOS expands beyond Google Pixel with Motorola foldables coming in 2027

Privacy-focused Android alternative GrapheneOS plans to support Motorola smartphones beginning in 2027, ending its longstanding dependence on Google Pixel hardware as its primary supported device family. The project says support will start with a conventional Motorola flagship before expanding to upcoming Razr Fold and Razr Ultra devices. Future models are expected to include hardware features GrapheneOS considers essential, including stronger memory protection and secure-element capabilities. Motorola is actively contributing engineering work needed to bring the operating system to its devices.

The partnership is significant for the small but influential market for privacy-oriented mobile operating systems. GrapheneOS has built a reputation around hardened Android security, strict app isolation and reducing dependence on Google services, but its hardware choices have historically been limited. Support from a major handset manufacturer could make alternative mobile operating systems more accessible while pressuring other device makers to expose the security features independent software projects require. Foldables also give GrapheneOS access to a premium hardware category where consumers previously had almost no serious privacy-first operating system option.

Why It Matters: Motorola’s cooperation with GrapheneOS could give privacy-focused mobile computing a path beyond niche Pixel installations and into mainstream premium hardware.

Source: The Verge.

STMicroelectronics and NUS launch Singapore lab for energy-efficient edge AI chips

STMicroelectronics and the National University of Singapore have launched a four-year research laboratory focused on developing more energy-efficient semiconductors for edge AI. The collaboration will investigate technologies that can run artificial intelligence workloads directly on devices rather than sending every task to large cloud data centers. The program brings together STMicroelectronics’ chip-design and manufacturing expertise with university research in computing, electronics and AI systems.

Edge AI is becoming increasingly important as models move into cars, robots, industrial equipment, cameras, wearables and consumer electronics. Processing information locally can reduce latency, lower cloud-computing costs and keep sensitive information on the device, but it requires processors capable of performing increasingly sophisticated inference under strict energy constraints. The Singapore project also reflects the country’s broader effort to strengthen its position in advanced semiconductor research as governments across Asia compete for higher-value roles in the chip supply chain. For STMicroelectronics, improvements in low-power AI processing could strengthen its position in automotive, industrial and embedded markets where Nvidia-style data-center GPUs are often impractical.

Why It Matters: The next phase of AI will increasingly happen outside massive data centers, making efficient edge processors a critical battleground for chips powering robots, vehicles and connected devices.

Source: TechNode Global.