LectureRechercheRadarCadre d'investissement
Connexion / Inscription
ToutTechnologieProduitsInvestissementBourseAutres

Lecture

2026-11-0956 publications

Actualiser

Roborock Posts Strong Q3 Growth While Ecovacs Profit Plunges: Robot Vacuum Market Enters Divergence Phase

Sujet · 全球扫地机器人市场分化Matériel
2026-11-09 15:00 GMT+8

According to a report by South Korea's ETNews, the global robot vacuum cleaner market witnessed a significant divergence in financial performance among top manufacturers in the third quarter of 2026. Chinese listed company Roborock was the only leading player to achieve substantial growth in both revenue and profit, while its competitor Ecovacs faced a severe decline in profitability.

Data shows that Roborock's Q3 revenue reached 5.614 billion yuan, a year-on-year increase of 34.85%. Net profit attributable to shareholders was 479 million yuan, up 33.01% year-on-year. In contrast, Ecovacs reported revenue of 4.647 billion yuan, growing 10.61%, but its net profit fell sharply by 72.43% to 121 million yuan.

Ecovacs attributed the profit collapse to rising prices for memory semiconductors and losses from a stronger yuan. Although its service robot shipments increased by 44.1% in the first half of the year, related revenue growth was only 26.8%, indicating that volume expansion did not translate into proportional profit gains. Analysts suggest that changes in product mix towards mid-to-low-end models and overseas markets, combined with price competition, squeezed margins.

Meanwhile, Dreame, another major manufacturer that does not disclose quarterly results, is reportedly restructuring its business to integrate duplicate product lines between its main brand and Mova to address inventory pressures. IDC data indicates that Roborock holds approximately 23.7% of the global robot vacuum market share, ranking first.

Regulatory changes also pose uncertainties. The U.S. Federal Communications Commission (FCC) has restricted equipment certification for new robots produced abroad since July. Manufacturers with local supply chains or existing certifications, such as SharkNinja (which reported a 21.2% revenue increase and a 55.1% rise in adjusted net profit), may benefit. Overall, the diverging performance of top firms suggests the global robot vacuum market has entered a maturity phase, with competition shifting from market share expansion to cost control and profit quality.

Sources :미래를 보는 창 - 전자신문

Recherche

Open-source AI tool 'Craft Anything' released to clone closed-source web and app projects

Prise rapide
2026-11-09 05:56 GMT+8

Japanese tech media outlet TechnoEdge reported on November 9, 2026, the release of "Craft Anything," an open-source AI tool designed to analyze the mechanisms of existing websites, applications, and games and use AI to generate independent versions with similar functionality, even when the original project's source code is not public.

Unlike traditional code generation assistants, Craft Anything provides a complete automated workflow consisting of four steps: "Study," "Build," "Check," and "Ready." Users simply specify the target they wish to emulate and describe their requirements, and the tool automatically executes the project generation process. During the build phase, the AI does not just write code but also compares the generated result visually and functionally against the original target (such as webpage display across different screen widths or game self-tests), iterating corrections until it closely resembles the original.

The tool includes 300 core engine tools, totaling 320 when including AI integration and setup utilities, capable of handling various objects such as PC and mobile apps, websites, and game files. For actual programming tasks, Craft Anything primarily relies on external AI agents, currently supporting Claude Code. Additionally, it can operate as a command-line interface (CLI) or MCP server, allowing developers to call its parsing engine directly from existing AI toolchains.

Craft Anything offers two main analysis modes: the "Clean room" mode analyzes only visible information such as screenshots, layout measurements, and colors without reading any internal code; while the "Deep" mode allows reading internal code and JavaScript using built-in engines or Ghidra, suitable for private use. The tool is currently in Alpha testing and is open-sourced under a dual MIT or Apache-2.0 license.

Sources :テクノエッジ TechnoEdge

Recherche

YC and Olam Labs Experiment: 50 AI Agents Get Stuck in Bureaucracy Instead of Building Civilization in Minecraft

Sujet · 多智能体官僚主义Prise rapide
2026-11-09 00:55 GMT+8

According to a report by German tech outlet t3n.de, researchers from Y Combinator and Olam Labs conducted an experiment where 50 AI agents based on OpenAI's GPT-6.1-Sol model attempted to build a thriving civilization in Minecraft within three hours. The results showed that the agents did not achieve their stated goals but instead fell into a trap of "bureaucracy" similar to human organizations.

The study found that the agents spent considerable time rationing scarce food resources in crowded accommodations and spontaneously formed a governance structure including written rules and approval processes. A notable detail was that one group of AI agents spent 11 minutes solely to appoint a "fertilizer manager." This behavioral pattern suggests that current multi-agent systems, when lacking explicit external directives, tend to allocate computational resources to internal coordination and management rather than direct productive execution.

Additionally, the research team compared behavioral differences between models. In another adversarial test, 10 GPT-6.1-Sol agents faced off against 10 Claude-Opus-5.5 agents. The report states that GPT agents adopted more aggressive and sophisticated strategies, while Claude agents behaved more defensively. When handling tool errors, GPT agents diagnosed the issue and shared workarounds, whereas multiple Claude agents concluded that "breaking blocks was impossible."

The experiment was published via a blog post by Y Combinator and Olam Labs and remains preliminary. While the environment is limited to a virtual game world, the revealed bottleneck in multi-agent collaboration—excessive focus on procedural justice and resource allocation over final output—provides specific reference points for understanding potential failure modes of current AI agents in real-world workflows.

Sources :t3n.de

Recherche

Japanese AI Firm Preferred Networks Releases Open-Weight Financial LLM PLaMo 3.0 Fin for On-Premise Use

Matériel
2026-11-09 15:00 GMT+8

Japanese artificial intelligence company Preferred Networks (PFN) released "PLaMo 3.0 Fin," an open-weight large language model specifically designed for the financial industry, on November 9, 2026. The model features 8 billion parameters and aims to address the regulatory constraints faced by banks and securities firms that cannot transfer customer data to external clouds. By providing a lightweight model capable of running on local servers or in closed network environments, PFN offers an alternative to cloud-based generalist LLMs for financial institutions bound by strict data compliance rules.

PLaMo 3.0 Fin was developed through continued pre-training, supervised fine-tuning (SFT), and direct preference optimization (DPO) based on a foundation model co-developed by PFN and Japan's National Institute of Information and Communications Technology (NICT). Notably, the SFT phase utilized PFN's proprietary AI semiconductor, "MN-Core 2," as the computing infrastructure. PFN reported that code porting to MN-Core 2 was completed within one week using an AI coding assistant, achieving training speeds comparable to GPU environments. The model supports a context length of 262,144 tokens, includes reasoning capabilities with Japanese output for thought processes, and can call external tools such as web search.

In terms of performance, PFN's published benchmark results show that PLaMo 3.0 Fin achieved an average score of 55.8 across five tasks in the "japanese-lm-fin-harness" test set, an improvement of 7.8 points over the base model. This score surpasses Gemma 4 E4B (45.6) and gpt-oss-20b (43.0). However, it remains behind similarly sized models like Qwen3.5-9B (60.1) and Ministral 3 14B (62.8). In text generation quality assessed by "pfmt-bench-fin-ja," the model scored 8.35, and its tool-calling capability averaged 71.3 points, both lower than the current leader, Gemma 4 E4B. These results are self-reported by the vendor and have not yet been independently reproduced.

Unlike Anthropic and OpenAI, which offer cloud-based financial services solutions, PFN adopts a strategy of distributing small, specialized models for clients to run locally. While this approach sacrifices some absolute performance ceiling, it meets the rigid data sovereignty requirements of specific industries. PFN explicitly noted in the model card that validation is currently limited to Japanese and that hallucination risks exist. The company plans to continue developing more competitive finance-specific models to narrow the gap with leading international open-source alternatives.

Sources :PC Watch

Recherche

Japanese AI Firm Shizuku AI Launches Full-Duplex Voice Model 'Shizuku-Duplex' for Real-Time Interaction

Prise rapide
2026-11-09 14:35 GMT+8

Japanese AI startup Shizuku AI announced "Shizuku-Duplex," a full-duplex Japanese audio dialogue model, on November 8. The model is designed to enable natural human-AI interaction for AI characters, allowing them to listen and respond simultaneously, even starting replies before the user finishes speaking.

Unlike traditional cascaded systems that sequentially process speech recognition, text generation, and synthesis, Shizuku-Duplex uses a full-duplex architecture operating in 80-millisecond ticks. In each tick, it listens to user audio while generating its own. The system also features an asynchronous background model for tasks like web search, maintaining conversation flow while waiting for results.

According to Shizuku AI's internal evaluation, the model achieved the highest conversational naturalness among six compared models, with a median response time of 0.57 seconds after the user stops speaking. It stopped within two seconds of user interruption in 89% of cases. The company plans to expand language support and add image recognition capabilities.

Sources :PC Watch

Recherche

TechOrange Columnist Spencer Liu: AI Output Has Increased, but Who Is Responsible for Understanding?

Matériel
2026-11-09 14:12 GMT+8

Over the past few months, we have likely used AI to do quite a bit—writing reports, organizing meeting minutes, generating proposals. The volume of output has indeed increased, but have you ever stopped to ask yourself: Did any of this actually make any decision within the company better? Spencer Liu raised this question in his TechOrange column, noting that mathematics is undergoing a similar shift: when AI can produce proofs in bulk, the standards previously used to measure quality suddenly become unreliable.

In September 2026, OpenAI released ten thousand AI Agents, directing them to search in different directions simultaneously and exchange intermediate results with one another. The entire process transmitted 2.7 million messages, took 88 hours, and cost approximately $15 million, ultimately solving the Navier-Stokes equations—one of the seven Millennium Prize Problems listed by the Clay Mathematics Institute. Twenty years earlier, Russian mathematician Grigori Perelman independently developed new proof tools along the Ricci flow method, spending nearly a decade to solve the Poincaré conjecture; he published his paper directly on arXiv for global mathematicians to verify. In 2006, he declined the Fields Medal, and in 2010, he again refused the $1 million Millennium Prize. Both events were considered successes, but Liu points out the key difference: Perelman left behind a methodology that provided new tools for future researchers, whereas OpenAI left behind only an answer.

This is the exact same problem encountered by enterprises implementing AI. When a proposal is generated by AI in just a few minutes, who is responsible for confirming its correctness? More often than not, the answer is no one—because the speed of verification cannot keep up with the speed of production, and gatekeeping is rarely regarded as a performance metric. The report is submitted, but has the team gained deeper understanding as a result?

Liu cites another Fields Medalist, Terence Tao, whose judgment during his October 2026 "Math 2.0" speech at the California Institute of Technology was pivotal: the mathematics community has long prioritized being the first to solve difficult problems as the core standard for measuring achievement, but when AI can mass-produce proofs, this standard has become "misaligned"—deviating from what it was originally intended to measure. Tao went further, stating that continuing to blindly optimize solely for problem-solving is now substantially harming the long-term health of mathematics. He delivered a line that Liu considered the most important of the entire speech: "The problem is the lighthouse, not the destination." If you rush straight to the base of the lighthouse at top speed, you have indeed arrived, but you have also abandoned the voyage itself.

When the Clay Mathematics Institute acknowledged that OpenAI's result "appears to be solved," it specifically emphasized that the procedures for verification and attribution of credit would be "deliberately slow." Liu reflected deeply on this deliberate slowness: in an era where everyone competes to reach the summit first, why does someone argue that success should be attributed to taking things slowly? Because if AI produces a thousand mathematical proofs in a day, yet there are insufficient researchers capable of reading, understanding, or extending these findings, we have not achieved a thousand-fold scientific progress.

Sources :TechOrange 科技報橘

Recherche

NVIDIA Board Launches Independent Investigation into CEO Jensen Huang Over Foreign Intelligence Risks

Sujet · 英伟达黄仁勋间谍指控Matériel
2026-11-09 12:44 GMT+8

The board of directors of U.S. chipmaker NVIDIA announced on November 8, 2026, the initiation of an independent investigation to examine whether Chief Executive Officer Jensen Huang has become a target of foreign intelligence operations. According to media reports citing an NVIDIA statement, the investigation will be conducted by a special committee composed of independent directors, assisted by external legal counsel. NVIDIA emphasized that there is currently no reason to believe that U.S. national security or the company's commercial secrets have been compromised, but the board will continue to review all available information.

Following the news, global financial markets experienced volatility. As reported by Bloomberg, Nasdaq 100 index futures fell by as much as 1%. However, Asian stocks rebounded subsequently, with the MSCI Asia Pacific Index rising 1.7% and Japan's Nikkei Index up 2.8%, indicating that the market did not panic broadly due to this single event. Analysts noted that tech stocks had already been under pressure due to concerns about potential U.S. congressional regulation of the AI industry and returns on data center investments.

Stacy Rasgon, an analyst at Bernstein, told Bloomberg that the investigation highlights NVIDIA's high dependence on Huang personally, known as "key-man risk." Since NVIDIA has not publicly disclosed a clear succession plan and its management structure is highly centralized, investors are increasingly focused on the company's governance structure and leadership stability.

Regarding the impact on the industry, Zhang Jianyi, President of the Taiwan Economic Institute, stated in a media interview that although related national security issues remain to be clarified—including whether AI technology has flowed into China and whether chip acquisition channels are compliant—he believes this event will not hinder the long-term development of the AI industry. Zhang pointed out that NVIDIA currently maintains a high gross margin (75% in Q2 FY2027), and despite future pressure from competitors like AMD, the overall impact on the AI supply chain is expected to be limited.

Additionally, chairs of three committees in the U.S. House of Representatives requested on November 5 that Huang attend a closed-door briefing on November 18 to explain relevant national security concerns and NVIDIA's security management mechanisms. NVIDIA is scheduled to release its next quarterly earnings on November 17 U.S. time, where data on AI data center demand, revenue, and profit outlook will serve as important basis for the market to assess its operational prospects.

Sources :TechOrange 科技報橘

Recherche

Japan's AIST Develops Automated Experiment System for Solid-State Battery Materials

Sujet · 产总研固态电池自动化系统Prise rapide
Vérifié 2026-11-09 13:17 GMT+8

The National Institute of Advanced Industrial Science and Technology (AIST) in Japan announced on November 5, 2026, the development of a "collaborative automated experiment system" in partnership with DENSO WAVE, Aichi Institute of Technology, and the Aichi Prefectural Industry Science Center. The system is designed to automate the entire process of mechanochemical synthesis for all-solid-state battery materials, from powder weighing to sample storage.

All-solid-state batteries are considered a next-generation energy storage technology offering high safety and energy density. However, developing their core solid electrolytes and electrode materials requires testing numerous combinations of compositions and synthesis conditions. Traditionally, these experiments had to be performed manually inside inert-gas gloveboxes, a labor-intensive and time-consuming process. The new system integrates small collaborative robots with specialized modules to automate key equipment such as planetary ball mills.

According to AIST, the system can process up to 10 samples continuously with a single command (handling two samples per milling cycle across five cycles). It not only replicates researchers' experimental techniques but also centrally manages parameters like reagent amounts, milling speed, and processing time as digital data. This data can be directly utilized for Materials Informatics (MI) and Process Informatics (PI) analysis, accelerating AI-driven closed-loop material discovery.

Sources :aist.go.jp

Recherche

Acronis Discloses BossAI Attack Framework: Integrating 12 AI Agents and MCP Protocol Across 11 Countries

Sujet · BossAI攻击框架Matériel
2026-11-09 11:14 GMT+8

The Threat Research Unit (TRU) at cybersecurity firm Acronis has released a report detailing the operational mechanics of an attack framework named "BossAI." Operated by an attacker believed to be located in Hualien, Taiwan, China, the framework integrates 12 AI agents, nine of which were actively used during the observation period. These agents did not establish direct communication channels; instead, the attacker manually passed work outputs between them, using the Model Context Protocol (MCP) to chain penetration testing tools and executing attacks with locally hosted AI models stripped of safety restrictions.

In attacks targeting emergency services in Taiwan, China, the attacker first used HalberdStrike to scan websites and obtain vulnerability reports. These reports were then analyzed by Hermes for potential attack paths, with the results provided to Cursor for further exploitation. This division of labor indicates that while some tasks could be automated, coordination between different AI agents remained heavily dependent on human intervention.

Attackers also utilized "Skills" to preserve attack workflows and past experiences. For instance, a skill named gov-api-penetration, specifically designed for government fire dispatch APIs, was frequently invoked, accumulating 51 uses. These SKILL.md files recorded operational instructions, scripts, and reference materials, allowing AI agents to reuse existing outcomes across different sessions or even after switching underlying models. Additionally, configurations such as HERMES_YOLO_MODE=1 were enabled to allow agents to automatically approve operations and load hook scripts, reducing the need for manual oversight.

However, Acronis's research also highlighted reliability issues in AI agent vulnerability assessment. HalberdStrike misidentified connection failures as Remote Code Execution (RCE) vulnerabilities, and more than half of its reported high-risk vulnerabilities were false positives or misclassifications. This suggests that while AI agents can execute large-scale reconnaissance tasks quickly and at low cost, the resulting noise increases the burden on attackers, and higher automation does not necessarily equate to successful intrusion.

When target systems implemented effective authentication, session access controls, and network endpoint filtering, the AI agents failed to breach them. Acronis emphasized that the security weaknesses exposed in this incident were pre-existing; the significant development is that AI agents can discover these weaknesses and exfiltrate data faster and more systematically.

Sources :iThome

Recherche

Threat Intel Firm Vulone Discloses Large-Scale Cyberattacks on Asian Governments Using AI Agent Framework Cairn

Sujet · CairnStructurel
2026-10-20 07:14 GMT+8

Threat intelligence firm Vulone published an investigation report on October 18, 2026, revealing that a China-linked, financially motivated hacker organization utilized the open-source AI agent framework Cairn to launch large-scale network intrusions against government agencies and enterprises across multiple Asian countries between September 24 and October 14, 2026. The report indicates that the group conducted 66 attack operations, 48 of which successfully obtained access to target systems, covering regions including Taiwan, Vietnam, Indonesia, and India.

The core aspect attracting researchers' attention was the deployment of autonomous AI agents to execute actual intrusion tasks. The attack platform was built on the Cairn framework, where a central Dispatcher scheduled multiple Kali Linux-based containerized worker nodes according to set goals, leveraging Large Language Models (LLMs) to assist in system reconnaissance, exploitation, and post-intrusion data collection. According to Vulone's analysis, the model used by attackers was DeepSeek V4.1 Flash. The platform ran up to 116 containers simultaneously at peak times, logged approximately 230,000 events over 12 days, and consumed about 36.7 billion tokens cumulatively. To reduce AI computing costs, the attackers also rotated through six LLM subscription accounts and sixteen CodeBuddy accounts to obtain free compute credits.

Taiwan was one of the primary targets. Regarding government institutions, hackers exploited CGI command injection vulnerabilities to breach the legal system of Taiwan's Legislative Yuan, obtaining system command execution privileges. They also uploaded WebShells to Ministry of Justice-related systems via unauthenticated file management weaknesses. Additionally, attackers penetrated third-party vendor applications and conducted activities on outdated CentOS servers within the Investigation Bureau. On the enterprise side, the bulk messaging service platform Every8D became a key target; hackers exploited the Shellshock vulnerability (CVE-2014-6271) to compromise its mail gateway, stealing Apache host configurations and internal SQL Server database credentials. China Airlines' cargo-related systems faced up to 1,001 intrusion attempts, briefly resulting in JSP Shell access.

Vulone noted that this operation was not part of the previously reported "SOUL" campaign, although they shared the same exploitation engine. Attribution evidence points to a Chinese-speaking, financially motivated operator linked to China, though no direct evidence of state sponsorship was found in the examined artifacts. This discovery highlights the potential of AI agents in automated cyberattacks and the new security challenges facing critical infrastructure.

Sources :VULONE

Recherche

TechOrange Author Oshawn: Every Industry Is Becoming a Semiconductor Industry

Matériel
2026-11-09 11:22 GMT+8

Most companies will never design their own chips in their lifetime, but Oshawn's analysis at TechOrange points out that advancements in semiconductor technology are increasingly determining what products companies can create, how they innovate, and how they compete. The key to this shift lies in the changing deployment of AI: it no longer exists solely in data centers and the cloud, but is moving toward edge devices—vehicles, medical instruments, industrial machinery, consumer products, robots, and drones are all embedding increasingly sophisticated chips. This trend is driven by corporate demand for real-time responsiveness, lower latency, better privacy protection, and more energy-efficient computing.

The automotive industry offers the most obvious example. In the past, cars were defined primarily by mechanical engineering and horsepower; today, consumers evaluate vehicles based on driver-assistance features, in-car entertainment, connectivity, and the level of autonomous driving. As Oshawn writes, the numerous chips in vehicles are defining the products themselves, with computing power becoming as important as, if not more important than, engine performance. According to market estimates, the global edge AI market size was $24.91 billion in 2025 and is projected to grow to $29.98 billion in 2026, making the importance of chips unprecedented.

Similar changes are occurring in other industries. Data center infrastructure is shifting from one-size-fits-all computing architectures to those that can adjust according to workloads and energy needs; medical devices are becoming smarter and more connected, supporting AI-assisted imaging, robotic surgery, and portable diagnostics; industrial robots, which previously only executed fixed functions, are now expected to sense, analyze, learn, and continuously improve productivity.

As chips begin to define products, the source of competitive advantage also shifts downward. Oshawn notes that technical decisions made below the application layer are increasingly affecting AI performance, energy efficiency, security, cybersecurity, and user experience. Customers may not see these technologies, but they feel the results in product functionality every day. This has also changed how companies view technology ownership: where purchasing off-the-shelf components was once sufficient, vertical integration—where companies control more upstream and downstream links—is gradually becoming the norm.

In the data center sector, the race for custom chips is no longer limited to semiconductor companies like AMD, Intel, and NVIDIA; hyperscale cloud providers such as Amazon, Google, Meta, and Microsoft are also directly investing in chip R&D. The same applies to the automotive industry, where electric vehicle manufacturers like Tesla, BYD, XPeng, NIO, and Li Auto are competing based on software, computing power, and AI. Oshawn mentions that Apple's acquisition of chip design company P.A. Semi nearly 20 years ago was an early signal of this trend, which has continued to accelerate since then.

Sources :TechOrange 科技報橘

Recherche

AWS, Visa and Partners Build Stablecoin Payment Rails for Autonomous AI Agents

Sujet · AI代理稳定币支付Matériel
2026-11-09 11:22 GMT+8

As AI agents evolve from chatbots into autonomous task executors, their lack of independent payment capabilities has become a critical bottleneck. Traditional credit card systems, with their high fees and complex processes, are ill-suited for the hundreds or thousands of micro-transactions (often just cents) generated by an AI agent during a single task.

To address this, major technology and finance companies are accelerating the deployment of stablecoin-based "machine-to-machine" (M2M) payment infrastructure. According to TechOrange, Amazon Web Services (AWS) has co-developed Amazon Bedrock AgentCore Payments with Coinbase and Stripe. This service allows AI agents to autonomously access and pay for APIs, web content, and other AI agent operations by verifying their wallets and sending stablecoins. When an AI agent receives an HTTP 402 "Payment Required" response code, it can automatically complete the payment and continue its task without human intervention at checkout.

Additionally, Visa announced a transaction system combining AI, stablecoins, and tokenized payments, offering AI agent verification and transaction controls. Visa stated that its stablecoin settlement activities have reached an annualized run rate of approximately $7 billion, emphasizing the role of stablecoins in cross-border settlements and machine-level micro-payments.

Ken Nizam, co-founder of AsiaTokenFund Group, notes that the significance of stablecoins lies in providing programmable digital dollars that operate 24/7 on blockchain infrastructure with lower costs. However, permission management remains the biggest challenge for AI agent payments. Enterprises must use smart contracts to set strict spending limits, whitelists, and daily caps to prevent unauthorized fund usage. The maturation of this infrastructure could fundamentally change online content monetization and pave the way for autonomous payments in scenarios like IoT devices and self-driving cars.

Sources :TechOrange 科技報橘

Recherche
Suivant →
Connexion / Inscription
LectureRechercheRadarCadre d'investissement

For businesses, the logic is identical. Liu believes that great breakthroughs often do not end a problem but instead open up more new research directions; true AI transformation lies in achieving what was previously impossible, rather than counting how many Agents are deployed. He ultimately returns the question to a specific managerial judgment: Does your team have someone responsible for understanding? Without understanding, there can be no judgment; without judgment, decisions merely become signatures on a stack of outputs. AI can help us reach the moon faster, but those stars still require someone to look up and see them.

However, Oshawn also emphasizes that not every company needs to design its own chips. Companies can integrate commercial chips, commission custom designs, or co-design with partners, but none can treat semiconductor innovation as a technical decision completely unrelated to business. Meanwhile, the direction of innovation has become two-way: in the past, companies would identify business problems first and then seek solutions from suppliers; now, it is often advancements in semiconductor technology that create entirely new possibilities first, with companies racing to incorporate them into their products and services. In many cases, capabilities appear before innovation follows.

This makes semiconductor literacy a crucial leadership capability. Oshawn believes that executives do not need to become chip designers, but must possess sufficient understanding to grasp how chip advancements affect product strategy, customer expectations, long-term competitiveness, and operational performance. Many companies still treat semiconductors as an invisible layer handled by suppliers, but as products become increasingly intelligent and defined by digital technology, the risks of this mindset are growing. The winners of the next decade will clearly know where chips create advantages and where collaboration is smarter, placing semiconductor strategy at the core of overall corporate strategy.