LesenForschungRadarAnlageframework
Anmelden / Registrieren
AlleTechnikProdukteInvestierenBörseSonstiges

Lesen

2026-11-1673 Beiträge

Aktualisieren

Nvidia's CUDA Moat: How AI Code Porting Undermines Software Lock-In

Wesentlich
2026-11-16 19:00 GMT+8

The most formidable competitive barrier Nvidia built over the past decade was not its chips themselves, but CUDA—a software architecture that allows developers to easily write programs and build AI models for Nvidia hardware. Once developers invest significant engineering time into CUDA, the cost of migrating to other hardware platforms is extremely high, creating what is known as a 'moat.' In an analysis published on November 16, 2026, WIRED senior writer Will Knight points out that this moat is being eroded by the very technological wave Nvidia itself has driven: AI models have become proficient enough at reading and writing code to port AI models onto competitors' chips within hours.

A concrete example of this mechanism is Wafer, a startup company. Emilio Andere, CEO of Wafer, stated that whenever a major open-weight model is released, Wafer can get it running on AMD hardware at least 80% efficiently within hours. Andere noted that AI has become so capable that his engineers rarely need to edit or review the code. This means that manual porting work, which previously took weeks or even months, is now drastically compressed by AI automation. For Nvidia, which relies on the CUDA lock-in effect, the friction costs for developers to migrate are dropping sharply.

At the same time, the rise of AI agents has made inference—running already trained models—increasingly important. Patrick Moorhead, chief analyst at MoorInStrat, pointed out that the emergence of AI agents has heightened the importance of inference, which is precisely an area where Nvidia holds a weaker advantage compared to hyperscale cloud providers and startups. Google, Amazon, and Microsoft are all developing more specialized chips to drive training and serving services for AI models, each gaining similar market share increases. Although Nvidia acquired a specialized chip startup named Grok in December 2025, it still faces structural shortcomings in inference and general-purpose CPU domains.

Nvidia's contraction in the Chinese market constitutes another layer of pressure. According to Nvidia's own filings, China's revenue share has fallen from 20% in 2024 to 8% in 2026. Washington has restricted exports of the most advanced chips to China, while Beijing pushes for the development of its domestic chip industry. Huawei has developed new types of supercomputers that utilize many more chips working in coordination. Chinese AI company Z.ai claims to have fully trained its models using Huawei chips. If Chinese open-weight models gain adoption in other countries, they could create overseas markets for Huawei chips.

However, whether these pressures constitute an existential threat remains controversial. Moorhead acknowledges that Nvidia also possesses AI tools and is not lagging behind in AI chip design. Dan Ives, partner at Yorkville Ives, stated in a memo to investors that Nvidia remains about two to three years ahead of its nearest competitors, and its strategic roadmap, driven by 42,000 employees, will not change due to the founder's departure. But chip industry analyst Dan Hutchinson likened Nvidia's current situation to Intel's 'Bob Swan moment'—in 2019, after engineering-focused CEO Brian Krzanich resigned amid an employee relationship scandal, CFO Bob Swan took over, shifting the company's focus toward sales, leading to bureaucratic bloat and declining revenue. Hutchinson believes 'Nvidia is now an Intel waiting to happen.'

Quellen:WIRED

Forschung

SolidWorks CEO Manish Kumar: As AI Generates More Designs, Verifying Correctness Becomes More Valuable

Thema · SolidWorksWesentlich
2026-11-16 17:44 GMT+8

In an interview with a Korean press delegation in November 2026, Dassault Systèmes SolidWorks CEO Manish Kumar offered a clear assessment of the relationship between AI and engineers: AI will make design creation extremely easy, but as the number of outputs increases, proving which result is correct becomes increasingly important. He believes this is precisely where the value of engineers emerges.

This assessment is not a vague 'AI will not replace humans.' Manish Kumar provided specific mechanisms. He used the REACT (Reasoning and Acting) framework as an example to explain how AI works in the design process: when a user requests an engineering drawing, the AI faces nearly infinite combinations of possibilities regarding view selection, dimension placement, and section arrangement. The AI does not directly output a single answer; instead, it uses a reasoning model to decide 'why dimensions are placed in this view,' executes the action, observes the result, and adjusts as needed. This process itself is iterative and autonomous.

But here lies the key turning point: AI can execute this iterative process only if a human tells it what is required. Manish Kumar used the example of designing a lawnmower robot. Suppose an engineer receives a task: reduce the robot's weight by 15% without increasing costs while maintaining its fatigue life. This requires modifying the design, running simulations, and adhering to cost constraints. He said that most engineers know how to do this, but it takes time because it involves numerous commands, clicks, and procedural steps.

AI can take over these steps. However, Manish Kumar argues that the core value of engineers has never been in 'clicking commands,' but in 'finding the correct intent'—defining goals, setting constraints on what the model should and should not do, and then evaluating the pros and cons based on the results provided by AI.

He compares this relationship to a driver and an engine: the engineer sits in the driver's seat, and AI is the engine. The engine makes the car move, but the direction and destination are the driver's responsibility.

This argument has a specific product context. This year, Dassault Systèmes is advancing SolidWorks' generative interactive AI to the 'agentic AI' level: rather than simple Q&A, humans provide specific tasks, and multiple specialized AI agents complete the work in the background according to learned processes. This system includes three 'virtual companions'—Aura, Leo, and Marie—each corresponding to different professional domains. Leo possesses 'Competencies' such as mechanical design, performance analysis, and mechanical programming; each competency is further subdivided into 'Skills,' such as conceptual design, detailed design, and drafting.

On the commercial front, Dassault Systèmes provides users who purchased SolidWorks in the past two years with 2,100 Blue Tokens per person for AI features. Each Token is priced at five cents, deducted based on executed skills and output value rather than text volume. Manish Kumar stated that within five months of commercialization, customers had already used millions of tokens, citing one customer who used the generative drawing function to increase dataset productivity by 50-60%, with a goal of reaching 90% within six months.

Quellen:바이라인네트워크

Forschung

The Emission Reduction Impact of Net-Zero Commitments: Three Years of Data from 1,200 Major Companies Shows No Significant Effect

Thema · 净零承诺减排效果Wesentlich
2026-11-16 15:01 GMT+8

At the end of 2019, only 70 out of roughly 2,000 major global companies had committed to long-term net-zero targets. By the end of 2024, this number had risen to more than 1,200 companies, accounting for two-thirds of the combined market capitalization of the sample. The proliferation of net-zero commitments has been astonishing, but have these commitments themselves led to measurable emission reductions? A study published in 2026 by Simon Dietz and Nikolaus Hastreiter provides a statistical test of this question.

The researchers compared changes in absolute emissions and emission intensity before and after companies adopted net-zero targets. In terms of direction, the results are encouraging: three years after adopting a target, average absolute emissions decreased by about 15%, and average emission intensity dropped by nearly 10%. The trend is indeed moving in the right direction. However, the magnitude of this change is insufficient to reach statistical significance. In other words, we cannot rule out the possibility that these declines were driven by random fluctuations or other concurrent factors.

For investors, this means that the environmental impact of current corporate net-zero commitments lacks rigorous statistical evidence. The spread of commitments alone does not equate to actual emission reductions. Of course, this conclusion has a clear temporal boundary: net-zero targets are typically set for 2040 to 2050, and the study only observed data for three years post-adoption. As author Joachim Klement pointed out, as emission reductions increase over time, statistical significance may change; but at least for now, we cannot conclude that net-zero commitments have significantly reduced greenhouse gas emissions.

Quellen:klementoninvesting.substack.com

Forschung

‘Big Short’ Michael Burry Warns Investors to Sell Nvidia Stock After Jensen Huang Resignation

Wesentlich
2026-11-16 18:16 GMT+8

Michael Burry, the investor famous for his role in the 'Big Short', publicly warned investors on November 16, 2026, to sell their shares in Nvidia, arguing that the true crash has yet to begin. According to Finbold, Burry stated that the resignation of former CEO Jensen Huang has removed the key driver behind the high demand for the semiconductor giant's chips. He claimed that the leadership change will enable customers to slow down deals, renegotiate terms, or switch to competitors for the first time in years.

Additionally, Burry cited a report by Culper Research which claims that at least 20% of Nvidia's annual revenue comes from sales in China, contradicting Huang's previous assertions that the firm's market share in the country is close to zero. With Nvidia's stock already down 31.93% from its October 6 highs and erasing $1.84 trillion in market capitalization, Burry believes this fundamental shift, combined with potential scrutiny over export restriction evasion, could trigger a catastrophic chain reaction.

Quellen:Finbold

Forschung

Greenpeace Activist Benjamin Stephan: Why AI Expansion Accountability Points to Nvidia

Thema · 绿色和平问责英伟达Wesentlich
2026-11-16 18:00 GMT+8

The expansion of AI data centers is encountering increasingly strong social resistance. Data Center Knowledge reports that residents, local protest groups, and large organizations are all expressing dissatisfaction with data centers. Greenpeace has chosen a path different from typical community opposition: holding chip supplier Nvidia directly accountable. In an interview with Data Center Knowledge, Benjamin Stephan, Greenpeace International’s Big Tech campaigner, fully elaborated on the logic behind this strategic choice.

Stephan’s argument does not start with the broad criticism that 'data centers consume too much electricity,' but rather focuses on Nvidia’s unique position within the economic chain of AI infrastructure. He points out that Nvidia now does more than just sell chips; it organizes financing models that make it easier for major tech companies to secure funding to build new data centers and purchase its chips, further fueling unchecked data center expansion. In Stephan’s view, this means Nvidia has transformed from a mere supplier into a driver and beneficiary of expansion, making it an accountable node.

Another link in the accountability chain comes from the supply chain. Stephan notes that Nvidia’s supply chain largely runs on fossil fuels, exacerbating the climate crisis while creating significant hardship for communities in Taiwan, where Nvidia relies on TSMC to manufacture its chips. He argues that given Nvidia’s profit scale, switching its supply chain to renewable energy would cost only a fraction of its earnings.

The third link involves military use. Stephan cites reports that Nvidia chips have been found in Russian autonomous drones deployed in Ukraine, deeming this unacceptable. He demands that Nvidia implement independent audits across all global distributors, subsidiaries, and third-party vendors to ensure that no Nvidia technology is used to violate human rights or diverted into military equipment and autonomous weapons.

These three threads converge on a specific event: the Taipei promise. Stephan recalls that at a previous GTC conference in Taipei, Greenpeace East Asia activists presented Huang with a five-layer cake topped with wind turbines, borrowing his metaphor that 'AI is a five-layered cake' to point out that AI and Nvidia were missing the layer of renewable energy. When asked if he could invest in renewable energy together with Taiwanese suppliers, Huang gave a positive response. Stephan said, 'We were in Berlin to hold him accountable because nothing had happened.'

The specifics of the Berlin protest also corroborate this accountability logic. Stephan, who was present as a spokesperson, waited outside by Huang’s car after the CEO left the venue, informing security that Huang had said on stage they would 'talk later.' However, the vehicle drove off empty, and the conversation did not take place. Stephan emphasized that this was not an action organized by his group; as a Berlin-based activist, he was simply seeking a direct dialogue in his capacity as a spokesperson.

Quellen:DataCenterKnowledge

Forschung

YC Partner Ankit Gupta: In the AI Era, the Value of Technical Training Is Not Writing Code, But Asking Models More Interesting Questions

Wesentlich
2026-11-16 17:40 GMT+8

YC Partner Ankit Gupta recently participated in Y Combinator Startup School activities in Taiwan and gave an interview to INSIDE, systematically outlining his views on the skills required of founders in the AI era. His core judgment is that AI is changing the competitive landscape between startups and large enterprises, and this shift is particularly favorable to those with solid technical training.

Gupta believes that large enterprises with thousands of employees find it difficult to completely overhaul their existing operations and decision-making processes to fully harness the power of AI models. Some companies provide employees with ChatGPT accounts but may not have truly redesigned their ways of working. In contrast, two or three technically proficient individuals familiar with AI tools can now rapidly develop complex products that would have taken much longer to complete in the past. He suggests that some CEOs of large enterprises are currently worried about competitors who may not be other large corporations, but rather small technical teams capable of solving problems in entirely new ways.

What is the mechanism behind this judgment? Gupta offers a definition of 'technically minded' that is broader than having a degree but stricter than simply 'knowing how to use tools': it means you can build what comes to mind. Previously, this required being an engineer; now, 'vibe coding' might suffice. However, he immediately adds a key observation: people with solid technical training, on average, are able to ask more interesting questions of models.

He further explains why he still recommends hard disciplines like computer science: the fundamental concepts taught in school are not software engineering—which is mostly learned on the job—but rather how to think. Taking courses in data structures and algorithms teaches logical thinking, how to build things, and how computers work at a low level, thereby revealing what is possible. His advice is to do both well: combine rigorous academic training with practical hands-on experimentation using models.

However, Gupta explicitly acknowledges that this is not an absolute rule. He states that YC invests in many non-technical founders who often possess deep domain knowledge and can use AI models to create attractive and genuinely effective products. He hopes AI will ultimately bridge the gap, allowing more people to turn ideas into reality and become great founders.

At the business model level, Gupta uses his experience founding an AI drug discovery company to illustrate a more specific change: instead of developing software and trying to persuade large pharmaceutical companies to change their existing R&D workflows, startups can directly use AI to develop products. He believes AI is giving entrepreneurs the opportunity to bypass some traditional software sales models and enter existing industries directly for competition.

During the same Startup School event, YC Operating Partner Jon Xu supplemented the discussion on finding entrepreneurial opportunities from another angle: besides observing market demand, one can also study tasks that current AI models are not yet good at completing. He specifically mentioned that models currently have shortcomings in basic technical tasks such as hardware design and chip verification, which, for engineers familiar with these fields, represent potential entrepreneurial opportunities worth exploring. YC President and CEO Garry Tan advised hard-tech startups with higher capital needs to work backward from their ultimate goals, identify technical milestones that can be achieved first, and secure subsequent funding through phased results.

Quellen:INSIDE

Forschung

Ambarella Technical Marketing Engineer Cicalese: Perception Reports Must Explicitly Retain Position, Time, Uncertainty, and Provenance

Thema · AmbarellaWesentlich
2026-11-16 14:09 GMT+8

Ambarella Technical Marketing Engineer Pietro Antonio Cicalese proposed a warehouse inspection scenario: one robot reports a fallen cardboard box in Aisle 14, and twenty seconds later, a second robot approaches from the opposite end of the aisle, reporting an obstruction near the entrance. The server now faces an awkward question—are the two robots seeing the same box? If treated as two separate obstacles, it may close more aisles than necessary; if merged into one, it might miss a second, actual obstruction. Cicalese points out that both descriptions can be accurate, yet the coordinator still cannot make a useful distinction.

He argues that the root of the problem lies in how information is handed off from perception to application. The phrase “a fallen box is blocking Aisle 14” is useful for a supervisor glancing at an alert, but for a coordinator comparing reports, it leaves substantial work unresolved: Where along the aisle is the box? How much space does it occupy? Did the second robot see it recently, or did its report spend time waiting for a connection while becoming stale? An accompanying image can help answer some of those questions, provided the recipient can interpret its geometry and capture conditions; a language model may also extract a richer description. But Cicalese emphasizes that neither approach automatically supplies a shared position nor establishes that two observations concern the same object—those relationships need to be represented and evaluated explicitly.

Therefore, the handoff requires an agreed specification between parties: what information is provided, and what behavior is expected when it is incomplete. Software engineers often call this an interface contract. Its value is practical: the receiving application should be able to distinguish a valid position from an unknown one without guessing the sender’s intent. For the warehouse, the report needs to preserve the object’s location and occupied area in a named coordinate frame, with stated units. It also needs the capture time, separate indications of uncertainty across dimensions, and a way to trace the report back to its source observation. A readable summary can sit alongside that structured information for the supervisor to quickly review.

Position conversion itself carries error. Detectors commonly locate the box within a camera image; those pixel coordinates describe where the box appears on the image plane. The coordinator needs a location on the warehouse map, and the conversion requires more than knowing where the robot is standing. A range measurement or a valid geometric constraint (such as a calibrated floor plane) is needed to locate the relevant points in three-dimensional space. The robot’s estimated position and orientation at capture time allow those points to be transformed into the map frame. The box’s extent also needs suitable geometry; the apparent width of its image alone does not establish the space it occupies on the floor.

Cicalese cites the ROS project’s REP 103 standard, noting that inconsistent units and coordinate conventions are precisely what cause integration errors. A camera’s optical frame and a robot’s body frame use different axis orientations; naming the frame and documenting the transformation prevents two components from interpreting the same numbers as different physical locations. Once both robots report in a shared map frame, the server can compare the estimated occupied regions. Overlapping locations and compatible dimensions provide evidence for a common obstacle, but they support an association hypothesis rather than proof of identity. Two similar boxes could be close together, or a worker could have moved one between observations. When evidence is insufficient to justify merging, the coordinator should retain unresolved alternatives.

Quellen:Ambarella

Forschung

Independent Technology Analyst Richard Windsor: Megaspeed Probe Unlikely to Pose Substantive Regulatory Risk for Nvidia

Wesentlich
2026-11-16 14:51 GMT+8

Independent technology analyst Richard Windsor published his assessment on Radio Free Mobile, stating that the Megaspeed investigation is unlikely to deliver a direct regulatory blow to Nvidia. After spending 11 years covering global technology equities at Nomura Securities, he founded this independent research firm. He currently holds a half-position in Nvidia and plans to increase his stake.

Windsor’s core argument is legal in nature: current United States export controls do not prohibit Chinese companies from renting computing power hosted in overseas data centers located in neighboring countries, provided four conditions are met—the data center is not owned by Chinese entities, the Chinese tenant does not take ownership of the facility, the chips are not subsequently shipped into China, and the AI trained is not used for military purposes. Under this framework, it would be entirely legal for a data center owned by a Malaysian telecommunications company to upgrade to Nvidia chips and sell computing capacity to Alibaba.

He cites conclusions from Megaspeed, Inspur, Nvidia’s internal audit, and Bloomberg’s 2025 investigation, noting there is no evidence that any of the 136,000 chips sold to Megaspeed were re-exported to China. While there are signs, inconsistencies, and rumors, there is no hard evidence in this specific instance. Windsor acknowledges that some banned Nvidia equipment has indeed entered China, but these are primarily smaller, portable DGX solutions rather than the massive NVL72 systems. Companies such as SuperMicro have been sued or pleaded guilty to smuggling Nvidia chips, but Megaspeed is not among them.

Windsor contends that the collective media siege against Nvidia following the leak of the memo is biased, making objective assessment difficult. He judges that the only verifiable offense is the leaking of the classified memo itself; everything else remains conjecture. If a large-scale, successful smuggling operation were underway, China would have no need to adopt the complex and risky path of renting overseas data centers.

Regarding the actual impact on Nvidia, Windsor’s reasoning chain is as follows: the outcome of the Megaspeed investigation determines whether Nvidia gets dragged into a compliance quagmire. If Megaspeed is found to have engaged in smuggling, Nvidia and CEO Jensen Huang would appear negligent, and the resulting regulatory burden would impair their ability to rapidly launch new products—precisely the foundation of their long-term competitiveness. However, he believes the probability leans toward avoiding direct penalties, with the most likely outcome being further tightening of chip export restrictions and stricter rules on renting overseas data centers, rather than enforcement action specifically targeting Nvidia.

Windsor links this judgment to valuation: at a share price of $160, this risk appears fully priced in, giving Nvidia at least one year to navigate any rule changes before financial performance suffers. He also notes that Jensen Huang’s unique management style is the bedrock of Nvidia’s competitive advantage; while this personnel risk remains, it constitutes a separate issue.

Quellen:radiofreemobile.com

Forschung

Sierra Releases Personal Agent Protocol (Poppy) 1.0; Shopify and Meta Muse Adopt Standard

Wesentlich
2026-11-16 15:29 GMT+8

Sierra, a company developing enterprise AI agents, announced on November 13, 2026, the public release of version 1.0 of the Personal Agent Protocol (commonly known as Poppy). This open standard defines how personal AI agents securely interact with corporate websites, APIs, and enterprise-side agents on behalf of users. According to gihyo.jp, e-commerce giant Shopify has enabled automatic support for the protocol across all its stores, while Meta's AI agent Muse has made Poppy available to all its users.

The core of the Poppy protocol addresses the challenges of identity verification and permission control when AI agents access business services. Under the released 1.0 specification, companies must publish a /.well-known/poppy.json file on their domain detailing the endpoints and functions available to agents. Personal AI agents read this file to choose connection methods, including web navigation, API calls (supporting MCP and OpenAPI standards), or dialogue with the company's own agent. For account access, the protocol relies on OAuth, allowing users to grant specific read-only or write permissions and revoke them later.

The update also introduces an extension mechanism, notably the Universal Commerce Protocol (UCP) extension co-developed with Shopify. This extension combines Poppy's connection discovery with UCP's handling of product search, cart management, and payments, enabling AI agents to shop as guests without logging in or to use customer accounts with explicit user authorization. Additionally, Sierra provided an implementation kit for enterprises, containing skills documentation to guide coding agents through system adaptation and validation tools.

Beyond Shopify and Meta, financial services firm Mercury, Turkish Airlines, and Synchrony have begun adopting Poppy. For instance, Muse users can now query Mercury account balances or categorize expenses via the protocol, though transfers require approval based on bank rules. Turkish Airlines allows users to check flight status and mileage upgrade options through agents. Currently, over 100 companies, including OpenAI, Stripe, Visa, Walmart, and Salesforce, are listed as partners in the standard's development.

Quellen:gihyo.jp

Forschung

AI Infrastructure Startup Mount Thor Secures $45 Million to Build Mac Data Centers for AI Agents

Thema · Mount ThorWesentlich
2026-11-16 09:05 GMT+8

AI infrastructure startup Mount Thor has announced it has secured $45 million in equity funding to build data centers running racks of Apple computers, specifically designed to support the development and deployment of artificial intelligence agents. Founded by former OpenAI employee Peter Voell, the company is headquartered in San Francisco, United States.

As AI agents increasingly need to navigate desktop applications, manage files, and perform tasks dependent on specific operating system environments, traditional Linux cloud servers often fall short. Mount Thor's service provides remote access to full macOS environments, sparing customers the expense and operational complexity of maintaining extensive Mac fleets themselves.

The financing spans Seed and Series A rounds. CRV and Conviction co-led the Series A, while Chad Byers and Offline Ventures co-led the Seed. CSC Leasing, Enea Capital, Long Journey Ventures, and Upper90 also participated. Additionally, the company secured non-dilutive capital to support hardware purchases.

Mount Thor stated that its first data center was deployed within six months. The new capital will be used to expand computing capacity, hire staff, and establish a laboratory to test inference capabilities of open-source AI models optimized for Apple processors. The company argues that Apple's long-standing investment in accessibility and human-computer interaction makes macOS particularly suitable for AI agents requiring human supervision.

Quellen:TechRadar

Forschung

Fed Report: AI Data Center Debt Issuance Up 5x in Three Years, Off-Balance-Sheet Commitments Exceed $3 Trillion

Thema · AI基建债务风险Strukturell
2026-11-16 15:36 GMT+8

The U.S. Federal Reserve released its semi-annual Financial Stability Report on November 13, 2026. In surveys conducted by the New York Fed between September and October with 23 market participants including brokers, banks, and investment funds, artificial intelligence (AI) emerged as the most frequently cited risk. Nearly 90% of respondents mentioned AI, compared to about 50% in the spring survey. The report notes that these findings reflect respondents' views and do not represent the official position of the Federal Reserve.

A dedicated section of the report analyzes debt issued for advanced computing infrastructure. It states that such debt issuance exceeded $680 billion between January and October 2026, roughly five times the total for all of 2023 ($130 billion). This borrowing now accounts for approximately 31% of non-financial investment-grade corporate issuance, up from less than 6% in 2023. As of October, outstanding public market debt related to advanced computing was estimated at over $1.1 trillion.

Beyond on-balance-sheet debt, the report reveals significantly larger off-balance-sheet commitments. Seven major tech companies—Amazon, Alphabet, Meta, Microsoft, Oracle, Broadcom, and NVIDIA—have undiscounted amounts exceeding $1 trillion in leases not yet commenced, plus over $2 trillion in other commitments such as take-or-pay contracts and residual value guarantees. These commitments link high-rated hyperscalers with weaker-credit developers and GPU manufacturers, creating complex "circular financing" dynamics.

Regarding risk distribution, mutual funds, pension funds, and insurance companies primarily hold the on-balance-sheet debt of hyperscalers and GPU firms. However, higher-risk subordinated debt and equity positions are held by high-yield bond funds, life insurers, and business development companies (BDCs). Additionally, the largest U.S. bank holding companies have nearly $1 trillion in exposure to large tech firms, data center developers, and related special purpose entities (SPEs), which will be assessed in annual regulatory stress tests.

Quellen:TechOrange 科技報橘

Forschung

Survey: One-third of Japanese BtoB Marketers Use AI Agents Continuously; Success Linked to Customer Understanding

Kurzfassung
Verifiziert 2026-11-16 16:38 GMT+8

Market research firm PRIZMA released a survey on November 16, 2026, targeting 1,007 Japanese BtoB marketing professionals. The data shows that 33.4% of respondents are continuously using AI agents. When including those who tried and stopped, approximately 40% have experimented with the technology in their work.

Results regarding effectiveness vary. Among experienced users, 33.2% reported an increase in valid sales opportunities, while 48.1% stated there was "almost no change." This indicates that simply adopting the tools does not guarantee business growth.

The survey highlights time reallocation as a key differentiator. Compared to groups where deal volume remained flat or decreased, those seeing growth were more likely to spend time saved by AI on "customer interviews and market understanding" (31.3% vs. 9.1%) and "creating unique information and case studies" (44.5% vs. 25.9%). In contrast, teams without significant gains tended to use the freed-up time to increase the volume of existing activities.

The survey was conducted by PRIZMA, part of the Japanese video content marketing company Mujin, targeting marketing staff at companies with 10 or more employees. The data reflects a trend in the early stages of generative AI adoption in Japanese BtoB sectors: efficiency gains from technology must be converted into deeper customer insights to translate into commercial results.

Quellen:プレスリリース・ニュースリリース配信シェアNo.1|PR TIMES

Forschung
Weiter →
Anmelden / Registrieren
LesenForschungRadarAnlageframework

AMD's market share rose from 2% in 2022 to between 7% and 10% in 2026, while Nvidia still accounts for 80% of sales in the AI accelerator market. Nvidia's revenue grew from $26.91 billion in January 2022 to $215.94 billion in January 2026. These figures indicate that Nvidia's lead remains massive, but the loosening of the CUDA moat, the shift in inference workloads, and the shrinkage of the Chinese market are structural forces that were already operating before Jensen Huang's personal controversy.

These figures come from vendor self-reports and a single customer case, lacking publicly reproducible benchmarks. However, Manish Kumar's core argument does not depend on these numbers: when generation costs approach zero, the scarce resource shifts from 'being able to create designs' to 'being able to judge which design is correct.' This logic applies to any engineering field using AI generation tools, regardless of the specific product.

The boundary that must be preserved is: Manish Kumar is the CEO of the SolidWorks product, and his discourse serves as both a technical viewpoint and a commercial positioning. His premise that 'AI makes design easier' assumes that their own tools can indeed achieve this—there is currently no independent verification indicating that the reliability of generative AI under complex engineering constraints has reached the level of fully replacing human execution. His argument describes a direction, not a completed fact.

Based on this argument, Stephan put forward two specific demands: First, Nvidia must commit to achieving 100% renewable energy across its supply chain and data investments by 2030, ceasing reliance on fossil fuels and nuclear power. Second, it must invest in new renewable energy capacity globally, particularly in key supplier regions like Taiwan and in areas where Nvidia invests heavily, such as Southeast Asia.

The boundaries of this accountability logic are also worth noting. In its 2026 Sustainability Report, Nvidia stated that it had matched 100% of its global electricity consumption with clean energy for the second consecutive year, and noted it is designing platforms like Vera Rubin to advance powerful computing alongside sustainability. Regarding chips flowing into Russian drones, Nvidia stated that the relevant products are consumer-grade systems intended for students, developers, and startups. Stephan’s argument positions Nvidia as a bottleneck node in the entire AI ecosystem, but there is debate over how much actual control chip suppliers have over how downstream customers deploy infrastructure—this is precisely the weakest point in his accountability framework.

Stephan stated that the Berlin protest is just part of Greenpeace’s new global Big Tech campaign. Just days after the GTC protest, Greenpeace activists joined the Stop Dirty Data Centres coalition, conducting two days of localized protests in Europe and around the world. He explicitly stated, 'Peaceful protest and nonviolent direct action are part of Greenpeace’s DNA.'

For the AI industry, the value of this interview lies not merely in recording another protest, but in presenting a specific accountability framework: when a company acts simultaneously as a technology supplier, financing driver, supply chain organizer, and public promise-maker, environmental groups choose to concentrate pressure on a single node rather than dispersing it across the entire industrial chain. This strategic choice itself offers a lens through which to understand issues of social acceptance regarding AI infrastructure.

Gupta’s entire argument has a direct implication for AI industry investment: when evaluating an AI startup team, 'knowing how to use AI tools' may not be the most critical screening criterion; rather, 'the ability to ask deep questions of the model' is. The latter has a statistical positive correlation with the founder's technical training background, although it is not a sufficient condition.

It should be noted that these are attributive viewpoints from YC partners, not independently verified research conclusions. Gupta did not provide quantitative data to support the effect size of the observation that 'technical training leads to better questions,' nor did he specify on which concrete tasks this advantage is most pronounced. His judgments on competitive dynamics are similarly based on qualitative observations from YC’s portfolio rather than systematic cross-industry comparisons.

Time also needs to be distinguished. Capture time tells the receiver when the scene was observed; processing completion and message arrival describe later events in the pipeline. They remain useful for diagnosing delays, but using arrival time as the observation time can make an old scene appear current. Comparing capture times across robots requires clocks with a known relationship. If clock error is large relative to the event being analyzed, apparent timing differences may be misleading. The contract should make the time reference clear and define when an observation has become too old for a particular decision—an obstacle report may remain useful for scheduling after it has become unsuitable for deciding whether a narrow route is currently clear.

Uncertainty also needs more than a single score. The system may be confident that an object is a carton while having a poor estimate of its distance. A precise depth reading may accompany an uncertain object label. Glare may affect part of an image without invalidating every measurement from it. Combining all of these conditions into one confidence value leaves the receiving application unable to respond to their different consequences. For inventory, this distinction is especially tangible: a shelf label can establish which SKU belongs at a location, but it cannot reveal cartons hidden behind an obstruction. The count should cover what the robot actually observed; unobserved portions remain unknown, potentially requiring another observation or manual count to complete the task.

Provenance makes such questions possible to investigate. The report can identify the robot, sensor, and source observation, along with the processing version used to produce it. Where access and retention rules permit, a reference can lead to an image crop or recorded frame. That evidence need not travel with every message, but its availability and eventual expiry should be understood by the recipient.

As more sensors join, more measurements need reconciling. A camera may suggest the object’s type, while radar contributes range and radial velocity. Those readings can be combined only with appropriate spatial calibration, temporal alignment, and an assessment of whether they concern the same target. A radar return with no reliable camera match can remain an object of unknown type; inventing a label would conceal the disagreement. Fusion can take place at different stages, depending on what the downstream task needs and what each processor can handle. Cicalese cites Jason Huang’s account of Ambarella’s radar demonstration: a CV3 system-on-chip in the front radar unit processes data from five Oculii radar heads, then sends a combined point cloud to a zone controller for sensor fusion. A point cloud represents individual detections by their positions in three-dimensional space. That design keeps multi-radar processing together and gives the next stage a spatial representation. Other fusion tasks may need a different boundary, depending on which measurements the receiving stage requires.

Finally, Cicalese proposes that the interface contract should be tested by replaying reports from known scenes: starting with two observations of one box, then using two nearby boxes with similar appearances; introducing an old report after a newer one; removing the range estimate, changing the coordinate frame, and simulating a clock offset. Expected results should include how the coordinator handles uncertainty in each case. A useful test checks the downstream decision: if two reports cannot be associated confidently, the coordinator should preserve that ambiguity and follow the specified scheduling behavior, such as requesting another observation. A missing distance must not become zero, and a region hidden from the camera must not become clear space. Passing a message through a parser establishes much less than this.

When two robots report from opposite ends of Aisle 14, the server should have enough evidence to make a justified association—or enough information to recognize that another observation is needed. The coordinator’s assignment remains a proposal evaluated by the robot application runtime; motion remains subject to the machine’s separately validated protective control constraints. A well-designed report improves coordination without taking over those responsibilities.

This analysis distinguishes between the different legal statuses of overseas computing rentals versus chip smuggling under current regulations, rather than relying solely on stock price fluctuations or media sentiment. Its limitations are obvious: the author holds a position, and his conclusion relies entirely on the current legal framework not being retroactively tightened or reinterpreted. Windsor himself admits that China holds negotiation leverage, including rare earth minerals and agricultural purchases, meaning the ultimate trajectory of export controls remains uncertain.