Can Industry Research Really Predict the Future?

Industry researchers are often asked to predict the future: next quarter’s market share, five-year growth trajectories, the next destination in the global supply chain. But are such expectations realistic? Without systems for timely feedback, institutional validation, or long-term credibility building, can industry analysis truly bear the burden of forecasting?

This essay reframes the issue from a structural perspective. It argues that the difficulty in making accurate predictions stems not from a lack of skill or effort, but from the absence of institutions capable of supporting, verifying, or rewarding such predictions. In this context, the real value of research may not lie in calling future events. It may instead reside in identifying early misalignments between belief and reality, and in preserving records of those invisible fractures before they surface.

When no system exists to reward or remember, the role of the researcher shifts. We are not prophets, but witnesses. We leave behind observations not because they are guaranteed to be remembered, but because someone, somewhere, will need them when the narrative begins to turn.

In the course of doing research, we are often asked questions like:

“What do you think this company will look like five years from now?”

“Do you expect this industry to reverse course next quarter?”

“How much longer can Taiwan hold its position in this supply chain?”

These questions are hard to avoid. In fact, they seem perfectly natural. After all, we have grown used to thinking of research as a way to forecast the future. It often feels as if the ability to see ahead is the ultimate source of value.

But this article begins from a different place.

It is not about the accuracy of investment models or the precision of specific forecasts.

While industry research is often referenced by investors and can influence capital flows, our focus here is not on returns. It is on something else:

When forecasts cannot be institutionalized, and when there is no system for validation or feedback, is industry research still worth doing? And if so, what kind of value does it leave behind?

This article tries to answer that question by asking something deeper:

In a world where systems fall short, how can researchers find their place and understand their responsibility?

1.  Why Are We So Drawn to Prediction?

Across industries, in investment circles, and even in media and academia, there is a persistent obsession with forecasting the future.

What will the market share look like next quarter?

Can this company double its growth over the next five years?

Which country will the supply chain move to next?

We often hope that industry analysis can offer answers as precise as a weather forecast. The expectation seems reasonable. After all, the more data we have and the more sophisticated the models become, the more accurate our predictions should be.

But in reality, moments of true predictive clarity are rare.

If we take an honest look at how industry analysis works in practice, we often find that forecasts are vague, tentative, and filled with assumptions. This isn’t because researchers aren’t trying hard enough. It is because the environment they operate in has never been built to reward precision in prediction.

2.  The Trouble with Prediction Is Really a Problem of Institutions

The difficulty of industry research is not just a matter of technical limitations. It is also a consequence of institutional gaps.

We do not have systems in place that allow predictions to be received, verified, or translated into lasting credibility.

• No feedback or verification mechanism: Unlike financial markets where price serves as real-time feedback, industry forecasts are rarely evaluated. No one is held accountable for being right or wrong in a measurable way.
• No space for revision or reputation-building: Most industry reports end once they are published. There are few opportunities to revisit, revise, or track their accuracy over time. Even if a prediction turns out to be correct, it is hard to prove that the research got it right.
• A mismatch between forecasting timelines and institutional expectations: Many forecasts aim to capture trends over three to five years. But institutions and markets often expect results on a quarterly or even monthly basis. This misalignment marginalizes long-term observations and makes it difficult for them to carry weight.

Some have suggested using crowdsourcing or prediction markets to close these gaps. Even in areas with high information flow and strong incentives, such as finance or elections, these mechanisms remain difficult to implement. In industry research, which is far less structured, they are even harder to sustain.

And so we return to a central question:

Without institutional support, are we still making predictions at all?

Or are we actually doing something else entirely?

3.  If We Can’t Predict Events, What Can We Do Instead?

Perhaps it’s time to let go of the expectation that industry research should predict specific events. Instead, we can begin to see its role in a different light. The value of research may not lie in telling us what will happen next, but in helping us see where the current structures are starting to show signs of strain or misalignment.

This way of seeing is closer to George Soros’s theory of reflexivity:

• Markets reflect not reality itself, but the beliefs shared by many.
• When those beliefs drift too far from reality, that’s when reversals tend to occur.
• What matters most is not the exact timing of the reversal, but the ability to notice the divergence early.

From this perspective, industry research doesn’t need to promise precision.

Instead, it should focus on recognizing when the market starts to believe in a story that may never come true.

As we saw in the case of Wolfspeed, trust collapsed before the industry fundamentals did. And in Broadcom’s story, structural consistency allowed the company to maintain credibility without leaning on exaggerated narratives.

Will the market eventually correct this misalignment? There is no way to know for sure. But if it does, the shift may come faster than expected.

4.  Outside the System: The Researcher’s Role and Responsibility

Some may ask: if predictions lack institutional support and cannot be verified or reinforced, what is it that researchers are still doing?

This, I believe, is precisely where the researcher’s role becomes clearest.

We are not prophets of the market. We are witnesses and quiet observers of the narratives that shape it.

Our responsibility has never been to predict the most accurately. Rather, it is to ask:

• Can we recognize the break between belief and reality before others do?
• Can we remember what the supply chain used to look like, and explain why the narrative turned when it did?
• Can we remain that steady pair of eyes when institutions grow short-sighted?

This kind of work is not rewarded by the market. When capital retreats, narratives collapse, and systems are rewritten, only a few people will look back and search for those who once spoke with clarity and remembered the details.

The value of research does not lie in predicting future numbers. It lies in preserving our sensitivity to change and our understanding of structure.

These observations may never be fully acknowledged by formal systems. But perhaps that is what allows them to endure.

Extended Conclusion: If Prediction Fails, What Remains of Industry Analysis?

If there is anything you choose to take away from this piece, perhaps it could be these four layers of reflection:

1.  At the level of knowledge: Understanding why predictive systems struggle to take root

You might see more clearly that the difficulty of institutionalizing industry forecasts does not stem from a lack of analytical effort. Rather, it comes from the absence of a foundation that can hold judgments, verify perspectives, and build trust over time.The issue is not that predictions are too weak, but that systems are too shallow.

2.  At the level of method: Reframing what we expect from research

The value of research has never been about precision in prediction. It lies in recognizing when belief and reality begin to drift apart. What matters is not who made the most accurate call, but who first noticed the fracture forming.

3.  At the level of reflection: Rethinking the role of the researcher

For those who do this work, this essay may serve as a quiet reminder. Even when systems offer no feedback and our judgments go untested, we can still be the ones who remember the structure and can explain why the narrative shifted. This may not earn rewards from the market, but it may be remembered by a few who matter, over the long term.

4.  At the level of worldview: On systems, trust, and the flow of knowledge

Finally, you might begin to ask different questions. What kind of knowledge is worth preserving? How is knowledge really accumulated? When systems cannot hold truth, are we still willing to remain observers?

If no one is tasked with judgment, then what we leave behind are small and persistent traces. These are observations that continue to be recorded and quietly passed along.

We do not know if they will be remembered. They may fade into the background, or one day be rediscovered in a moment no one expected.

This is what research looks like when there is no system to respond. It is lonely. But it may also be the most honest form it can take.

We leave these notes behind because, perhaps, you will be the one who finds them.

This article is part of our Future Scenarios and Design series.
It explores how possible futures take shape through trend analysis, strategic foresight, and scenario thinking, including shifts in technology, consumption, infrastructure, and business models.

See more in this category, or explore more notes here.

Why Do Good Companies Struggle to Tell Good Stories? The Case of UiPath’s Narrative Mismatch

Some companies perform steadily and enjoy strong customer loyalty, yet never quite resonate with the market. UiPath is one such case worth observing. It has evolved from an RPA tool into an AI automation platform capable of orchestrating complex enterprise workflows, but it still lacks an easy-to-grasp story or a breakout use case that captures attention.

This reflects a common kind of narrative mismatch. When a company is too practical, too hard to visualize, or simply not shiny enough, the market struggles to form belief or commit capital. As George Soros once suggested, markets are not driven by reality, but by belief. This article does not aim to recommend a company, but to explore a deeper question: Why do some good companies struggle to tell a good story?

Introduction: When Market Value Follows Imagination, Not Execution

As industry analysts, we used to believe that solid performance would naturally earn a company market recognition. But financial markets tend to operate more like systems of belief than systems of merit. Only stories that can be understood, imagined, and retold with ease are rewarded with valuation and capital.

In George Soros’ theory of reflexivity, markets are not driven by reality but by belief. Beliefs often attach themselves to simple, concrete, and easily spread narratives.

Put differently, the market prefers narratives that require the least cognitive effort. The strongest stories are the ones you can explain in a sentence or visualize in your mind:

• NVIDIA: “Chips that make AI real”
• Palantir: “AI helps governments fight invisible wars”
• UiPath: “We help automate business workflows” — a line that struggles to create a clear mental picture

What we often see is that when narrative strength falls short of a company’s real value, a quiet mismatch begins to surface.

• Customers are satisfied (strong NRR and stable ARR),
• Investors stay indifferent (low valuation and limited interest),
• New products are misunderstood (still seen through an outdated lens)

This is the kind of delayed recognition that Soros might describe as reflexivity waiting to activate.

Unpacking the Market’s Collective Misjudgment

In recent years, UiPath has been quietly boxed into a narrative: a company that once disappointed investors, entered public markets with an overhyped valuation, and has since failed to reinvent itself for the AI era.

Some of the most common market perceptions sound like this:

• It is just another RPA (Robotic Process Automation) tool.
• Copilot-style AI assistants will make it obsolete.
• SaaS growth is slowing, and even positive free cash flow hasn’t restored investor confidence.

Together, these beliefs form a self-reinforcing loop. It is not just about sentiment cooling. It is about price and belief spiraling downward in sync, creating a textbook case of reflexive deterioration.

Yet, that very consistency in pessimism may be obscuring something more interesting: a growing mismatch between what UiPath is becoming and how it continues to be valued. This is no longer just a workflow automation tool. UiPath is quietly evolving into an AI automation platform for enterprise execution. And the market hasn’t noticed.

From Tool to Platform: A Quiet Transformation

Today, nearly every enterprise claims to be “integrating AI.” But the more important question has become: so what?

Most of today’s AI is framed as a conversational assistant. It is good at summarizing meetings, drafting content, or answering questions. But for most businesses, the real pain point has never been about knowing things. It’s about doing things. Specifically, who handles the repetitive, rule-based, cross-system tasks that still drain productivity?

This is the area where UiPath has quietly built expertise.

1. Beyond conversation, toward execution: Traditional AI copilots act like advisors. They suggest what you should do. UiPath is building something different: an AI assistant powered by automation. t can read an email, log into the ERP system, fill out a form, update inventory, and send a reply without any human intervention. That’s a fundamental shift in what AI can actually execute.
2. The edge lies in knowing how businesses really work: Big AI players like OpenAI, Google, and Microsoft may have the best models. But they don’t have deep access to enterprise workflows, cross-system process logic, or decades of deployment data. UiPath has spent years building that foundation. Its automation network is trained not on internet-scale data, but on the actual operational DNA of thousands of businesses.
3. More than smarter bots, it’s orchestration at scale: UiPath isn’t just improving individual bots. Its long-term vision is to create an orchestration layer that understands entire workflows and assigns tasks to the right bots at the right time. Think of it as an AI-powered conductor, coordinating enterprise execution across systems and teams. This level of process orchestration is still rare in the market and often misunderstood.

UiPath’s current client base includes Wells Fargo, insurance leader Generali, NTT Data, Japanese municipal governments, and the U.S. Department of Veterans Affairs. These organizations are using UiPath to automate critical back-office functions, including compliance, data management, and logistics.

Yet despite this real-world traction, UiPath is still perceived as a legacy automation tool, not as the emerging enterprise AI infrastructure it is steadily becoming.

The Overlooked Reality

UiPath’s narrative has already started to shift, although market sentiment has not yet caught up. Several key friction points help explain why this disconnect remains unresolved.

• Lack of breakout use cases: Products like Maestro and Autopilot offer clear enterprise value, but they remain abstract to the broader market. There is no Copilot-style a widely understood, easily adopted use case.
• Unclear narrative role: Is UiPath a tool, a SaaS company, a platform, or an infrastructure layer? This ambiguity makes it difficult for investors to quickly categorize and believe in its long-term positioning. It lacks the tight narrative framing that companies like Palantir or Snowflake have achieved.
• Limited narrative visibility: Even though UiPath integrates with major AI models including OpenAI, Azure OpenAI, and Google Vertex AI, its brand presence and narrative momentum still lag behind companies like Copilot, Anthropic, or Palantir.

UiPath also faces structural challenges that make narrative ignition harder and valuation recovery more elusive:

• High enterprise adoption barriers: Deploying UiPath requires upfront investment and long implementation cycles, which limits the ability to show viral growth or rapid onboarding.
• Success cases are hard to replicate: Most customer wins involve highly customized workflows tailored to internal processes. These victories are not easily packaged into replicable modules or shareable visual demos, reducing their storytelling power.
• Lack of differentiated narrative labels: While often compared to Palantir and Snowflake, those companies benefit from strong identity hooks. Palantir is framed as an ‘AI battlefield operating system,’ and Snowflake as a ‘data cloud.’ UiPath, however, has yet to offer a framing that resonates immediately with investors that helps the market grasp its value at a glance.

In today’s capital markets, investors prefer companies with high growth potential, strong platform dynamics, or API-first architecture. These attributes imply scalability and ecosystem leverage. UiPath, by contrast, is often positioned as an efficiency-first AI company, emphasizing automation and productivity gains. However, it tends to generate less excitement from a narrative perspective.

So far, no trigger event has emerged to reset this perception. The gap between what the company is becoming and how the market values it remains wide. Faith and price have not yet gone through a process of re-synthesis.

Will the market eventually correct this misalignment? There is no way to know for sure. But if it does, the shift may come faster than expected.

Conclusion: The Narrative Counteroffensive That Has Yet to Begin

UiPath may not be the kind of AI company that grabs attention at first glance. It is better understood as a quiet, stable, deeply integrated enterprise AI platform. Rather than chasing trends, it focuses on the repetitive tasks and system coordination challenges that truly hinder organizational efficiency.

The company may already be laying the groundwork for the next layer of enterprise infrastructure. Its strength lies not in developing the flashiest models, but in understanding operational realities, navigating enterprise workflows, and knowing how work actually gets done. This kind of value is hard to capture in a promotional video. It rarely fits into a single slide of a pitch deck. And that is exactly why it remains misunderstood and undervalued.

UiPath does not present itself as a flashy or futuristic AI player. And that may be exactly why its potential remains under-recognized. From a reflexivity perspective, it represents a case where belief has not yet caught up with execution.

If a turning point in the narrative emerges, such as a breakthrough enterprise use case, a shift in AI infrastructure priorities, or a clearer articulation of its orchestration platform value, UiPath may be reconsidered in a different light. If that moment comes, the market may begin to reassess its position. Whether this shift will happen—or how quickly—is uncertain. But the question remains: what happens when belief starts to match reality? In the end, this is less about one company and more about how the market processes stories. When execution outpaces imagination, value can go unnoticed.

The mismatch between belief and reality we see in UiPath also appears in the Wolfspeed case, where trust unraveled as the company’s story lost clarity. You can read more in our reflection on Wolfspeed’s narrative breakdown.

While UiPath is still searching for a story the market can believe in, Broadcom has quietly rebuilt its identity through a focused AI infrastructure narrative. For more on that shift, see our insight on Broadcom’s transition into a belief-driven platform company.

This article is part of our Future Scenarios and Design series.
It explores how possible futures take shape through trend analysis, strategic foresight, and scenario thinking, including shifts in technology, consumption, infrastructure, and business models.

See more in this category, or explore more notes here.

When AI Redefines the Interface: Jony Ive, OpenAI, and the Future of Display Strategy

OpenAI’s collaboration with Jony Ive is more than just a hardware announcement. It marks a fundamental rethinking of how humans interact with machines. The AI device they are developing, designed without a screen, challenges the long-standing role of displays as the central interface and compels the display industry to rethink its value proposition.

This article explores the structural implications of this shift, including how display modules must be reimagined, how value chains may be restructured, and how display technologies must respond to the new requirements of AI-native devices. If displays are no longer permanent fixtures but instead summoned by context, then the display industry may find itself shifting from competing on shipment volume to excelling at semantic timing and integration. This is both a challenge and an opportunity for reinvention.

Recently, news of OpenAI collaborating with former Apple design chief Jony Ive on a new AI device has drawn significant attention across the tech industry. More than a move from software into hardware, it feels like the start of a broader conversation, one that asks how we will interact with intelligent systems in the years ahead.

What stands out most is that this upcoming device is reportedly designed without a screen, relying instead on voice and environmental sensing. In a world where touchscreens have dominated our digital lives for over a decade, this design decision is more than a technical curiosity. It may signal a deeper shift in how we understand the role of the display itself.

This article does not aim to report the collaboration alone. Rather, it explores a larger question: are we witnessing a structural transformation in the role of displays within AI-native devices?

1.  From Software to Hardware: OpenAI’s New Direction

When OpenAI filed trademarks for consumer electronic products, many saw it as a natural extension of its growing business ambitions. But with Jony Ive joining the effort, it is clear that something deeper is taking shape. This is no longer just about branding or prototypes. It is a reimagining of how humans will interact with intelligent systems.

For OpenAI, this is not simply about launching a new product. It is about rethinking what interaction means when the system already understands, predicts, and responds. As we noted in a previous analysis, OpenAI is not just entering the consumer market. It is trying to rewrite the basic language of human-computer interaction, turning devices from passive tools into intelligent partners.

2.  The Displacement of the Screen

In traditional electronics, the screen has always been the center of interaction. We read, control, and adjust through visual output. But AI-native devices challenge that premise. Their design does not start with what needs to be displayed, but with how the system understands the user.

According to public information, OpenAI and Jony Ive are working on a small, elegant device roughly the size and aesthetic of an iPod Shuffle, worn around the neck. It contains a camera and microphone to perceive the environment but lacks any built-in screen. All visual output would be delivered via connected smartphones or PCs.

This redefines the role of the display. No longer a default conduit, the screen becomes a summoned tool, an optional layer of trust and interpretation. The screen is no longer the interface itself but rather an assistant to the AI’s ability to persuade, explain, or reassure.

In this context, we are no longer talking about a physical display always present on the device. Instead, the display becomes a semantic trigger, something that appears when the situation calls for it and disappears when it is not needed.

This is why the partnership with Jony Ive matters. Ive has always focused less on screen brightness and more on emotional rhythm and the flow between people and products. In an AI-centered world, his approach helps redefine the core question. When a device has no screen, how do we understand it, and how do we trust it?

3.  The Display as a Bridge of Trust

Despite the shift, displays will not disappear overnight. During this transition, screens remain critical trust-building tools for AI devices.

• Users still need visual confirmation of AI decisions and intent
• Summaries, options, and alerts are more quickly grasped visually than audibly
• For new users especially, screens provide psychological safety

In early-stage AI devices, visual modules such as small OLED panels, projection displays, or wearable or ambient formats will likely remain essential. But the design philosophy will shift away from always-on panels toward low-latency, high-readiness screens that appear just in time and vanish without intrusion.

This requires a new way of thinking about what makes a display valuable. It is not just brightness or resolution, but the ability to activate quickly, respond to context, and align with conversational flow.

4.  Structural Changes to the Display Value Chain

While displays will still play a role in AI-enabled products, this shift may be the most fundamental transformation of the display value chain in over a decade.

The question is not whether screens will disappear. It is whether they will remain central sources of value or become modular components that are easily replaced or bypassed.

Here are three structural shifts already underway:

4.1  From Device Integration to Modular Design

Displays used to be tightly coupled with entire devices such as laptops, TVs, or phones. In AI-native logic, screens become optional add-ons. This weakens the fixed relationship between screen and host device.

Display makers will need to explore:

• How to build detachable, summonable, or deployable display modules
• How to integrate with SoCs, sensors, and voice engines to serve as semantic output layers

4.2  From Resolution Wars to Semantic Responsiveness

Display technology has long focused on size, brightness, resolution, and color range. In AI-driven scenarios, the value shifts toward how fast and precisely the screen can deliver relevant visual cues.

This means:

• Screens must support content-driven rendering rather than just pre-loaded visuals
• The emphasis shifts from image quality to timing, coordination, and contextual fit

Some traditional display benchmarks may lose strategic relevance. New priorities such as startup latency, edge-awareness, and energy efficiency will define the next generation of valuable display technologies.

4.3  From Scale-Driven Supply Chains to Design-Centric Collaboration

The display industry has historically relied on economies of scale and standardized panel formats. But AI-native devices may not come from a single vendor or follow uniform design rules.

Instead, display suppliers will need to:

• Participate in upstream scenario planning with device brands
• Offer modular, customizable, on-demand displays
• Build research and development capabilities that synchronize display behavior with voice, sensor, and chip architectures

This means moving from a manufacturing mindset to a design-and-context mindset.

Conclusion

Jony Ive’s collaboration with OpenAI is a provocation rather than just a product reveal. It challenges us to rethink what a device is, what a screen means, and how humans and machines build mutual understanding.

In a world where AI is ambient and ever-present, the display’s job is no longer to shine. Its role is to appear at the right time, in the right way, and help us trust what the system knows.

If the past 20 years of display innovation were measured by shipment volume and pixel count, the next era will be shaped by:

• How well displays support semantic flow
• How fast they respond to AI cues
• How gracefully they appear and disappear

Displays will not vanish. But they will lose their monopoly. What comes next will not necessarily be brighter or bigger. It will be better at knowing when to be seen.

This article is part of our Future Scenarios and Design series.
It explores how possible futures take shape through trend analysis, strategic foresight, and scenario thinking, including shifts in technology, consumption, infrastructure, and business models.

See more in this category, or explore more notes here.

The Age of Semantic Recommendation: Are We Choosing, or Simply Being Understood?

As generative AI and semantic recommendation technologies become increasingly mainstream, the way consumers search, choose, and place trust in products is quietly changing. What counts as visibility, and what we perceive as freedom of choice, are being redefined. The shift moves from keywords to semantic intent, from fixed prices to institutional constraints, from browsing pages to being guided by platform recommendations.

This article begins with subtle adjustments in platform features and language use, examining how semantic systems are steadily infiltrating everyday shopping experiences. It identifies three key trends:

1. Search behavior is evolving from keyword-based actions to narrative-based intent. Brands without a clear semantic presence are becoming harder to identify and recommend.
2. Institutional conditions set by platforms, such as tax systems, shipping policies, and fulfillment access, are increasingly shaping whether a transaction takes place.
3. AI-driven recommendation systems are becoming more centralized, turning visibility into a new form of market control.

As recommendation slots shrink from ten options to three, semantic logic is taking over the role of determining which products get seen. For smaller brands, the competition is no longer just about bidding for ads. It now revolves around data quality, narrative clarity, and the ability to integrate technically with the platforms themselves.

For consumers, what appears to be a more efficient and personalized recommendation experience may quietly come at the cost of exploration and the ability to deviate from default paths. As we grow increasingly accustomed to being understood and more willing to trust the options presented by AI, it becomes worth asking: can we still say with confidence, “This was my choice”?

As the global marketplace enters a period of heightened uncertainty, what is changing is not only economic data, but also the way people search for, choose, and trust the products they buy.

AI-powered search, semantic recommendation systems, and automated classification mechanisms on digital platforms may appear to be technical advancements. In reality, they are redefining what counts as a visible product, who gets recommended, and who is deemed trustworthy.

This shift is not loud. It is quiet, almost invisible. Most consumers may not even notice. Yet the moment we type a question into an AI system, asking what we might want to buy, our intent is already being translated, categorized, and filtered through a semantic logic predetermined by the platform.

This article does not focus on any single company or crisis. Instead, it starts with the language and actions of platforms to uncover a deeper transformation embedded in our everyday shopping behaviors. It is intended for marketers, small brand owners, and digital platform researchers interested in the evolving dynamics of product discovery.

If the way platforms understand us has changed, can we still say the choices we make are truly our own?

1.  From Keywords to Semantic Intent: Rethinking How We Discover Products

Traditional search behavior was built on a familiar pattern: type in a keyword, match results, sort, and click. But with the rise of generative AI tools like ChatGPT, Perplexity, and You.com, consumers are beginning to search in more natural phrases, describing situations, aesthetics, or moods instead of specific terms.

Behind this shift lies the logic of semantic recommendation. Platforms are no longer just matching keywords on the surface. They are attempting to understand the context, motivation, and intent behind what we say, and recommend results that align with that semantic meaning. Semantic recommendation refers to AI systems that interpret user queries based on meaning, context, and intent, rather than literal keyword matching.

Search is starting to feel more like a conversation, less like a command. In North America, we’re already seeing signs of this change. Consumers are no longer typing “brass hair clip,” but asking for “a beautifully designed hair accessory for a friend who loves Japanese aesthetics.”

Language has moved from instructional to expressive, and the way search results are ranked has changed accordingly. What powers this new behavior is no longer just database indexing. It’s the semantic structure of product information, narrative coherence, and the platform’s ability to recognize intent.

Shopify’s recent earnings call hinted at this direction, with features like Sidekick, its AI assistant, and tariffguide.ai, a tool for classifying duties using natural language. These aren’t just technical upgrades. They are early signs of a deeper shift in how consumers think, search, and expect to find things.

When users no longer search for a single keyword but express a style, context, or set of values, brands without a clear narrative may not be found at all. As a result, advertising models are evolving from bidding-based rankings to what might be called contextual relevance.

Contextual relevance refers to whether a product or brand aligns with the overall setting, tone, and intent conveyed in a user’s query. For example, a search for “a minimalist gift for a design lover” might prioritize brands known for clean aesthetics and lifestyle storytelling. Recommendations are no longer based solely on literal keyword matches. They now depend on whether the response “makes sense” in light of how the user speaks and what they seem to be asking for.

For a brand to be recommended in the future, it must first be understood by AI. This will redefine the logic of who qualifies to be seen.

2.  Structural Conditions Are Quietly Reshaping Consumer Choice

In an age of geopolitical instability and shifting trade policies, consumers’ choices are no longer simply about what they want to buy. Increasingly, they are about how the platform allows them to buy it.

Consider elements like tariff calculation, tax-inclusive pricing, regional logistics, and warehouse integration. While these may seem like technical features designed for merchants, they are in fact structural responses by platforms to a changing regulatory landscape.

And these decisions ultimately play out in front of the consumer through every transaction:

• Is the listed price inclusive of tax in different countries?
• Can I pay in local currency and get local delivery?
• Will I encounter surprise fees or shipping delays at checkout?

These conditions influence not only whether a shopper completes a purchase, but whether they choose to return at all.

In the past, the brand was the primary architect of the consumer experience. Today, amid rising geopolitical risks, it is the platform’s architecture that determines whether the consumer feels safe enough to complete a transaction.

When shopping flows are transparent, calculations predictable, and processes reversible, they offer the kind of stability consumers seek in times of uncertainty.

This signals a deeper shift in consumer behavior from demand-driven choices to choices bounded by structural conditions. Whether a person buys is no longer just about preference. It is about what the platform permits, how much friction the process introduces, and whether the cost feels justifiable.

Once platforms move from being neutral transaction tools to becoming gatekeepers of what kinds of transactions can even occur, the boundaries of consumer choice are redrawn, often without anyone noticing.

3.  In the Age of Semantic Recommendation, Visibility Becomes a Privilege of the Few

Traditional keyword advertising, while competitive, still offered a degree of predictability and strategic maneuverability.

Smaller brands could gain visibility and conversions by bidding on niche keywords, refining their copy, and optimizing budget allocation.

But in a semantic search environment increasingly driven by AI, that model is quietly breaking down.

AI platforms no longer show users a page of ranked results. Instead, they directly recommend two or three options that most closely match the user’s semantic intent.

This mechanism, known as intent-based recommendation, demands far more than traditional search.

It requires consistent language, high-quality content, and alignment with the user’s context, tone, and implicit intent.

The recommendation slots are fewer and the bar for inclusion is much higher.

To be considered by an AI’s shortlist, a brand must offer highly structured data, a coherent narrative, and technical integration with platform systems.

For many small and emerging brands, these are thresholds that money alone can’t overcome.

Looking ahead, AI-driven advertising may no longer be priced by impressions or clicks, but by a new form of semantic eligibility:

• Are you well-known enough to be included?
• Have you provided language models with semantic-rich material that can be understood and classified?
• Do you appear in the open datasets of the platform’s partners?

The path to visibility is shifting from a bidding war to a semantic capital race. Semantic capital refers to the structured, machine-readable language, metadata, and narratives that make a brand easily discoverable and classifiable by AI.

Those who are recommended will increasingly be those who can offer the most consistent, rich, and machine-readable narrative, such as brands with data depth, platform fluency, and narrative clarity.

And as the recommendation model consolidates around fewer slots, visibility will cluster even more tightly around those already well positioned.

The result is a quiet exclusion. Small brands are not erased. They just aren’t surfaced.

Consumers rarely notice this shift. But when a search experience moves from “choose one of ten” to “trust one of three,” being recommended becomes synonymous with being seen.

Conclusion: The Illusion of Choice and the Structural Bias of Semantic Machines

We think we are making choices, but in truth, it is the semantic models choosing for us.

For consumers and small brands alike, future competition will not revolve around visibility, but around the depth and clarity of meaning. The platform economy is undergoing a structural rewrite. The logic of how demand is formed and distributed is no longer what it once was.

What looks like freedom of choice is increasingly guided, and constrained, by the semantic frameworks set by platforms. This pattern of surface-level openness and underlying concentration is not limited to e-commerce search. It also shapes AI-driven news feeds, video platforms, and social media algorithms. It is a quiet but far-reaching transformation.

As semantic search and AI recommendation systems continue to shape the architecture of choice, consumers appear to have more options, yet increasingly rely on the language, filters, and classifications defined by platforms.

Whether a product is discovered no longer depends solely on a brand’s effort. It depends on whether the platform understands the product and is willing to recommend it.

This shift imposes higher visibility thresholds for small brands. Narrative consistency, structured data, system compatibility, and platform alignment all become prerequisites for being seen.

As recommendation slots shrink to just a few, access to those spots becomes scarce and gated. The competition moves from an ad bidding war to a contest over semantic capital.

Perhaps we should ask this instead:

In the future, what determines whether a product is worth recommending?

Is it the brand that bids the highest, or the one that can be best understood and trusted?

More fundamentally, only those who hold semantic capital, those who can be clearly read, categorized, and interpreted, will be seen. In a semantic recommendation system, being understood is the first qualification for being recommended, and even for being allowed to bid for attention.

Our Reflections and Questions

Semantic recommendation and AI search have certainly improved efficiency and made the consumer experience feel more intuitive. But this shift toward being understood by algorithms and guided by platform recommendations is quietly eroding our capacity for exploration and comparison. This article outlines the trade-offs between discoverability and autonomy in a recommendation-driven world.

When we grow used to AI telling us, “These are the top three brands for you,” do we still feel curious about the fourth or fifth option?

Are we still willing to go beyond what the system defines as our intent, to discover voices, values, or products that fall outside the model’s view?

Consumers may not have designed this system, but they still hold the freedom to go with the current or to veer slightly off course.

In a world where semantic recommendations are shaping how we shop, choosing is no longer just a transaction. It becomes a quiet act of resistance, a moment of reflection:

How are we being recommended?

Are we comfortable letting our intent be predicted by a platform?

Do we still have the instinct to search for the unexpected?

If small brands hope to challenge the growing concentration of semantic capital through stories and meaning, then consumers may be their strongest allies in this new landscape, as long as we are willing to relearn how to choose.

This article is part of our Future Scenarios and Design series.
It explores how possible futures take shape through trend analysis, strategic foresight, and scenario thinking, including shifts in technology, consumption, infrastructure, and business models.

See more in this category, or explore more notes here.

Imagining Asia’s Supply Chain Collaboration: Toward Flexible, Real-Time Networks

Amid rising costs, shifting geopolitics, and increasing supply chain fragmentation, Asia’s supply chain networks are undergoing a quiet yet profound transformation.

This article explores how the next phase of supply chain collaboration will move beyond relocating production hubs to creating real-time, flexible, and information-driven systems.

While Taiwan is emerging as a critical logistics node, the real race lies in who can build the next-generation information infrastructure. The ability to control standardized, real-time data flows may ultimately determine the future leaders of regional and global trade.

In a previous analysis, we explored how tariffs, rising costs, and geopolitical shifts have been quietly reshaping Asia’s supply chain collaboration patterns, with Taiwan emerging as a key node in this evolving network.

Beyond the immediate relocation of manufacturing hubs, a deeper shift is now underway. Companies are seeking faster responses, closer regional collaboration, and more resilient ecosystems that can withstand an increasingly fragmented global environment.

This article imagines the next stage of Asia’s supply chain collaboration: a transformation from static networks into dynamic, living systems.

1.  The Invisible Needs Emerging in Asia’s Supply Chains

As companies recalibrate their operations across Asia, the first wave of changes has focused primarily on relocating factories and logistics hubs. Yet a more profound transition is quietly unfolding beneath the surface.

Today, supply chains are no longer judged solely by cost efficiency or geographic proximity. Enterprises now demand real-time adaptability, localized resilience, and the ability to pivot quickly across fragmented regional markets.

Traditional models built for large, centralized systems struggle under the pressure of faster market cycles and rising regulatory complexity.

2.  Imagining a Real-Time Supply Chain Collaboration Platform

If the old logic of supply chains was about maximizing economies of scale, the new logic favors flexibility, speed, and proximity to markets.

Imagine a real-time collaboration platform spanning Taiwan, Southeast Asia, and Northeast Asia. Companies could instantly access information on available manufacturing capacities, warehousing options, light-processing facilities, and cross-border logistics solutions.

Instead of committing to rigid, long-term routes, businesses could dynamically reconfigure their production and distribution strategies based on current demand, regulatory shifts, or logistical conditions.

A system like this would allow enterprises to design modular, adaptive supply chains that that flex and adapt to local realities almost in real time.

3.  Beyond Logistics: Rethinking Regional Resilience

To meet future demands, rearchitecting the supply chain across Asia is no longer optional. The future of supply chains will not be defined solely by the movement of goods, but by the synchronized optimization of information flows and physical logistics.

Today, information across Asia’s supply chain networks remains highly fragmented. Many small agents, freight forwarders, customs brokers, and transport companies in the region still rely heavily on manual operations, communicating via phone, fax, and email. There is a lack of standardized data flow, leading to high trust costs and a significant lag compared to other industries that have embraced real-time digitalization, such as financial services.

Compounding this challenge is the region’s geopolitical complexity, which makes it difficult for network effects to take root quickly. The path toward supply chain technologization in Asia will not be smooth and will face considerable hurdles.

While Taiwan stands as an important physical logistics node, the question of who will emerge as the true platform builder remains open. The ability to create a new, real-time supply chain collaboration platform will define the next phase of competition.

The transformation underway in Asia’s supply chains is not simply about shifting cargo routes. It is about rewriting the foundational layer of collaboration itself. Whoever succeeds in building a new information coordination infrastructure, not just for logistics channels but also for data synchronization, resource allocation, and standardized interfaces, will have the opportunity to lead the next era of supply chain development.

The eventual platform builders may emerge from American tech giants, or they may come from new players across Asia. The race is quietly taking shape.

Whoever controls the infrastructure layer for global trade data flows could secure a winner-takes-all advantage, much like Stripe for financial infrastructure.

In the future, leadership in supply chains will not belong to the companies that move the most containers, but to the platforms that manage the largest volumes of supply chain data and transaction flows.

These platforms will connect companies, systems, and nations through standardized, real-time, and interoperable networks, becoming the API layer, the living interface of the future supply chain. Those who build this information network will set the new rhythm for Asia’s supply chains.

Closing Thought

In a world where supply chains are no longer static, perhaps the greatest strength lies not just in moving faster, but also in moving together in ways that are smarter, closer, and more resilient than ever before.

This article is part of our Future Scenarios and Design series.
It explores how possible futures take shape through trend analysis, strategic foresight, and scenario thinking, including shifts in technology, consumption, infrastructure, and business models.

See more in this category, or explore more notes here.

Optical Actuators: The Overlooked Risk Node in a Geopolitically Fragile Supply Chain

As global tech competition intensifies and rare earth elements become increasingly strategic, a quiet yet critical vulnerability is emerging within the imaging module supply chain: optical actuators. Although these components account for only 15–20% of camera module costs, they are essential to core functions such as autofocus and image stabilization, with applications ranging from smartphones and AR headsets to autonomous vehicles and medical systems. Their deep dependence on Chinese-sourced rare earth magnets, particularly neodymium (Nd), praseodymium (Pr), and dysprosium (Dy), makes them highly susceptible to geopolitical and material risks.

This report analyzes the structure of these actuator-related dependencies, explores the potential channels of disruption, and outlines actionable strategies for companies and policymakers navigating the next wave of global supply chain realignment.

1.  Background: A Critical Component Hiding in Plain Sight

While chips and EV motors have dominated the rare earth risk narrative, optical actuators are quietly emerging as the next chokepoint. These miniature motion control devices are essential to modern camera modules and are deployed across smartphones, AR/VR systems, automotive cameras, robotics, and medical imaging equipment. With rising U.S.–China tensions, the actuator’s dependence on China-dominated rare earth materials places a structurally underestimated burden on midstream and downstream players.

2.  Application Scope and Rare Earth Dependency

Optical actuators use high-performance magnets, typically neodymium-iron-boron (NdFeB), to achieve precise lens movement. These magnets require a stable supply of rare earth elements, almost exclusively processed in China. Table 1 outlines actuator use across industries and the varying levels of rare earth dependency. Applications with high design precision, such as AI vision systems, flagship smartphones, and surgical endoscopes, face particularly high exposure.

Table 1   Key Optical Actuator Applications and Rare Earth Dependency

Source: Researcher and Research LLC

3.  Impact Assessment: Strategic Risks from Rare Earth Export Constraints

3.1  Cost Pressure and Supply Fragility

Export controls on Nd, Pr, Dy, and Tb would cause significant price volatility in magnetic materials. Actuator prices could rise 15–40%, impacting bill of materials (BOM) costs in mid-to-high-end modules. Most voice coil motor (VCM) suppliers lack long-term hedging mechanisms, amplifying the impact of raw material shocks and weakening downstream pricing power.

3.2  Product Development Delays

Non-Chinese rare earth production remains limited. A sudden supply shock could prevent ODMs from meeting delivery schedules, forcing OEMs like Apple and Samsung to redesign products or adopt less proven actuator alternatives such as micro-electro-mechanical systems (MEMS). This shift could lengthen development cycles and increase production risk.

3.3  Competitive Realignment and Dual Bifurcation

China may strengthen its internal actuator supply chain, leveraging material access as a competitive edge. This risks creating a dual bifurcation scenario, where both design standards and materials sourcing diverge regionally. Over time, this separation may lead to regionally incompatible supply chains between the U.S. and China.

4.  Strategic Response Options

4.1  Short-Term Mitigation

Temporarily scale back high-end module r.ollouts and substitute fixed-focus modules paired with algorithmic enhancement.

Build inventory buffers and adopt lower rare-earth-content actuator designs.

4.2  Mid-to-Long-Term Strategies

Accelerate MEMS and ceramic actuator R&D.

Incorporate more software-driven image control to reduce mechanical dependency.

4.3  National-Level Interventions

Diversify sourcing through mining investments (Japan, Australia, U.S.).

Launch rare earth stockpiles and subsidies for alternative technologies.

Conclusion

Though often underestimated, optical actuators are central to imaging precision and functional stability. In an era of geopolitical fragmentation and strategic resource competition, their vulnerability to rare earth volatility is no longer a niche concern. It represents a frontline risk. Companies that invest in alternative technologies, predictive monitoring systems, and diversified sourcing will be best positioned to thrive amid global supply chain rebalancing.

Glossary of Key Terms:

VCM (Voice Coil Motor): An electromagnetic actuator used for autofocus and stabilization.

MEMS (Micro-Electro-Mechanical Systems): Miniaturized devices that combine electrical and mechanical functions.

NdFeB: Neodymium-Iron-Boron, a type of powerful rare earth magnet.

BOM (Bill of Materials): Comprehensive list of parts and costs in a product.

Dual Bifurcation: Simultaneous divergence in technical standards and material sourcing paths.

This article is part of our Future Scenarios and Design series.
It explores how possible futures take shape through trend analysis, strategic foresight, and scenario thinking, including shifts in technology, consumption, infrastructure, and business models.

See more in this category, or explore more notes here.

NVIDIA’s Leadership in AI: Key Insights from Jensen Huang’s GTC Keynote

We’ve explored the evolution of AI, NVIDIA’s strategic positioning, and its impact at each stage. The breakthrough of the GeForce 5090 will drive the shift from Perceptual AI to Generative AI.

Next, Agentic AI will evolve into Physical AI, and these two will eventually merge, creating a profound real-world impact.

While NVIDIA has established itself as the dominant player in the AI ecosystem, the varying hardware needs across industries and use cases will spur competitors to find more cost-effective alternatives. As technology advances, this competition will only intensify.

Our Perspective

1.  What Did We Learn from Jensen Huang’s Keynote?

Last year, NVIDIA made its return to in-person events, creating an atmosphere akin to a rock concert. Some even called it the “Woodstock of AI,” while this year’s event has been dubbed the “Super Bowl of AI.” As the GPU Technology Conference (GTC) continues to grow in scale and influence, AI has firmly positioned itself at the forefront of global technological advancement.

In Jensen Huang’s keynote, we mapped the evolution of AI, NVIDIA’s strategic positioning, and its industry impact at each stage. As illustrated in Table 1, AI development unfolds in several key phases, with NVIDIA playing a crucial role at each stage by providing the essential computing power through its GPU technology.

Table 1   NVIDIA’s AI Strategy

Source: Researcher and Research

NVIDIA is evolving into the central supplier of the AI ecosystem, moving beyond its role as a GPU manufacturer. By actively integrating technologies such as CUDA, Omniverse, AI agents, and robotics, it is creating an irreplaceable software-hardware moat.

This shift suggests that Agentic AI could disrupt SaaS, enterprise software, marketing models, and decision-making processes, with significant implications for e-commerce strategies. More importantly, the next phase of AI development will extend beyond software, influencing hardware devices, automation applications, and physical-world AI technologies. For example, Physical AI is poised to transform industries like manufacturing, logistics, autonomous vehicles, and other automation sectors.

Next, we will delve deeper into these topics.

2.  How NVIDIA Bridges Perceptual AI to Generative AI

In his keynote, Jensen Huang highlighted the technological advancements of the GeForce 5090, including a 30% reduction in size, a 30% improvement in cooling efficiency, and performance that far exceeds the 4090. We view this as a critical “computational infrastructure” for NVIDIA, bridging the development path from Perceptual AI to Generative AI.

Through advanced chip fabrication, packaging technologies, and optimized architectures, the 5090 significantly boosts GPU performance, accelerating AI computations across applications such as gaming rendering, 3D design, and AI-generated content.

NVIDIA has deeply integrated AI into the core of its GPUs, incorporating AI-assisted rendering techniques like 100% path tracing and AI-based pixel completion to demonstrate AI’s evolving role in graphics processing. This has the potential to disrupt workflows in the gaming and animation industries, while also transforming computational methods used across software and content creation sectors. This breakthrough is poised to make a lasting impact on game development, animation production, professional computing, and AI training.

3.  Will the Next Wave Be Agentic AI or Physical AI?

The next phase in the evolution of Generative AI will likely move toward either Agentic AI or Physical AI. While there is no strict order, their relationship can be understood from a technological development perspective.

Agentic AI refers to AI systems with autonomous decision-making capabilities. These systems can perform tasks based on goals, environmental changes, and context, rather than simply responding to commands. It primarily develops in the “digital world,” providing AI with autonomy but without direct influence over the “physical world.”

In contrast, Physical AI involves AI systems capable of performing actions in the real world, interacting physically with their environment. Common applications include robots, autonomous vehicles, and smart factories. Physical AI depends on Agentic AI for decision-making and incorporates technologies like sensors and motion control to facilitate real-world interactions. Therefore, Physical AI can be seen as the natural evolution of Agentic AI.

Ultimately, these two areas will converge. The future of Physical AI will likely be driven by advanced Agentic AI, enabling robots and autonomous vehicles to make independent decisions and exert real-world influence. This is why NVIDIA is initially focusing on developing technologies related to Agentic AI before advancing into Physical AI.

3.1  Agentic AI Will Redefine Software Companies

Agentic AI represents the next evolution in AI, shifting from a passive “responder” to an active, autonomous decision-maker. In the future, AI will not only react to commands but will also perceive its environment, understand context, and engage in reasoning, planning, and action. It will even use tools to execute tasks. With the ability to browse the web, read texts, watch videos, and learn from these interactions, Agentic AI will evolve beyond relying on fixed datasets and will possess “self-learning” capabilities.

This transition will have a profound impact on industries like search engines, digital marketing, SEO, and e-commerce recommendation systems, as AI moves from being a supportive tool to becoming the ultimate decision-maker. For instance, AI could autonomously optimize marketing campaigns or adjust e-commerce strategies based on real-time data and insights.

Furthermore, this shift will challenge traditional SaaS systems, such as CRM and ERP software, which businesses currently rely on. AI will no longer be just an enhancement to these systems; businesses will require AI agents with decision-making abilities to directly execute tasks. Traditional software, which relies on manual inputs and predefined workflows, will no longer meet the demands, pushing companies toward more adaptive and intelligent solutions driven by Agentic AI.

3.2  Physical AI and the Robotics Revolution: A Turning Point for Manufacturing and Logistics

NVIDIA envisions Physical AI not only enabling AI to understand data but also allowing it to influence the physical world. This technological leap will drive advancements in robotics, transitioning from rigid, fixed production lines to more flexible, intelligent robots. Physical AI’s ability to understand concepts like friction, inertia, causality, and object permanence (i.e., objects don’t disappear, they are just temporarily hidden) demonstrates how AI is beginning to comprehend the physical world in a way that empowers robots to learn and adapt autonomously.

This evolution will significantly impact industries such as smart logistics, autonomous vehicles, warehouse management, and even urban planning. As AI becomes more adept at interacting with the real world, it will enhance systems reliant on precise physical interactions. Autonomous vehicles and robots within logistics and warehouse operations will be able to navigate complex environments, performing tasks with greater efficiency, safety, and flexibility.

Consequently, logistics and supply chain companies must closely track the development of robotics and AI computing solutions. These innovations are set to reshape operations and business models in industries that depend on physical tasks and movements.

4.  Conclusion and Discussion

4.1  NVIDIA: From GPU Manufacturer to AI Ecosystem Leader

Jensen Huang highlighted how GeForce played a key role in promoting CUDA globally, catalyzing the rise of AI, and now, AI itself is revolutionizing the world of computer graphics. This signifies that NVIDIA has evolved beyond its roots as a graphics computing company, positioning AI computing as its central competitive advantage.

In the past, real-time graphics rendering relied heavily on path tracing technology, where each pixel was rendered mathematically, with AI inferring the other pixels. Today, AI directly participates in graphics creation, suggesting that future GPUs will likely integrate AI computing even more deeply. This shift is not only transforming gaming but is poised to extend into fields like medical imaging, scientific simulations, and 3D design.

NVIDIA has fully integrated AI with its GPUs, creating a powerful synergy between hardware and software. Moving from CUDA to AI computing, and now to AI-enhanced graphics, NVIDIA has transitioned from being a graphics card company to a leader in AI computing. AI is no longer just influencing computation methods; it now plays an active role in decision-making and data processing, understanding context, generating responses, and retrieving information to enhance understanding.

NVIDIA’s software strategy—spanning CUDA, Omniverse, and generative AI models—has become its core strength, shifting away from a hardware-centric business model. Instead, NVIDIA has locked industries into its software ecosystem, creating a robust competitive advantage. Specifically:

• CUDA forces developers to run AI workloads on NVIDIA chips, reinforcing its market dominance.
• AI Agents enable businesses to deploy AI decision-making systems directly on NVIDIA platforms, enhancing operational efficiency.
• Omniverse leads the charge in the industrial metaverse, making NVIDIA the exclusive leader in this space.
• Physical AI opens new opportunities in robotics, positioning NVIDIA as a key player in the automation future.

Through this integrated hardware-software strategy, NVIDIA has evolved from simply “selling GPUs” to controlling the entire AI industry’s computational and development landscape, making it increasingly difficult for competitors to challenge its dominance.

4.2  How Will the AI Industry’s Competitive Landscape Evolve?

The AI industry has evolved into a core technology domain, transitioning from a specific application to a disruptive force across multiple sectors. Each wave of AI innovation presents three primary challenges: data, training, and scalability. These factors will play a pivotal role in shaping the future competitive landscape, particularly in areas like data acquisition, computing performance, and large-scale deployment.

NVIDIA has established a near-monopoly in the AI training market, competing with industry leaders such as Google, Meta, and OpenAI in both AI training and inference. While NVIDIA maintains its dominance in AI training, the focus of future competition will shift towards inference—reducing costs, enhancing efficiency, and developing specialized chips. As ASIC (Application-Specific Integrated Circuit) technology progresses, competitors are intensifying efforts to create cost-effective solutions and secure breakthroughs in the inference market.

4.2.1  High-Performance Inference Chips (ASIC vs. GPU)

NVIDIA’s GPUs remain dominant in the high-performance computing sector. However, with the rise of specialized AI chips, particularly in cloud AI inference and edge computing, the cost advantages of ASIC technology are becoming more apparent. Notable examples include Google’s TPU, AWS’s Inferentia, Meta’s custom AI chips, and Tesla’s Dojo, all of which are increasingly challenging NVIDIA’s market leadership. As these competitors’ solutions mature, GPUs may face increasing challenges from ASICs, especially in inference applications, and potentially even in training.

4.2.2  Software-Hardware Integration and Ecosystem Development

Beyond hardware competition, the AI software ecosystem has become a crucial battleground. NVIDIA leverages its ecosystem, including CUDA and TensorRT, to strengthen its market position. However, Google, AWS, and Meta are actively developing their own AI software frameworks, such as TensorFlow and PyTorch, to reduce reliance on NVIDIA’s technology. In the future, as AI software and hardware become more tightly integrated, major companies will focus on building their own ecosystems while attempting to weaken NVIDIA’s influence in the software domain. The competition will center not only on hardware performance but also on the ability to establish seamlessly integrated software ecosystems.

4.2.3  Cloud vs. Edge AI Computing

Cloud computing remains the primary platform for AI training, but the rise of Edge AI is driving a significant shift. As applications like smart vehicles, automated factories, and IoT devices expand, the demand for edge computing continues to grow. Products such as NVIDIA Jetson and Tesla’s FSD chips are increasingly establishing strong footholds in the Edge AI market. As Edge AI progresses, these chips—designed for autonomous driving, smart devices, and industrial automation—will compete with large-scale cloud AI training platforms, collectively accelerating the adoption and evolution of AI technology.

4.3  Why Do So Many Companies Challenge NVIDIA Despite Its Unshakable Dominance?

NVIDIA’s unshakable dominance in AI stems from its highly integrated ecosystem, making it difficult for competitors to surpass in the short term. However, many companies continue to challenge its position, reflecting the complexity of the industry’s competitive landscape.

Technology evolves rapidly, and different applications have distinct hardware requirements. In some areas, specialized ASICs offer more cost-effective solutions than GPUs. Consequently, companies like Google, Meta, and AWS are developing their own custom chips to reduce reliance on NVIDIA’s products.

In summary, while NVIDIA’s leadership in AI remains formidable, competitors persist in challenging its position due to the diverse hardware demands across industries and applications, as well as the drive for more cost-efficient alternatives. This means NVIDIA will continue to face competition, particularly in the AI inference market.

This article is part of our Future Scenarios and Design series.
It explores how possible futures take shape through trend analysis, strategic foresight, and scenario thinking, including shifts in technology, consumption, infrastructure, and business models.

See more in this category, or explore more notes here.

AI chip market evolution: Cloud vs. edge in training and inference

Part 2: Edge training, inference, and market trends

The AI chip market is undergoing significant transformations, which can be understood through two key dimensions: deployment environments (cloud vs. edge) and market segments (training vs. inference).

Cloud-based training currently dominates the market and is expected to maintain strong growth in the future. Training is critical for AI model development, requiring immense computational power to process vast amounts of data, which is why it is primarily concentrated in cloud data centers. NVIDIA is the dominant player in this space, but competition from AWS, Google TPU, and Meta could reshape the market landscape in the coming years.

Alongside training, the cloud inference market is also experiencing rapid growth. Inference refers to applying trained AI models to real-world scenarios and making decisions based on real-time data. The increasing demand for inference is driven by advancements in AI models and emerging application scenarios. While inference requires lower computational power than training, it demands high accuracy and low latency, making cloud-based inference expansion essential for scaling AI applications.

On the edge, both edge training and edge inference cater to demands for low latency and data privacy. Compared to cloud-based training, edge training remains a smaller market but is expected to grow with the rise of smart devices and autonomous vehicles. Meanwhile, edge inference focuses on real-time, on-device processing, which is crucial for applications such as autonomous driving and other latency-sensitive use cases. While the edge market remains a niche segment, its importance will increase as these specialized applications continue to evolve.

In conclusion, training and inference in both cloud and edge environments present unique challenges and opportunities. Companies capable of delivering high-performance, cost-effective solutions in these areas will have the potential to challenge existing market leaders.

Following our previous discussion on the cloud AI training and inference market, this article will focus on the on-premises AI chip market for training and inference.

Our Analysis

Compared to the cloud market, on-premises AI solutions offer distinct advantages in low latency and data privacy. As emerging applications such as autonomous vehicles and smart devices grow, on-premises AI training and inference are expected to be key drivers of future market expansion. This article will analyze the trends shaping this segment and explore efficient, cost-effective solutions to address competitive challenges. Finally, we will summarize the key findings from both articles in a comparative table and discuss the future trajectory of the AI chip market, highlighting strategic areas that potential competitors and market players should closely monitor.

3. Edge Training

In the AI chip market, edge training represents the smallest segment, accounting for approximately 3–5% of the market. Edge training requires high computational efficiency and is highly sensitive to latency. These applications must perform AI training at the edge rather than relying on cloud-based computing resources. Examples include autonomous vehicles that require real-time learning and adaptation of their perception systems, industrial machines and robots that train and adapt based on local data, wearable devices such as smartwatches and health monitoring systems that continuously learn and adjust based on user data in real time, and smart city applications that require immediate data processing.

With growing concerns over data privacy and security, businesses are increasingly opting for edge training to ensure sensitive data remains protected. In this market, NVIDIA remains the dominant supplier, while AMD holds a smaller share. If emerging players like DeepSeek can provide cost-effective AI training solutions optimized for edge devices—especially amid growing demands for data privacy and real-time processing—they could become serious challengers to NVIDIA.

4. Edge Inference

Edge inference accounts for approximately 10–15% of the AI chip market, primarily serving enterprise inference needs by running trained models on endpoint devices or edge computing systems. These applications demand low latency and real-time responsiveness, with significantly lower computational requirements than AI training. Key use cases include:

1. Smart security: real-time image analysis for detecting suspicious behavior or individuals
2. Smart home: rapid response to user commands and real-time environmental adjustments
3. Smart transportation: traffic monitoring, autonomous driving, and intersection surveillance
4. Drones – real-time image analysis for navigation and filming
5. Healthcare monitoring – real-time data processing to assess user health conditions
6. Industrial IoT – data collection and analysis to ensure smooth production operations

As concerns over data privacy and security grow, more businesses and institutions are shifting inference processing to local devices to prevent sensitive data leakage. This trend is especially prominent in industries requiring strict privacy protection, such as finance, healthcare, and government.

NVIDIA remains the leading supplier, while AMD continues to gain traction with advancements in its products. Competitors offering high-efficiency, low-power edge inference solutions—tailored to smart home, security, and industrial IoT applications—could challenge NVIDIA and AMD in this evolving market.

Conclusion

The AI chip market is undergoing rapid transformation, with different market segments showing varying growth trends and challenges, as shown in Table 1.

Table 1   AI chip market structure analysis and comparison

Source: Researcher and Research

The growth of the edge market and its divergence from the cloud market compel us to consider the potential impact of emerging competitors. Among these, DeepSeek’s technological innovations are particularly noteworthy, as they have the potential to disrupt the current market landscape. Although large companies like NVIDIA and Google currently dominate the market, DeepSeek’s rise—whether in hardware acceleration or breakthroughs in AI training—could significantly alter this dynamic.

Despite the ongoing emergence of Cloud Service Providers (CSPs) developing their own ASICs, ICs, and other competitors, we believe that NVIDIA continues to maintain a strong market position and technological advantage. The primary reasons for this are:

1. GPU Advantage

NVIDIA has long dominated the GPU market, with accelerators like the A100 and H100 becoming industry standards for AI training and inference. NVIDIA’s GPUs not only support training but also handle large-scale inference, playing a critical role in many AI applications. As a result, even with the competition from CSPs developing their own ASICs and ICs, NVIDIA maintains a strong advantage in applications requiring general-purpose and high-performance computing.

2. Robust Software Ecosystem

NVIDIA boasts a comprehensive developer ecosystem, including tools like CUDA, cuDNN, and TensorRT, making it easy for developers to build AI applications on NVIDIA hardware. In contrast, competitors like CSPs with in-house ASICs and DeepSeek need to invest significant time and resources in developing their own software ecosystems, giving NVIDIA a clear edge.

3. Efficient AI Computing Platform

NVIDIA’s high-performance computing (HPC) and AI platforms offer highly optimized hardware and software integration, providing powerful acceleration for a variety of AI workloads, such as natural language processing and image recognition. The optimization of these platforms gives NVIDIA a performance advantage in processing large datasets and models, surpassing other competitors.

However, NVIDIA also faces significant challenges, primarily from two forces in the specialized competition space:

1. Development of CSP-Developed Chips

In an effort to reduce reliance on third-party chip suppliers and achieve cost control and hardware customization, many CSPs have opted to develop their own chips. For example, Google’s TPU focuses on neural network inference, while Amazon’s Inferentia is optimized for inference scenarios. These in-house chips offer more efficient, cost-competitive solutions for specific applications.

2. Breakthroughs in Specialized ASICs

Some competitors may pose a threat to NVIDIA by achieving breakthroughs in specific areas, such as low-latency inference or other specialized acceleration needs, and developing highly specialized ASICs. Especially in cost-sensitive markets or those with niche requirements, NVIDIA’s high-end GPUs (such as the A100 and H100) may not be as attractive as these specialized ASIC solutions due to their higher price points.

Therefore, NVIDIA currently maintains a significant competitive advantage in the general AI training and inference market. However, if CSPs aggressively develop in-house ASICs or competitors make breakthroughs in specialized areas, NVIDIA will face increased competitive pressure. Future competition will depend on whether these challengers can surpass NVIDIA products in terms of performance, efficiency, price, and ecosystem support.

In conclusion, when discussing the AI chip market, the training and inference needs in both cloud and edge markets each present different challenges and opportunities. The development of these four areas is interwoven, and the characteristics of each can influence the future direction of the market. As AI technology evolves, businesses that can provide higher-performance, cost-effective solutions in these areas will not only effectively address current market challenges but also capture growth opportunities in the future. Such solutions have the potential to challenge the current market leaders and carve out a strong position in these rapidly developing sectors.

This article is part of our Future Scenarios and Design series.
It explores how possible futures take shape through trend analysis, strategic foresight, and scenario thinking, including shifts in technology, consumption, infrastructure, and business models.

See more in this category, or explore more notes here.

AI chip market evolution: Cloud vs. edge in training and inference

Part 1: Cloud training and inference

The AI chip market is undergoing significant transformations, which can be understood through two key dimensions: deployment environments (cloud vs. edge) and market segments (training vs. inference).

Cloud-based training currently dominates the market and is expected to maintain strong growth in the future. Training is critical for AI model development, requiring immense computational power to process vast amounts of data, which is why it is primarily concentrated in cloud data centers. NVIDIA is the dominant player in this space, but competition from AWS, Google TPU, and Meta could reshape the market landscape in the coming years.

Alongside training, the cloud inference market is also experiencing rapid growth. Inference refers to applying trained AI models to real-world scenarios and making decisions based on real-time data. The increasing demand for inference is driven by advancements in AI models and emerging application scenarios. While inference requires lower computational power than training, it demands high accuracy and low latency, making cloud-based inference expansion essential for scaling AI applications.

On the edge, both edge training and edge inference cater to demands for low latency and data privacy. Compared to cloud-based training, edge training remains a smaller market but is expected to grow with the rise of smart devices and autonomous vehicles. Meanwhile, edge inference focuses on real-time, on-device processing, which is crucial for applications such as autonomous driving and other latency-sensitive use cases. While the edge market remains a niche segment, its importance will increase as these specialized applications continue to evolve.

In conclusion, training and inference in both cloud and edge environments present unique challenges and opportunities. Companies capable of delivering high-performance, cost-effective solutions in these areas will have the potential to challenge existing market leaders.

This article, “AI Chip Market Evolution: Cloud vs. Edge in Training and Inference,” is structured into two parts to provide a comprehensive analysis of the market’s diverse landscape and future trends.

Part 1: Cloud Market
This section explores Cloud Training and Cloud Inference, examining their growth potential, key industry players, and evolving competitive dynamics.

Part 2: Edge Market
This section delves into Edge Training and Edge Inference, focusing on low latency, data privacy, and emerging applications, while assessing their technological advancements and market opportunities.

By adopting this structured approach, the article offers a clearer understanding of the AI chip market’s distinct segments, delivering valuable insights to industry stakeholders for strategic planning and competition.

Our Analysis

When examining the future of the AI industry, understanding the market structure is essential. The AI landscape can be broadly categorized into four key markets: Cloud Training, Cloud Inference, Edge Training, and Edge Inference. While each of these markets operates with distinct mechanisms and applications, they are deeply interconnected, forming the foundation of AI technology.

• Cloud Training: The Core of AI Development
  Cloud training serves as the backbone of AI advancement, responsible for large-scale data processing and model training. As data volumes continue to grow and computational demands rise, cloud training remains the cornerstone of AI’s rapid technological progress.
• Cloud Inference: Enabling Real-Time AI Applications
  Once AI models are trained, cloud inference plays a crucial role in applying them to real-world scenarios. Unlike training, inference focuses on enabling AI to deliver rapid, real-time responses. The evolution of cloud inference is tightly linked to cloud training, and as demand for AI-powered solutions expands, the cloud inference market continues to grow.
• Edge Training & Edge Inference: Addressing Privacy and Latency Needs
  The edge AI market—comprising edge training and edge inference—is gaining momentum, particularly in applications requiring high data privacy and low latency. By shifting training and inference processes from the cloud to local devices, edge AI enhances security and reduces response times. In certain use cases, edge solutions offer distinct advantages over cloud-based alternatives.

These four market segments not only highlight how AI technologies are applied across industries today but also signal the future direction of AI development. Despite their unique operational models, their interdependencies will continue to drive AI innovation and industry-wide progress.

This article will provide an in-depth exploration of the cloud sector within the AI chip market, focusing specifically on cloud training and cloud inference markets. As artificial intelligence technologies advance, these two markets are rapidly becoming key growth areas. Given the critical role of training and inference in the AI model lifecycle, both will significantly influence the structure of the market and future competitive dynamics. In this article, we will analyze the current state and future growth potential of the cloud training market, along with the competitive landscape of key players. Additionally, we will explore the development trends of cloud inference and its impact on emerging application scenarios. This section will lay the foundation for the subsequent discussion on the edge market and provide a comprehensive understanding of the current landscape and challenges within the AI chip market.

Among these four markets, the cloud training market forms the foundation of AI development, as large-scale data processing and model training require substantial computational power, which cloud platforms provide. With the advancement of AI technologies and the explosion of data, the cloud training market has become the core driver of AI technology growth. In the cloud training market, the deep learning process of AI models demands vast computational resources, typically relying on the infrastructure provided by large cloud service providers. The successful completion of this process sets the stage for subsequent inference and applications, fueling the rapid growth of the cloud inference market.

1. Cloud Training

The cloud training market holds the largest share of the AI chip market, estimated at 50%-70%. The demand for training large AI models such as GPT-4, Gemini, and Claude allows NVIDIA to maintain its market leadership. The H100’s competitive edge lies in its CUDA ecosystem and TensorRT, while Google TPU leverages its massive internal training clusters and networking technologies. Additionally, companies like AWS with Trainium, Meta, and Anthropic are developing their own training ASICs. While these companies are unlikely to challenge NVIDIA’s market dominance in the short term, they could change the market landscape in the long term.

The market will continue to grow, but the growth rate may slow as AI training remains expensive, and companies are looking for ways to reduce costs. These include techniques such as sparsity, mixed-precision training, more efficient AI acceleration chips, and the reusability of pre-trained models (with improvements in fine-tuning technology).

However, even if technological optimizations reduce the cost of AI training, key considerations remain:

1. companies still need high-performance AI training chips,
2. generative AI use cases are rapidly expanding, and
3. advancements in training technologies and changes in market demand remain uncertain.

Therefore, the training market led by NVIDIA and Google TPU will continue to experience strong demand in the short term.

If DeepSeek can provide innovative AI acceleration technologies and efficient training solutions in the cloud training market, it could challenge the established leaders, such as NVIDIA and Google TPU. Cloud training relies heavily on high-performance chips and energy-efficient technologies to handle large-scale datasets and models. If DeepSeek can offer low-cost, high-performance AI training solutions, particularly by making breakthroughs in sparsity techniques and mixed-precision training, it has the potential to disrupt the existing market structure.

After discussing the development of the cloud training market, we now turn to explore the cloud inference market. Although these two markets may seem distinct, they are actually closely related. Cloud training involves large-scale computations based on vast datasets, while cloud inference applies the trained models to real-world business scenarios. Therefore, the development of both markets is interdependent.

2. Cloud Inference

The rapidly growing cloud inference market has become the second-largest segment of the AI chip market, accounting for approximately 15-25% of the market share. This includes applications such as chatbots, speech recognition, recommendation systems, computer vision (e.g., autonomous driving perception systems), financial risk assessment, medical diagnostics, and industrial inspection.

Currently, the market is dominated by NVIDIA, although AMD has started gaining some market share through its partnerships with Meta and Microsoft. Meanwhile, Intel is maintaining its competitive edge in the inference market with its Habana Gaudi AI chips and Xeon CPUs. Tenstorrent is actively developing its own AI accelerator, attempting to challenge NVIDIA’s leadership position. Qualcomm is focusing on low-power AI inference applications, launching the Cloud AI 100 accelerator, while Mythic explores analog computing technologies, which may further impact the AI inference market in the future. Cerebras and Groq are also competing by renting out inference compute power through self-built small cloud service providers (CSPs), adding more options to the market.

Looking ahead, the cloud inference market is expected to grow significantly, driven by the following key factors:

1. Advances in AI models: This is the most critical factor. As more advanced AI models (such as OpenAI’s GPT-5 and DeepSeek’s R1) are released, there will be a significant increase in the demand for inference computing, which will, in turn, drive the overall market development.
2. Emerging application scenarios: Applications like autonomous driving, financial risk assessment, and medical diagnostics will greatly drive the demand for inference computing. These applications not only expand the scope of AI use but also increase the demand for efficient inference, further stimulating market growth.
3. Reduction in inference costs: As inference costs decrease, more companies will enter the market and apply AI technologies in more fields, which will be a long-term driver of market growth.
4. Innovation and improvement in hardware infrastructure: As specialized hardware for AI inference (such as AI ASICs and FPGAs) is developed, inference efficiency will greatly improve, further reducing costs and enhancing performance.
5. Stricter data privacy and security requirements: As AI is applied in sensitive fields like finance and healthcare, increasing requirements for data privacy and security will drive demand for more efficient and regulation-compliant inference technologies. This could also promote the development of related technologies, such as encrypted inference and edge computing.
6. Enhancement of traditional software by AI: AI’s enhancement of traditional software can significantly improve the performance of existing systems, providing more efficient tools for traditional businesses and possibly giving rise to new business models. Although the impact of this factor is indirect, it plays a stabilizing role in driving long-term market growth.

As AI technology continues to evolve, the demand for more efficient and advanced inference computing will intensify, leading to a surge in demand for inference models. We expect that the computational demand for inference models could exceed the current demand for large language models (LLMs) by more than ten times. This change in demand will drive CSPs to accelerate the development of self-designed ASIC chips to improve inference performance and reduce reliance on third-party hardware (like NVIDIA). For example, AWS aims for 50% of its chips to be self-designed ASICs, Meta plans 70%, and Microsoft could reach up to 80%.

Moreover, as the demand for efficient inference computing increases in AI applications like autonomous driving and medical diagnostics, DeepSeek has the potential to become a strong competitor by introducing accelerators or software solutions that surpass existing technologies.

This would have a significant impact on NVIDIA, which currently dominates the market. As Google’s TPUs, AWS’s Inferentia, and other specialized ASIC technologies gradually increase their market share, more competitors will join the market, making the competition increasingly fierce.

After understanding the market dynamics of cloud inference, the next piece will focus on the development of the edge market, another field that complements cloud inference. While edge training and inference are independent of the cloud market, they offer more efficient and low-latency solutions in many situations, especially in industries with high data privacy requirements, where edge solutions hold unmatched advantages.

This article is part of our Future Scenarios and Design series.
It explores how possible futures take shape through trend analysis, strategic foresight, and scenario thinking, including shifts in technology, consumption, infrastructure, and business models.

See more in this category, or explore more notes here.