The race to dominate generative video has quietly shifted from experimental curiosity to strategic battlefield, and at the center of this transition stands VEO. Google’s flagship video generation model has moved far beyond early demos into something that increasingly resembles a production-grade creative engine. For a tech-savvy audience watching the intersection of AI, media, and infrastructure, VEO is not just another model release—it is a signal of where synthetic video is heading, and how quickly that future is arriving.

The State of VEO Today: From Prototype to Platform

VEO’s latest iteration represents a decisive leap from its earlier versions, which were already impressive but limited in consistency, temporal coherence, and narrative control. The newest release focuses heavily on stability over long sequences, semantic understanding of prompts, and cinematic realism.

Where earlier models struggled with maintaining identity across frames—faces morphing, objects drifting, physics breaking—VEO now demonstrates significantly improved temporal consistency. Characters persist. Lighting remains coherent. Motion flows naturally rather than jittering between frames. These are not cosmetic upgrades; they are foundational shifts that move AI video from novelty to usability.

Equally important is prompt fidelity. VEO understands not just what you ask for, but how elements relate within a scene. Complex prompts involving camera movement, emotional tone, and environmental interaction are now interpreted with far greater accuracy. A request like “a handheld camera tracking a runner through a foggy forest at dawn with cinematic depth of field” produces something recognizably intentional, not a collage of loosely related visual guesses.

This improvement in semantic grounding is what separates VEO from earlier generation tools. It signals a model trained not just on frames, but on storytelling structure.

Measuring Success: Adoption, Output Quality, and Industry Signals

The success of VEO’s latest version cannot be measured purely by benchmarks, because the most meaningful indicators are emerging from usage patterns and industry reactions.

Internally, Google has positioned VEO as part of a broader creative ecosystem, integrating it with tools that allow editing, sequencing, and multimodal input. Early adopters—ranging from marketing teams to independent creators—are using VEO not just for experimentation, but for pre-visualization, prototyping, and even final outputs in certain contexts.

What stands out is the reduction in friction between idea and execution. Creators no longer need to storyboard extensively before generating visual material. Instead, VEO allows iterative prompting, where concepts evolve dynamically through interaction with the model.

In terms of output quality, the latest version consistently produces:

Highly detailed environments with realistic lighting behavior
Smooth camera motion including pans, tilts, and tracking shots
Improved human anatomy and facial coherence
Better handling of complex physical interactions like water, smoke, and fabric

These capabilities place VEO firmly in the top tier of generative video systems, competing directly with leading models from OpenAI, Runway, and emerging startups.

What’s Coming Next: Roadmap and Emerging Capabilities

The most interesting aspect of VEO is not what it does today, but what its trajectory suggests.

Google’s development direction points toward three major expansions: longer-form video generation, deeper controllability, and integration with real-world production pipelines.

Long-form generation is the most obvious next step. Current outputs are still relatively short, but improvements in memory architecture and temporal modeling are paving the way for minute-long or even multi-minute sequences. This is not just about length; it’s about maintaining narrative continuity across extended timelines.

Control is the second frontier. While VEO already interprets prompts well, the next iteration is expected to introduce structured control layers. These may include timeline-based editing, scene segmentation, and parameterized adjustments for elements like lighting, camera lens, and motion dynamics. Essentially, VEO is moving toward becoming a virtual film engine rather than a simple generator.

The third area is integration. Google is uniquely positioned to connect VEO with its broader ecosystem, including cloud infrastructure, creative software, and possibly even YouTube workflows. This could enable seamless transitions from generation to distribution, fundamentally altering how video content is produced and published.

VEO vs. the Competition: A Shifting Landscape

The competitive landscape in AI video generation is evolving rapidly, and VEO’s position must be understood in relation to its rivals.

OpenAI’s Sora has captured significant attention for its ability to generate highly realistic, physics-aware scenes with strong narrative coherence. Sora’s strength lies in its understanding of real-world dynamics and its ability to simulate them convincingly. However, VEO is closing the gap, particularly in areas like prompt interpretation and stylistic flexibility.

Runway’s Gen-2 and subsequent models have focused on accessibility and creative tooling. Runway excels in offering user-friendly interfaces and real-time editing capabilities, making it popular among creators who prioritize workflow integration. VEO, by contrast, is still more model-centric, though this is likely to change as Google expands its tooling.

Other competitors, including Pika and Stability AI’s video initiatives, are pushing innovation in specific niches such as stylized animation and open-source experimentation. These models often prioritize flexibility and community-driven development over raw performance.

What distinguishes VEO in this crowded field is its balance. It does not dominate every category, but it performs consistently well across all major dimensions: realism, coherence, scalability, and prompt understanding. This makes it a generalist powerhouse rather than a specialized tool.

The Types of Video VEO Can Create

One of VEO’s defining characteristics is the breadth of video styles it can generate. Unlike earlier systems that were confined to short, abstract clips, VEO operates across a wide spectrum of visual genres.

Cinematic storytelling is perhaps its most compelling use case. VEO can produce scenes that resemble film footage, complete with dynamic camera work, atmospheric lighting, and emotional tone. This makes it valuable for pre-visualization in filmmaking, where directors can explore ideas before committing to expensive shoots.

Commercial and advertising content is another area where VEO excels. The ability to generate polished, brand-like visuals opens up new possibilities for rapid campaign development. Marketers can test multiple concepts in hours rather than weeks.

VEO also handles abstract and artistic styles effectively. From surreal landscapes to stylized animations, the model can adapt its output to match a wide range of aesthetic directions. This flexibility is crucial for creative experimentation.

Educational and explanatory videos represent a more functional application. VEO can generate visualizations of concepts, processes, and scenarios, potentially transforming how complex information is communicated.

Finally, there is the emerging category of hybrid content—videos that blend real and synthetic elements. VEO’s ability to integrate with other AI tools makes it possible to combine generated footage with real-world inputs, creating entirely new forms of media.

The Biggest Advantage: Semantic Intelligence Meets Scale

If one had to identify VEO’s single greatest advantage, it would be the combination of semantic intelligence and infrastructure scale.

Semantic intelligence refers to the model’s ability to understand context, relationships, and intent within prompts. This goes beyond recognizing objects; it involves interpreting how elements interact within a scene. VEO’s improvements in this area are what make its outputs feel intentional rather than accidental.

Scale, on the other hand, is about compute and data. Google’s infrastructure allows VEO to be trained on vast datasets and deployed at levels that smaller competitors cannot match. This translates into faster iteration cycles, more robust performance, and the ability to handle increasingly complex tasks.

The intersection of these two factors creates a powerful feedback loop. Better understanding leads to better outputs, which in turn generate more data for further training. Over time, this compounds into a significant competitive advantage.

Limitations and Challenges

Despite its progress, VEO is not without limitations. Understanding these is crucial for assessing its future trajectory.

Temporal coherence, while improved, is still not perfect. Long sequences can exhibit subtle inconsistencies, particularly in highly dynamic scenes. This is an area where further research is needed.

Another challenge is controllability. While VEO interprets prompts well, fine-grained control over specific elements remains limited. Creators often need to iterate multiple times to achieve the desired result, which can be inefficient.

Ethical and regulatory concerns also loom large. The ability to generate realistic video raises questions about misinformation, deepfakes, and content authenticity. Google has implemented safeguards, but the broader implications are still unfolding.

Finally, there is the issue of cost. High-quality video generation requires significant computational resources, which may limit accessibility for smaller users unless pricing models evolve.

Strategic Implications: Beyond Content Creation

The rise of VEO has implications that extend far beyond video production.

In media and entertainment, it challenges traditional production models. If high-quality video can be generated quickly and cheaply, the barriers to entry for content creation drop dramatically. This could lead to an explosion of new voices, but also increased competition and content saturation.

In marketing, VEO enables hyper-personalized campaigns. Brands can generate tailored video content for different audiences, optimizing engagement at a granular level.

In education and training, the ability to create dynamic visual content on demand could transform learning experiences. Complex concepts can be illustrated in ways that static media cannot achieve.

Even in enterprise contexts, VEO has potential applications in simulation, design, and communication. From architectural visualization to scenario planning, the possibilities are vast.

The Road Ahead: From Tool to Medium

What makes VEO particularly significant is that it is not just a tool—it is the foundation of a new medium.

As generative video models continue to evolve, they will enable forms of storytelling that were previously impossible. Interactive narratives, real-time content generation, and AI-assisted filmmaking are just the beginning.

The distinction between creator and audience may blur as users gain the ability to generate personalized content on demand. This raises profound questions about authorship, ownership, and the nature of creativity itself.

VEO’s role in this transition is still unfolding, but its trajectory is clear. It is moving from a model to a platform, from a feature to an ecosystem.

Conclusion: A Defining Moment for AI Video

VEO’s latest version marks a turning point in the evolution of generative video. It demonstrates that AI can produce not just visually impressive clips, but coherent, intentional, and usable content.

Its success lies in its balance—combining realism, flexibility, and scalability in a way that few competitors can match. Its future lies in deeper control, longer narratives, and tighter integration with creative workflows.

For those watching the convergence of AI and media, VEO is more than a technological milestone. It is a glimpse into a future where video is no longer captured, but created—where imagination is limited not by resources, but by ideas.

And if the current trajectory holds, that future is arriving faster than most are prepared for.

The financial world is no stranger to hype cycles. From algorithmic trading to blockchain and generative AI, each technological wave has promised to reshape markets—and often has. But every so often, a new system emerges that doesn’t just promise efficiency or speed, but something more unsettling: autonomy with intent. The latest entrant into this category is an AI system referred to as “Mythos,” and it is already sending ripples through Wall Street, venture capital circles, and regulatory bodies alike.

What makes Mythos different is not simply its technical sophistication. It is the growing perception that this class of AI is no longer just a tool, but an actor—one capable of independent strategic reasoning across complex systems. And that distinction is where excitement begins to blur into concern.

Beyond Generative AI: A Shift Toward Strategic Intelligence

For the past few years, the AI narrative has been dominated by large language models and generative systems capable of producing text, images, and code. Systems like ChatGPT and their competitors have demonstrated impressive fluency, but their limitations are well understood: they predict rather than plan, respond rather than initiate.

Mythos represents a departure from that paradigm. Instead of focusing on output generation, it is designed to operate across decision layers. In financial contexts, this means analyzing markets, identifying opportunities, executing trades, and dynamically adjusting strategies in real time—not as isolated actions, but as part of a coherent long-term objective.

This evolution reflects a broader shift in artificial intelligence toward what some researchers describe as “agentic systems.” These are AIs that can pursue goals, adapt to feedback, and operate with a degree of persistence previously reserved for human actors or tightly controlled algorithms.

The implications for financial markets are profound. Traditional algorithmic trading systems are powerful, but they operate within predefined rules. Mythos-like systems blur those boundaries, potentially creating strategies that evolve faster than human oversight can track.

Why Wall Street Is Both Investing—and Worrying

It’s not surprising that major financial institutions are paying close attention. The potential upside is enormous. An AI that can synthesize global data streams, anticipate market movements, and execute complex strategies could deliver a decisive competitive edge.

But with that potential comes a new category of risk.

One concern is opacity. As AI systems grow more complex, their decision-making processes become harder to interpret. Even today, firms struggle to fully explain the behavior of advanced machine learning models. With systems like Mythos, the challenge intensifies: if an AI is optimizing across multiple variables simultaneously, its reasoning may not map cleanly onto human logic.

Another issue is systemic risk. Financial markets are deeply interconnected, and the widespread adoption of similar AI systems could lead to synchronized behavior. If multiple institutions deploy comparable models, they may react to signals in the same way, amplifying volatility rather than dampening it.

This is not a hypothetical scenario. The Flash Crash demonstrated how automated systems can interact in unpredictable ways, triggering rapid market declines. Mythos-level AI could magnify such dynamics, especially if operating at greater speed and complexity.

The Question of Control

Perhaps the most unsettling aspect of Mythos is not what it can do, but how much control humans retain over it.

In traditional financial systems, human oversight remains a critical safeguard. Traders set parameters, risk managers enforce limits, and regulators monitor compliance. But as AI systems become more autonomous, the locus of control begins to shift.

If an AI is continuously learning and adapting, static rules may become insufficient. By the time a human intervenes, the system may have already moved beyond the original framework. This raises fundamental questions about accountability. If an AI-driven strategy causes significant losses—or worse, destabilizes a market—who is responsible?

These concerns echo broader debates in artificial intelligence, particularly around alignment. Ensuring that AI systems act in accordance with human intentions is a central challenge, and one that becomes more complex as systems gain autonomy.

Organizations like OpenAI and DeepMind have invested heavily in alignment research, but much of that work is still in its early stages. Applying those principles to high-stakes environments like finance adds another layer of urgency.

A Broader Context: The Convergence of AI and Capital

To understand the significance of Mythos, it helps to zoom out. What we are witnessing is not just a technological development, but a convergence of trends.

First, the scale of data available to financial institutions has exploded. From real-time market feeds to alternative data sources like satellite imagery and social media sentiment, the information landscape is richer—and more chaotic—than ever.

Second, computational power continues to grow, enabling more complex models to operate at scale. Advances in hardware, particularly GPUs and specialized AI chips, have lowered the barriers to deploying sophisticated systems.

Third, the competitive dynamics of finance create strong incentives for adoption. In an industry where milliseconds can translate into millions of dollars, the pressure to leverage cutting-edge technology is relentless.

Mythos sits at the intersection of these forces. It is not an isolated innovation, but a product of a broader ecosystem that is pushing AI toward greater autonomy and influence.

The Regulatory Catch-Up Game

Regulators are now faced with a familiar dilemma: how to oversee a rapidly evolving technology without stifling innovation.

Historically, financial regulation has struggled to keep pace with technological change. The rise of high-frequency trading, for example, prompted years of debate before meaningful frameworks were established.

With AI systems like Mythos, the challenge is even greater. Traditional regulatory approaches rely on transparency and auditability, but these are precisely the areas where advanced AI systems are weakest.

There is also an international dimension. Financial markets are global, and AI development is taking place across multiple jurisdictions. Coordinating regulatory efforts will require unprecedented levels of cooperation.

Some policymakers are already exploring new approaches, including requirements for model explainability, stress testing of AI systems, and real-time monitoring of algorithmic behavior. But these measures are still in their infancy.

Fear, Hype, and Reality

It’s important to separate legitimate concerns from exaggerated fears. Not every advanced AI system represents an existential threat, and much of the discourse around Mythos is shaped by speculation.

At the same time, dismissing these concerns outright would be a mistake. History shows that transformative technologies often bring unintended consequences. The key is not to halt progress, but to manage it responsibly.

In this sense, Mythos can be seen as a test case. It forces us to confront questions that extend beyond finance: how much autonomy should we grant to machines, how do we ensure accountability, and what safeguards are necessary in a world where AI systems play an increasingly active role?

The Strategic Implications for Investors and Institutions

For investors, the rise of systems like Mythos introduces both opportunities and challenges.

On one hand, firms that successfully integrate advanced AI could achieve significant performance gains. On the other hand, the competitive landscape may become more volatile, with rapid shifts driven by algorithmic strategies.

This dynamic could also reshape the role of human expertise. Rather than making direct decisions, traders and analysts may increasingly focus on supervising AI systems, interpreting their outputs, and managing risk at a higher level.

Institutions will need to invest not only in technology, but also in governance. This includes developing frameworks for oversight, ensuring diversity in model design to avoid systemic convergence, and maintaining a clear understanding of how AI systems operate within their organizations.

A Glimpse Into the Future

Looking ahead, it’s likely that Mythos is just the beginning. As AI continues to evolve, we can expect more systems that combine data analysis, strategic reasoning, and autonomous execution.

The financial sector will serve as a proving ground, but the implications will extend far beyond it. Similar systems could emerge in areas like logistics, energy management, and even geopolitical strategy.

The central question is not whether these systems will be developed—they already are—but how they will be integrated into existing structures.

Conclusion: Between Power and Prudence

The story of Mythos is ultimately a story about transition. We are moving from an era where AI augments human decision-making to one where it increasingly participates in it.

This shift brings immense potential, but also new risks that cannot be ignored. For Wall Street, the challenge will be to harness the capabilities of systems like Mythos without losing control over the processes they influence.

For the broader world, the stakes are even higher. As AI systems become more autonomous, the need for thoughtful design, robust oversight, and ethical consideration becomes paramount.

Mythos may not be the final form of this evolution, but it is a clear signal of where things are heading. And for those paying attention, it raises a simple but urgent question: are we building tools—or counterparts?

The most dangerous players in emerging tech are rarely the loudest ones. While much of the crypto-AI narrative is dominated by hype cycles, token pumps, and overpromised infrastructure, Seedance 2 has been moving with a very different rhythm—measured, deliberate, and increasingly dominant. Over the past months, whispers around the project have grown louder: internal upgrades, strategic partnerships, and a roadmap that—if even partially accurate—could reshape how decentralized intelligence networks are deployed at scale.

Seedance 2 is no longer just “one of the leaders.” It is becoming the benchmark.

From Underdog to Market Benchmark

Seedance didn’t start as the obvious frontrunner. Early iterations of the project were viewed as technically ambitious but commercially uncertain. The core thesis—combining decentralized compute, adaptive AI models, and tokenized incentive structures—was compelling, but so were dozens of similar narratives across the market.

What changed with Seedance 2 was execution.

The second-generation architecture stripped away much of the experimental overhead that plagued earlier decentralized AI systems. Instead of trying to solve everything at once, the team narrowed its focus: efficient compute allocation, scalable model orchestration, and real economic incentives for node operators. The result is a system that actually works under real-world load conditions—something many competitors still struggle to demonstrate convincingly.

Today, Seedance 2 is widely considered the most operationally mature platform in its category. Not the most hyped. Not the most speculative. But the most functional.

The Core Advantage: Adaptive Compute Markets

At the heart of Seedance 2 lies a concept that sounds simple but is extraordinarily difficult to execute: adaptive compute markets.

Traditional decentralized compute networks operate on static pricing or loosely optimized supply-demand matching. Seedance 2 introduces a dynamic layer where compute resources are continuously repriced based on real-time demand signals, model complexity, latency requirements, and network congestion.

This creates several cascading advantages.

First, it dramatically improves efficiency. Idle compute is minimized because pricing adjusts fast enough to attract demand. Second, it aligns incentives in a way that feels closer to high-frequency financial markets than traditional blockchain systems. Node operators are not just passive providers; they are active participants in a constantly evolving marketplace.

And third, it enables something most AI networks fail to deliver: predictable performance.

In decentralized environments, unpredictability is the norm. Seedance 2 flips that narrative by making unpredictability itself a variable that can be priced, hedged, and optimized.

Rumored Upgrades: What’s Coming Next?

While the team has remained relatively tight-lipped, several consistent leaks and insider discussions point to a series of major upgrades currently in late-stage development.

1. Modular AI Pipelines

One of the most talked-about upcoming features is the introduction of modular AI pipelines. Instead of deploying monolithic models, developers will be able to chain specialized micro-models across the network.

This is a significant shift.

Rather than running a single large model that handles everything from input parsing to output generation, Seedance 2 would allow distributed specialization. One node cluster might handle natural language understanding, another handles reasoning, and another handles output formatting.

The implications are massive. It reduces computational overhead, improves scalability, and allows for continuous optimization at each stage of the pipeline.

More importantly, it creates a marketplace not just for compute—but for intelligence itself.

2. Latency-Sensitive Routing

Another rumored feature is latency-sensitive routing, designed to address one of the biggest criticisms of decentralized AI: speed.

In centralized systems, latency is tightly controlled. In decentralized systems, it can vary wildly depending on node location, network conditions, and workload distribution.

Seedance 2 is reportedly implementing a routing layer that dynamically selects compute nodes based on latency thresholds defined by the application. This would allow high-frequency use cases—like trading bots or real-time AI assistants—to operate within strict performance constraints.

If executed properly, this could unlock entirely new categories of applications that were previously considered impractical on decentralized infrastructure.

3. On-Chain Model Reputation Systems

Trust remains one of the hardest problems in decentralized AI. How do you know a model is performing as advertised? How do you verify output quality in a trustless environment?

The answer, according to multiple sources, is an on-chain reputation system for models.

Each model instance would accumulate performance metrics over time—accuracy, response time, user feedback, and even economic efficiency. These metrics would be recorded and made accessible, allowing developers to choose models based on transparent performance histories.

This effectively introduces a meritocratic layer to the network. The best models rise not through marketing, but through measurable results.

Inside Signals: What Insiders Are Saying

While official announcements remain sparse, conversations among early contributors, node operators, and ecosystem partners paint a clear picture: Seedance 2 is preparing for a major expansion phase.

There are three consistent themes emerging from insider chatter.

The first is confidence. Not the speculative kind, but the operational kind. Contributors describe a system that is already handling workloads far beyond what is publicly disclosed. This suggests that much of the platform’s real capacity is still under the radar.

The second is institutional interest. While retail narratives dominate public discourse, there are increasing signs that enterprise players are quietly testing Seedance 2’s infrastructure. These are not headline-grabbing partnerships—at least not yet—but pilot programs, integrations, and backend experiments.

The third is timing. Several insiders hint that the next major update cycle is aligned with broader market conditions, suggesting that Seedance 2 is not just building in isolation but positioning itself strategically within the macro crypto cycle.

Performance Metrics: Quiet Dominance

Unlike many projects that rely heavily on token price as a proxy for success, Seedance 2’s real strength lies in its usage metrics.

Network throughput has reportedly increased several-fold over the past quarter, with a corresponding rise in active node participation. More importantly, the ratio between supply (compute providers) and demand (AI workloads) appears to be stabilizing—a key indicator of a healthy network.

In many decentralized systems, supply far exceeds demand, leading to underutilized resources and weak economic incentives. Seedance 2 seems to be approaching equilibrium, where both sides of the market are actively engaged.

This balance is what transforms a project from an experiment into infrastructure.

Competitive Landscape: Why Seedance 2 Is Pulling Ahead

The decentralized AI space is crowded, but most competitors fall into one of two categories.

The first group focuses heavily on theoretical capabilities—massive model sizes, complex architectures, and ambitious roadmaps. The problem is that these systems often struggle with real-world deployment.

The second group prioritizes simplicity but lacks the depth needed to handle advanced AI workloads.

Seedance 2 occupies a rare middle ground.

It is technically sophisticated enough to support complex applications, yet pragmatic enough to deliver consistent performance. This balance is difficult to achieve and even harder to maintain.

Another key differentiator is economic design. Many projects treat tokenomics as an afterthought. Seedance 2 treats it as core infrastructure. Incentives are not just aligned—they are continuously optimized.

This creates a feedback loop where network growth reinforces economic stability, which in turn attracts more participants.

The “King” Narrative: Is It Justified?

Calling any project the “king” of a fast-moving sector is always risky. Markets evolve quickly, and today’s leader can become tomorrow’s cautionary tale.

That said, the label is not entirely undeserved.

Seedance 2 currently leads in three critical areas: usability, performance, and economic coherence. These are not flashy metrics, but they are the ones that matter when moving from experimentation to adoption.

However, dominance brings its own challenges.

As the network grows, maintaining decentralization becomes more difficult. Larger players may attempt to consolidate control over compute resources. Regulatory scrutiny could increase, especially as institutional involvement deepens.

And perhaps most importantly, expectations rise.

Seedance 2 is no longer judged against its past—it is judged against its potential.

Strategic Implications for the Market

The rise of Seedance 2 signals a broader shift in the AI–crypto landscape.

We are moving away from purely speculative narratives toward systems that deliver tangible utility. The market is beginning to reward execution over ambition, and infrastructure over ideology.

This has several implications.

Developers are likely to gravitate toward platforms that offer reliability and scalability. Investors may start prioritizing usage metrics over token hype. And competitors will be forced to either catch up or differentiate in entirely new ways.

In this context, Seedance 2 is not just a project—it is a signal of where the industry is heading.

What to Watch Next

The next phase for Seedance 2 will be defined by its ability to scale without losing its core advantages.

If the rumored upgrades—modular pipelines, latency-sensitive routing, and reputation systems—are successfully deployed, the platform could extend its lead significantly.

At the same time, external factors will play a crucial role. Market conditions, regulatory developments, and technological breakthroughs in adjacent fields could all influence the trajectory.

But perhaps the most important variable is execution.

So far, Seedance 2 has demonstrated an ability to deliver where others have stalled. If that pattern continues, the project may not just remain at the top—it could redefine what “top” means in this space.

Final Take: Momentum With Substance

There is a difference between momentum driven by hype and momentum driven by substance.

Seedance 2 clearly belongs to the latter category.

It is not the loudest project. It does not rely on constant announcements or aggressive marketing. Instead, it builds, iterates, and quietly expands its footprint.

In a market often defined by noise, that approach stands out.

Whether it ultimately becomes the long-term leader of the decentralized AI ecosystem remains to be seen. But as of now, the combination of technical execution, economic design, and strategic positioning makes one thing clear:

Seedance 2 is not just participating in the race.

It is setting the pace.