BlueFocus CEO Letter to Investors

Dear Investors: This year marks the 30th anniversary of BlueFocus. We started as a small workshop of just over a dozen people and have grown into one of the world's top ten marketing communications groups. Thirty years has not been easy, but it has been extraordinary! We are grateful for this magnificent era of sweeping change, grateful for the trust of our clients, grateful for the support of our shareholders and partners, and grateful for every colleague who has fought alongside us. You are the ones who have forged the true soul of BlueFocus. Thirty years of achievement are only dust and earth. At 30, BlueFocus now has one core mission for the future: self-reinvention and a new entrepreneurial beginning! More clearly than ever, we understand that AI will either make us or destroy us. That is precisely why we must start again. We must have the courage to remake ourselves and break the paths and inertia that once made us successful. What we want to create is not an old company with higher efficiency, but a new BlueFocus built for the future: a genuine AI (marketing) technology company, a truly global company, a new culture, a new engine, a new governance model, and a new belief. Marketing is the industry most thoroughly rewritten by AI, without exception. If we execute well, then under the twin waves of AI and globalization, we have the opportunity to truly break through the boundaries and ceiling of BlueFocus's development. We will no longer be merely a marketing company, no longer merely an agency, no longer a labor-intensive company, but a global technology company. And we will have the opportunity to stand alongside the best companies in the world and create new value for the industry, for China, and for the world. That is exciting enough, and meaningful enough, to deeply inspire both me and my team. This year marks the completion of BlueFocus's third full year of All in AI. We moved beyond the shallow stage of using AI tools simply to cut costs and improve efficiency long ago. We did not linger for too long among the low-hanging fruit. Instead, we moved into the stage of using truly AI Native methods to reconstruct our business, build new models, build new organizations, and create new intelligence capable of autonomous decision-making. With that in mind, I will develop this year's letter around that standard and ground it in real data. I would also ask you to judge BlueFocus's early achievements by that same AI Native standard: are we moving closer to, or further away from, what it means to become a new technology company? In my mind, that matters far more than short-term net profit, and we will never become enamored with any short-term success. Today, BlueFocus has entered an era of trillion-scale token consumption, and tokens have become the new fuel and new infrastructure of intelligence. In 2025, our total token consumption through API access to large models exceeded 1 trillion for the first time. In the first quarter of 2026, that number continued to grow by multiples. In March alone, tokens consumed by reasoning models exceeded 200 billion, while the estimated token volume for generative models was roughly six to eight times higher. On that basis, our estimated token consumption for full-year 2026 will increase even further. There is a simple way to judge whether a company has truly embarked on the path toward AI intelligence: do not look at how many case studies it has published, how many articles it has written, or how many thin-wrapper products it has launched. Those things can all be packaged. Go directly to its API token consumption. Token volume does not lie. It is the most honest starting indicator of whether AI has truly entered the business process. Without token consumption, all claims of All in AI are empty. Only when token consumption is real can we begin to discuss depth, and only then can we begin to discuss intelligence. Behind these massive token numbers is the fact that AI has truly entered every capillary of BlueFocus's business, from social media insights, creator analysis, and advertising risk control, to winning-creative extraction, intelligent budget adjustment, and video content production. AI is participating with unprecedented depth and density. In building AI capability, we have gone through three stages: using tools, building tools, and building infrastructure. Many companies remain at the first level. Some have reached the second. Very few have truly entered the third. In 2026, BlueFocus will stand firmly at the third level and systematically build its own AI infrastructure. This year, we will also officially launch Blue AI Skills Market and build BlueFocus's own Agent Cloud. Over the past two years, I have imagined countless times whether BlueFocus could one day become a "cloud platform" company. My answer is that anything is possible, and the rest should be left to time. Beyond tokens, AI-driven revenue is also growing rapidly: USD 175.912 million in 2024 and USD 546.05 million in 2025. What is "AI-driven revenue"? It is revenue generated with minimal human input, extremely high levels of AI automation, greater gross margin potential, and newer business models, while also allowing data to be accumulated, annotated, iterated, and continuously learned from. Our objective going forward is very clear: to systematically reconstruct BlueFocus's existing business with AI Native methods so that every business scenario becomes AI-driven. Today, that proportion is only around 5 percent. It will grow substantially from here. The outlook is broad, the path is clear, and by this time next year, we will let the numbers speak again. What matters more than revenue and profit growth? Genuine AI intelligence. We are building a truly end-to-end AI marketing platform with autonomous decision-making, self-learning, and automation. Today, BlueFocus's AI is already capable of executing tasks at scale. The next step is to enable AI to make judgments and decisions on its own, moving toward truly end-to-end intelligence. Across BlueFocus today, two platforms are positioned to be the first to achieve genuine full-chain intelligence. One is the Blue AI platform that everyone already knows well. It is the "intelligent engine" built by BlueFocus people ourselves, and all of our strongest resources are concentrated there. In March alone, Blue AI's self-developed intelligent media-buying system experienced what we consider our own "iPhone moment": with no human intervention, Blue AI outperformed humans across most core operating scenarios, including strategy formulation, budget allocation, and delivery decisions, covering 85 percent of use cases. Put simply, you set the objective, AI generates the strategy and executes it, allocates the budget autonomously, and humans are responsible only for calibrating the final ROI. Another number we are proud of is that in 2025, A2A collaborative tasks on Blue AI reached 146 million. The meaning of that number is that a great deal of optimization and iteration no longer requires humans to pass instructions between steps. AI is evolving on its own. When AI can trace back through its execution process, evaluate the quality of its own output, and proactively adjust its next move, it is no longer just a tool being called. It becomes a genuinely evolving, intelligent, end-to-end system. Another platform likely to achieve a true full-chain closed loop is STARUNION AI, an AI Native platform focused on creator marketing. Today, STARUNION AI can already handle market insight, creator matching, content strategy, batch outreach, post-campaign tracking, and performance analysis, with the full chain no longer requiring human step-by-step driving. In 2025, order volume on the STARUNION AI platform exceeded USD 146.593 million, while order volume per media professional increased by 51 percent. This is not an optimization of labor productivity. It is a replacement of the operating model itself. As AI intelligence continues to advance, it is easy to imagine that the underlying logic of BlueFocus as a company will also change fundamentally. In the past, we were a company driven by labor intensity. In the future, we will be a company driven by AI intelligence. Blue AI has already become BlueFocus's new intelligent engine, and we hope it will go even further and become the world's best example of an "AI application engine." When the time is right, we will consider opening it to the broader industry in the form of a "cloud" offering and help drive the intelligent upgrading of the entire industry. We have never forgotten that as a content company, AI Native multimodal capability requires both extreme efficiency and advanced aesthetic quality. In the near term, we have only two goals in multimodal capability. The first is extreme intelligence and automation, allowing content to be generated quickly, accurately, and at scale. The second is top-tier AI production capability at a Hollywood level, capable of moving people and giving content soul, warmth, and emotional force. Both are indispensable. For the first, we have Blue AI Xinying Platform and the "DDD End-to-End" AI Native platform. For the second, we have Blue Master and Blue Minds Box. Last year, Blue AI produced 500,000 video assets, and in the first quarter of 2026 alone, it produced another 300,000. Behind these numbers is the fact that Blue AI, with greater stability and deeper contextual understanding, has truly entered the core processes directly tied to revenue. I would now like to focus on the second of those two goals: cinematic-grade AI content production. Over the past year, we have built a full-chain creative system connecting multi-model collaboration, 3D, and motion capture, while continuing to make breakthroughs in long-shot orchestration, emotional performance, and multi-character consistency. We delivered truly cinematic-level AI content for leading brands such as Pechoin, Jeep, and Dongfeng eπ. These were not demos. They were real works tested in the market. At the same time, our self-produced AI film "New Human Plan: Solon" won top awards at major international competitions in Hollywood, Cannes, and Rome. "Starlight Moment," a production we participated in, was nominated for Best Children's Film at the 38th China Golden Rooster Awards. Most recently, we reached a milestone moment in the fullest sense: the Blue Master team officially secured an S-tier AI content production project from the film arm of a platform company, billed by the minute. "Billed by the minute" means that our AI production capability has reached the professional standard required for cinematic-grade delivery. It also redefines the boundaries of AI multimodal capability. We are no longer just helping brand clients create content. We are beginning to move into theaters, and into the world's best films and television dramas. Let me now turn to the AI Native new models that matter most to me. What is an AI Native new model? It is a model that could not exist without deep AI involvement, or one that simply did not exist before. The first new model is Generative Engine Optimization. As of the first quarter of this year, 42 brand clients had already entered the service pipeline, including leading brands such as Alibaba, L'Oréal Group, Kimberly-Clark, and Starbucks. We expect revenue from this specialized business to reach USD 2.199 million to USD 2.932 million this year. This is a new battlefield where brands compete for mindshare in the AI Native world. Brands must not only win consumer consideration, but also ensure that AI has a correct and timely understanding of them. We will only engage in positive, long-term Generative Engine Optimization. It has nothing to do with the previously exposed so-called "AI poisoning," which is exactly why so many top-tier brands trust us. The second new model is AI-driven operational services. Blue AI and STARUNION AI are undergoing a fundamental shift, from selling projects to selling systems, and from one-time delivery to continuously online intelligent operating agents. What clients are buying is no longer a single project, but a continuously learning, continuously optimizing, continuously value-creating AI service system. This is the next true form of service after SaaS. The third new model is real-time generation, with a thousand people and a hundred million variations. This is the direction I find most imaginative, and it also represents the deepest rewrite of the marketing industry. The old logic was one version for everyone: content was produced at scale, and algorithms determined who saw it. Personalized recommendation was, in essence, only personalized distribution, while the content itself remained standardized. AI is now making something else possible: in that exact second, content generated in real time for each individual and for that individual alone. Based on this, we are exploring new traffic ecosystems, new forms of content interaction, and new forms of advertising. This is the true frontier that BlueFocus's AI strategy is pursuing. It is also the core driving force behind the AI investment I will discuss next: we want to find and build new models together with the world's best technology companies and the newest entrepreneurial companies. Last year was also the first year in which we strategically concentrated our investment on AI. We have currently invested in six companies, all of them AI Native companies. They span five directions: multimodal video generation, represented by PixVerse; AI creator marketing, represented by Aha Creator; digital employees and agent infrastructure, represented by OpenHex; Generative Engine Optimization, represented by PureblueAI; and AI advertising decision-making, represented by Pepr AI. This is not scattered betting. It is a complete closed loop for AI marketing. At the same time, we also made a strategic investment in AGI House, a fund focused on early AI unicorns, with the goal of helping us discover and connect with the most ambitious AI founders. Behind every investment is deep synergy with BlueFocus's core business. We are not only financial investors. We are also their clients, the proving ground for their models, and a bridge for their global expansion. Going forward, on one hand, we will continue to work with the world's best AI technology companies and large-model companies with an extremely open mindset. On the other hand, we will continue to identify and strategically invest in future AI unicorns, with sustained arrangements in agents, world models, multimodal generation, AI media-buying decision-making, and related areas. Both are critically important to BlueFocus's strategy, and we will stay committed to them. For us, investment has never been merely a financial act. We need to find the fastest-moving AI Native companies and stay close to them, remain in sync with the industry's most advanced capabilities. That in itself is one of the ways BlueFocus accelerates its own transformation. In 2026, on the road to becoming an AI company, we have only one principle: there will be no upper limit on AI investment, and every major decision will follow "AI First." In concrete terms, we will continue investing in the most AI Native companies, not traditional companies with an AI label attached, but genuinely new species that rebuild business logic from first principles. We will continue to work with top model companies and co-build deeply around real business scenarios. We have the scenarios, and they have the capabilities. This is not simple procurement. It is deep co-creation. Employee use of AI will continue without limits. Beyond our self-developed platforms, everyone will be free to choose the best models and tools, and the company will pay for them. Only when people do not have to worry about tool costs can creativity be fully released. We will continue investing in infrastructure, continue deepening our moat, and above all, AI talent density will be the highest priority, with no exception. In the AI era, the ceiling of a company begins with the ceiling of its people, and only then comes everything else! Last year, our pure R&D spending related to AI infrastructure and products reached USD 13.926 million, up 76 percent year over year. If AI technical talent investment is included, that number rises to USD 26.387 million. I can state with absolute clarity that BlueFocus's R&D spending will continue to grow rapidly in the years ahead, and we will continue to invest decisively in AI² and Globalization 2.0, our two critical strategies. I know that even all of the above is still not enough. That is why we are doing something even harder and more fundamental. We are reconstructing BlueFocus into a truly AI Native organization. That is the essential path and the sustainable path. What is an AI Native organization? It is one whose decision-making is AI Native, whose collaboration is AI Native, whose talent standards are AI Native, and whose incentive mechanisms are AI Native. AI is not an add-on. It is part of our genetic code. That is the fundamental divide between us and many other companies. Concretely, we are building a belief system that seeks answers through technology. Today, BlueFocus already has nearly 500 AI talents, including more than 350 people in product and technology-related roles and more than 150 talents who connect business and technology. It is easy to foresee that by the time of next year's letter, the substance and value of that number will have increased rapidly again. In terms of standards, we have established the highest measuring stick: AI Native + AI First. This is the highest standard by which we evaluate a person, a team, and a project. We do not ask whether you used AI. We ask whether you used AI to redefine the thing itself. Within management, 30 percent to 50 percent must be leaders with real AI capability. If they are not, they will be replaced. We have also written AI metrics into the formal performance commitments of every member of management. In other words, what determines compensation over the course of the year will no longer be traditional profit metrics, but AI metrics. At the same time, we have introduced the AI Business Partner mechanism so that AI capability truly penetrates every business line, rather than remaining on PowerPoint slides. On incentives, we mean what we say: "No AI, No Bonus." This is not a slogan. It is a hard rule written into our evaluation system. In culture, we are building an innovative culture that is self-driven, free, and accountable. Over the past year and more, I have flown nearly 180,000 kilometers and spent more than eight hours each day interacting with AI, all for one core purpose: to inject the spirit of entrepreneurship, vitality, and belief in the AI era into this thirty-year-old company! Next, let me turn to the progress of our global outbound business and our Globalization 2.0 strategy! Today, more than 80 percent of BlueFocus's revenue and more than half of its profit come from our global outbound business. Looking three to five years ahead, that business will remain on a trajectory of rapid growth. We have been deeply engaged in globalization business for 11 years. I know there are many people and many companies in the market today claiming leadership in this field. So let me clarify a few objective facts. First, as of today, by total outbound business volume and total market share, we rank No. 1 in China, based on 2024 revenue according to Frost & Sullivan. This leadership extends not only across Meta, Google, and TikTok for Business, but also across mid-tier and other small-to-medium global media platforms such as Pinterest and Snapchat. Second, we have already helped more than 100,000 Chinese companies go global, and our client penetration and industry penetration both remain in leading positions. Third, many of the overseas offices we began building last year have already started to achieve scaled profitability. Finally, we have built a highly capable globalization team, and that is where our confidence for the future comes from. Looking ahead, our global outbound business has only one core priority: fully executing the Globalization 2.0 strategy. What is that strategy? First, we must continue to grow rapidly and maintain our No. 1 market share. Extremely high market share and deep penetration into client budgets form the foundation of Globalization 2.0. Second, we must fundamentally change our revenue and profit structure. To that end, we have introduced the new 5-3-2 profit rule: 50 percent of profit should come from the top three media platforms, 30 percent from mid-tier media platforms, and 20 percent from self-built traffic and technology platforms. Third, we must shift from remote outbound operations to true growth driven by our own on-the-ground offices. Last year, we had seven overseas offices. This year, we expect to have at least ten. The logic behind our overseas office expansion is simple and clear. First, we must dominate the markets closest to us and become the No. 1 brand in Southeast Asia. I am very encouraged to see that our offices in Singapore, Vietnam, Thailand, and Indonesia are developing rapidly, with more than two of them already entering the phase of scaled profitability. Second, we must go to the markets farthest from us, namely Latin America. We expect the Brazil office to achieve new breakthroughs in both revenue and profit in 2026. Third, we must move into the most difficult markets, namely Europe and the United States. Today, our companies in the Netherlands and the United States already have landmark clients, and we have also found the model for breakthrough. We will next establish a UK company. Finally, there are markets with particularly distinctive civilizations and cultures, such as the Middle East. Fourth, we must move from agency-driven growth to technology-driven growth, and from agency traffic to self-built traffic. Our Blue X and Blue Turbo platforms, which we have incubated for more than a year, are expected to generate more than USD 14.659 million in revenue in 2026. It is important to understand that the traffic structure of globalization is different from China's. Beyond the three dominant platforms, at least 30 percent of traffic can still be restructured through technology. That provides fertile ground for our technology-driven and platform-driven growth. Today, self-owned traffic revenue may appear small relative to our overall business, but its quality is exceptionally high, because both its profit profile and technological sophistication are on an entirely different level from agency business. Fifth, we must fully integrate AI. I have said before that AI and globalization are like twin brothers. They are deeply intertwined, and reality has proven that true. In global business scenarios, AI has natural advantages across languages, cultures, and civilizations. In the process of going global, AI will become the first choice. I believe truly AI Native business models are likely to emerge first in the globalization track, and that is strategically significant for us. New AI traffic, new AI models, new AI culture, and more are already being practiced with increasing depth in the globalization arena. A person of great ambition should set out at dawn and still be chasing distant horizons by dusk! For the future of BlueFocus, we will always think from the perspective of the global market, and we will always seek answers through technology. In the era of AI and globalization, if we do not execute well, we do not deserve today's share price at all. If we do execute well, then USD 14.659 billion in revenue, USD 14.659 billion in market capitalization, and the possibility of becoming a respected technology company with global social value all become genuinely achievable. Being within AI, while keeping our minds above AI. Being global in presence, while remaining rooted in China! Thank you all! Fei Pan CEO, BlueFocus April 2026 *Based on the real-time exchange rate quoted by Morningstar at 02:56 UTC on April 17, 2026, USD 1 = RMB 6.8216.