Compiled by Gu Lingyu, Tencent Technology
DeepSeek recently completed its first round of external financing since its founding. The total fundraising amount exceeded RMB 50 billion (approximately USD 7.4 billion), with a pre-money valuation of about RMB 367.5 billion (approximately USD 54.3 billion). Among the investors, DeepSeek’s founder Liang Wenhong personally invested RMB 20 billion, Tencent invested RMB 10 billion, CATL invested RMB 5 billion, and NetEase, JD.com, andIDG Capital each invested RMB 3 billion, while the National Artificial Intelligence Industry Investment Fund contributed RMB 1 billion.
Previously, Liang Wenfeng proposed the principle of "no fundraising, no IPO, no commercialization." This large-scale funding round marks DeepSeek's formal entry into the capital market and has drawn widespread industry attention to its commercialization path and technological vision.
At a recent investor engagement event, Liang Wenfeng provided a detailed overview of DeepSeek’s organizational culture, open-source philosophy, technology roadmap, and perspectives on the industry’s competitive landscape.
The following is a curated transcript of Liang Wenhong’s remarks from a nearly four-hour session, obtained by Tencent Technology, organized by theme and totaling 118 entries. The text preserves the original meaning with minimal editing.
01 Vision and Restraint
1. When we first started this company, we never imagined that we would eventually make a lot of money, go to the capital market, go public, or anything like that. The first dozen or so people never thought about it—if they had, they wouldn’t have joined.
2. We are doing this with great goodwill toward the world, believing it is beneficial to humanity—it’s about more than money. Our original motivation, our vision, and the vision we’ve maintained to this day were never driven by maximizing commercial interests.
3. Managing a large company isn't about your rules and regulations—it's about vision. Vision isn't a slogan on the wall; it's about what you do, not what you say—it's how you actually operate.
4. We are unstructured—we are vision-driven, organized solely by a shared vision. We don’t operate based on “I need to achieve certain KPIs” or any form of performance evaluation; we only have a vision.
5. This vision isn’t even written down—it has never been put into writing. It lives in the way we do things and our attitude toward the world.
6. We don’t have many other advantages—we don’t have special skills, we’re not richer than others, and our team isn’t better than those at other companies. Actually, none of that is true. When we founded this company two years ago, we had little money, no major connections, no brand recognition, and no influence—we were just a group of very ordinary people.
7. The more restrained you are, the more likely you are to succeed—or at least, this has been true so far and remains the most plausible explanation. Otherwise, there’s no way to account for how we’ve succeeded: we had no advantages, started from a very low point, had minimal resources, and our team was essentially just a random group of ordinary people.
The AI matter is too big, and the potential rewards are too great. We are being very restrained—just by succeeding, the final rewards will be enormous. Even if you take just a small portion, the returns will be massive, so there’s no need to consider which part to take or how to take it; I don’t think this even needs to be considered, because the potential is simply too large.
9. Last year during the Spring Festival, we experienced a sudden surge in users, but we did not pursue retaining these users, monetizing them, or seizing commercial opportunities by exploiting them. We did not compete for users or try to profit from them; instead, we worked hard to find ways to serve them well.
10. We don’t have the mindset of trying to become the next super app, competing with someone, or aiming to be the next ByteDance or Tencent—we have absolutely no such intentions. I believe the opportunities presented by AGI going forward will be enormous, and those AGI opportunities will always remain enormous.
11. Discipline is a strategy—it means sometimes you give up certain things to gain more in other areas. Similarly, not open-sourcing this can be seen as both a pressure we face and a concession we make.
12. I understand this restraint as something that, in the long term, increases our probability of achieving AGI. When considering this, I have no doubt that AGI will hold immense commercial value. On that basis, my priority is not how to secure a larger share or take more for myself; rather, my priority is how to increase the likelihood of successfully achieving it.
13. We have always been very restrained and unwilling to become competitors with any large or small internet company. I hope I can empower him, or assist everyone in doing this, and help everyone achieve this.
14. I feel that, by maintaining this attitude all along, we haven’t lost anything as a result—we haven’t received less because we open-sourced, because of our goodwill, or because I’ve helped others. In fact, it may have even worked to our advantage. This might seem counterintuitive, but it’s truly the case.
15. Our goal is AGI, but we have consistently pursued commercialization, which is why we have C-end users and B-end revenue. Historically, this strategy has been successful.
02 AGI Roadmap
16. If you can describe a problem clearly, providing it with complete context and instructions, it has already surpassed humans. But there is a definition, a prerequisite: you must provide it with complete context and complete instructions.
17. AI cannot replace your employees. But if AI had the ability to learn continuously—just like your employees, who learn at the company for two months—then it could replace everyone. So we’re still one step away: continuous learning.
18. The development of AI can be understood as a staircase. Last year’s step was chain-of-thought, as we discovered that using chain-of-thought methods enables intelligence to reach a higher level.
19. This year’s ladder is Agent, because we’ve found that with an Agent approach, even more tasks can be accomplished—it has a broader capability range and a higher ceiling for intelligence. Agents rely on CoT, and CoT in turn depends on the previous rung of the ladder, which is the language model—so no step has been wasted.
20. After the Agent, we believe the next issue to address is continuous learning—how to enable the model to learn continuously over time, rather than requiring a single, intensive training session; it should be able to learn persistently, much like a human.
21. After continuous learning, we may reach a singularity. At this singularity, once the model can learn continuously, it will be capable of doing everything a human can do. It will be able to develop its own versions, conduct further research, and create its next iteration—developing even more advanced artificial intelligence models.
22. This singularity is not actually a singularity; it is also a gradual process. This process may be a relatively long transition, not a sudden change. However, by habit, we tend to think of it as a singularity.
23. This is our speculation: we believe the timeline should be this—first solve learning to learn, then reach the intelligent singularity capable of self-iteration, and only then will we have embodied intelligence. After achieving embodied intelligence, it will enter the real world and be able to help with household chores and elder care.
24. If we first address continuous learning, then the singularity of self-iteration, and finally embodied intelligence, the journey becomes much easier, because later stages can leverage technologies developed in earlier ones.
25. We are focused solely on the mainline of AGI. The AI field is broad, and there are many areas we believe lie outside this mainline—for example, 3D and video generation, which I don’t think are closely related to the core of intelligence—we won’t pursue those.
26. When video generation first came out, it became very popular, as if it was something you absolutely had to do—if you didn’t, you weren’t considered an AI company. So I found it strange, because if you think about it carefully, it has nothing to do with the roadmap of intelligence.
27. It’s a good business commercially, a good business commercially. But it has nothing to do with intelligence. We won’t do it just because it’s a good business; we’ll only do it if it’s part of the intelligent roadmap.
28. From our perspective, world models and intelligence are not the most important things at this stage. The most important things are AI training and how to address continuous learning after AI training. This is our company’s assessment, though of course each company’s perspective may differ.
29. We currently place more faith in the narrative that AI can accelerate AI research. That is, it’s not linear, because you can use AI to accelerate your own research, so it may become nonlinear over time.
30. I believe embodied intelligence will inevitably enter and ultimately become essential, because for an ordinary person, their needs aren’t centered around computers, right? Normal people don’t need computers for eating, drinking, entertainment, clothing, food, housing, or transportation—they need embodied intelligence to address real-world human needs.
31. What do we want AGI to do? It can help us iterate to the next version of the model, just as it helps us iterate to the next version of the model. Once embodied, we also want it to iterate on the next version of embodiment—to build the next generation of robots.
32. The core capability of the next-generation model must include continuous learning; otherwise, it cannot be called a next-generation model. Until then, all we can do is reduce costs, improve performance, and increase speed. But for a major breakthrough, it must have continuous learning capabilities.
33. The current capabilities of the Agent are limited because it cannot learn continuously; it cannot effectively engage in continuous learning. If continuous learning could be achieved first, AI’s capabilities would be extremely powerful, greatly enhancing the efficiency of our own research.
34. If we first focus on building something and then continue learning, general intelligence might become much easier, and using it would also be straightforward. So I say this is a result we’d much prefer—it requires less effort and is more effortless. Otherwise, currently trying to manually create general intelligence is a tiring, arduous process that’s data-intensive and labor-intensive, with low cost-effectiveness.
03 Team and Talent
35. My previous experience has taught me that the vision of AGI is incredibly powerful. This advantage in talent isn’t about my people being smarter than others, but about how I organize them, motivate them, and foster collaboration.
36. Bringing together smart people doesn’t mean they will naturally collaborate or be passionately driven toward a common goal—you need a vision.
37. Our greatest core interest is maintaining team stability. This is our greatest core interest—perhaps even the only one. As long as I can maintain team stability, I will succeed; I will definitely achieve AGI. It’s that simple.
38. Money is certainly not an issue, resources are not an issue, and all other elements are easily obtainable. For us, there is only one core priority, one non-negotiable: we must maintain team stability.
39. This is also a very significant challenge we face, or perhaps the greatest risk. Of course, this risk has been significantly mitigated due to our recent financing round, as everyone has received a relatively large number of options with substantial value.
40. From the perspective of team stability, as long as the most important and longest-tenured employees remain, others are unlikely to leave. Even if they have fewer options or lower income, they won’t leave, because they’re not solely motivated by money—everyone wants to be in an environment where AGI can be achieved.
41. The rest are just matters of time; they might delay us by six months or a year at most, but we won’t be unable to do them. We definitely don’t lack money, and we definitely don’t lack resources—none of these are lacking.
42. The main gap we have with the U.S. is in resources; the gap in talent is not significant. There’s almost no difference in talent, because it’s the same group of people—many of whom are Chinese. When Chinese people go abroad, some stay in China, some stay overseas, and some move abroad—it’s not the case that only smarter people go overseas.
43. People are not the bottleneck; resources are the biggest bottleneck. Resources first affect talent development, because with less computing power, we have fewer opportunities to conduct experiments, so our overall talent is still behind that of the United States. The gap in talent is essentially due to the gap in computing power.
The shortage of AI talent is also temporary, and we have already seen it significantly alleviated. In reality, there is no real shortage of AI talent—every company will quickly train its own people, and training talent happens rapidly.
45. There are too many companies in China currently building models—way too many. In the U.S., there might only be three, but in China, there are far too many companies working on foundation models. In the end, not nearly so many people will be needed to build foundation models—it will definitely consolidate.
46. Our company’s management actually follows two paths: one from top to bottom, and one from bottom to top. The bottom-up path means everyone decides for themselves what they want to do and does it, with no supervision and no KPIs.
47. Generally, we hope that employees have half of their time unassigned, allowing them to do whatever they want. This is a space for research, where they can explore whatever they consider important, without any predefined requirements.
48. We also don’t usually work overtime. There are two reasons for this. First, research requires a relatively relaxed environment. If you push too hard, you can’t do proper research. Since research requires personal interest and the habit of thinking about these issues in your daily life, it’s only possible in a relaxed setting.
49. The second is that we are extremely focused. Being extremely focused means we do very few things. Therefore, I have less to do and don’t need to work overtime. This is consistent with the restraint mentioned earlier.
50. Our company as a whole is built on consensus; I don’t mean that I decide everything on my own, but rather that I seek consensus. My authority and influence within the company are based on consensus.
51. This decision-making mechanism is essentially a consensus-seeking process; I won’t move forward with anything unless there is consensus, and only then will I take action.
52. As the team grows, we will make this adjustment. We should have made this adjustment immediately, because I am already doing it. Without this adjustment, many things cannot move forward. Indeed, many departments should have an organizational structure.
04 Hashpower and Resources
53. How many cards do we need? Right now, more is definitely better. Within our budget, the more cards we can get, the better—there’s no doubt about it. So our current strategy is to buy as many cards as possible at a reasonable price.
Actually, it’s very difficult to spend such a large amount of money—there aren’t enough cards available to buy, they’re hard to acquire, and the prices are very high. We can’t simply spend exorbitant amounts; we must ensure the prices are reasonable. If we can spend $20 billion this year, it would be considered an outstanding achievement by our procurement team.
55. The biggest gap between us and the United States lies in resources. On the one hand, we simply cannot purchase the necessary hardware domestically; on the other hand, our capital investment is significantly lower than that of the United States. Our capital investment is much lower, and the proportion of talent salaries within this is very small. For example, they offer salaries in the hundreds of millions of dollars, but even so, talent compensation still makes up only a small portion—the bulk of the cost is computing power.
56. All the differences we observe—including differences in talent, model capabilities, and applications—can be attributed to differences in computing power resources.
57. The gap between us and the United States may be 12 months behind, possibly 12 to 18 months behind, or 6 to 12 months behind. In simple terms, we are about two years behind the United States, yet we achieved this using only one-twentieth of the computing power the U.S. has.
58. This narrative is lagging by one to two years, but it uses only one-twentieth of the computing power. In the future, we aim to rewrite this narrative: using just a fraction of its computing power, but drastically shortening the timeline to six months or even three months—I believe this is our goal.
59. Scaling — we believe in scaling; the larger the scale, the better the results, and the more features we can unlock. What prevents us from scaling is not a lack of willingness, but rather insufficient computing power.
60. We trained such a large model not because I believe this size is sufficient, but because I happened to have that much resource. I determined the model size based on the resources I could afford and train with—not because this model size is adequate.
61. When Silicon Valley talks about scaling reaching its limit, that’s from Silicon Valley’s perspective; for Chinese people, we are still far from that point—we haven’t even reached that level of scaling yet. This scaling includes data scaling, model size scaling, and training costs.
05 Domestic Chips and Ecosystem
62. The moat around NVIDIA's CUDA is rapidly eroding. On one hand, with the advent of AI, it’s now much easier to build an ecosystem, because AI can write code.
63. The market for computing cards has now surpassed that of gaming cards, so there’s no reason for these two to remain coupled. The current trend is that they will no longer be coupled in the future. As a result, dedicated chips—whether from Huawei or NVIDIA themselves—will be specialized chips, nothing like the previous ones.
64. There is now a historic opportunity for domestic AI chip alternatives. We believe that within the next year, we will see one thing verified: the ecosystem for domestic chips is completely viable. Previously, people thought there were issues—that they couldn’t be used or were difficult to use—but I believe that within the next year, we will be able to reverse this perception, or rather, facts will do so.
65. The hardware and ecosystem of domestic AI chips are not problematic; the only issue is insufficient production capacity. There are no barriers to adapting domestic cards, and NVIDIA cannot stop this. In a normal commercial environment, if I could purchase NVIDIA cards, domestic alternatives would be much harder to adopt; but when NVIDIA cards are unavailable, everyone is forced to turn to domestic chips.
During the training of V3, NVIDIA GPUs were still used, but the NVIDIA ecosystem was no longer relied upon. V3 uses NVIDIA hardware, but instead of depending on NVIDIA’s ecosystem, we first developed a high-level compiler called TileLang, and then built all other components on top of TileLang’s ecosystem, thereby minimizing dependence on NVIDIA’s ecosystem almost entirely.
67. I am quite optimistic about domestic computing power. I believe NVIDIA is digging its own grave on this front. Huawei’s super nodes, such as the Huawei 950 super node, can fully replace NVIDIA’s GB200 and GB300 in both performance and price.
68. Four Huawei cards can match one NVIDIA card.
69. I believe the gap between us and the U.S. in chips will no longer exist in terms of ecosystem, but in chips themselves, it’s four times plus two years.
70.We are currently primarily collaborating with Huawei. Huawei is handling its own adaptation, but we will actively participate in this ecosystem and engage deeply within Huawei’s ecosystem. Huawei’s main issue remains insufficient production capacity.
71. I don’t really believe that we’ll still be stuck on capacity issues five years from now. Right now, we’re definitely held back by capacity—this year, next year, and the year after, I think we’ll likely still be facing capacity constraints—but five years from now, I’m not so sure. I remain fairly optimistic.
06 Competitive Landscape and Industry Assessment
72. The final performance differences among various models should be viewed holistically. When comparing model performance, it’s essential to do so under the same cost conditions—only then is the comparison meaningful. Just as you would compare two cars only when they are in the same price range.
73. Anthropic surpassing OpenAI now—is this long-term? I don’t think it is; it’s definitely temporary. OpenAI and Google will likely continue to alternate in leadership going forward.
74. When it comes to dominating the global AI division of labor, Chinese companies are likely to still play the role of the largest producer. Logically, we have the largest production capacity, including chips, and we also have the most electricity.
Chinese people will make this product the cheapest possible, and then focus on performance—after all, for many goods today, there’s little difference between products made in China and those made in the U.S. In the future, AI may be the same, but Chinese-made AI could be even cheaper. This lower price may stem from systemic cost advantages, similar to how services provided by China are often cheaper across other industries.
76. The final gap should be threefold: cost, time, and user experience. Beyond that, there may be no significant difference.
77. Cost is definitely a differentiator; I think cost is probably the primary distinction. The second is time—when you can achieve it. Whether you do it a few months earlier or later makes a difference.
78. OpenAI initially thought it could truly monopolize the world, but in reality, it will face many, many challengers. It will encounter competition and won’t be able to operate so easily. The United States will face challenges, and in the future, it may also face competition from China, because Chinese companies are willing to provide this service for less.
79. Those who aim to take more will be defeated by those who aim to take less. Even if you don’t actually take more, if your vision is to take more, you’ll be defeated by those whose vision is to take less. In reality, no one has actually made money—it’s just a vision. If your vision is to take more, you’ve already lost, and you’ll face greater challenges.
80. For us, it’s not about maximizing profits or setting prices to achieve the highest possible return—we simply aim for a reasonable profit. That’s the explanation. I truly believe this, and I’m not trying to justify it, because there’s no need to.
81. I think in many aspects of user experience, we may be able to outperform the U.S. In terms of product capabilities, we are not necessarily inferior to the U.S., and our costs should also be lower, so China will still remain competitive.
82. The cost is easy to understand—it’s because they don’t do it at all, so they never develop this capability. They certainly don’t prioritize this matter as much as we do. We can treat it as something extremely important, but for them, it’s not important.
83. A large model might not need to compare two big companies and two small companies; comparing just a few might already be sufficient. The difference comes down to only two things: time and cost. So it’s unlikely that any one company will make excessive profits—I don’t think excessive profits are possible. Those who manage costs well earn a bit more; those who manage them poorly earn a bit less—that’s all there is to it.
07 Model Development and Technology
84. Half of the people in our company might generally think OpenAI is better. In fact, Anthropic has a first-mover advantage, but this advantage should disappear quickly and isn’t something it can sustain long-term. All three companies are very strong, and among them, Anthropic has the highest efficiency—it spends the least amount of money and burns through the least capital.
85. Multimodal layout is something we've always been working on. It's important for products, and especially important for consumer-facing products. But when it comes to the upper limit of intelligence, it's just one component—not the core itself.
86. We will be introducing related models; our V4 and subsequent versions will support native multimodality. However, for us, multimodality is a component of intelligence, not intelligence itself.
87. I see no upper limit to the scaling of language models. Our current level of intelligence, or even the level of intelligence in the United States, shows no signs of an upper limit.
88. Many of us internally think like this: First, it must be useful to us; it must be for our own use first. And this is the fastest way to achieve AGI. When it works well for us, it likely means others will find it useful too—but first and foremost, we must ensure it works well for us.
89. The primary goal of the model we built isn’t for everyone to find it easy and useful, but for us to find it easy and useful ourselves. First and foremost, it must be useful to us. Once it’s useful to us, I’ll be able to develop the next version of the model much faster.
90. We call this “drawing lots.” The barrier to entry is low—anyone can participate—but I don’t know whether what you get depends on talent or something else. So we don’t need to allocate resources here. What sets us apart from other companies is that we take the time to discuss this issue, think about it, and treat it as something important.
08 Commercialization and Pricing
91. Our API pricing represents a reasonable profit—approximately enough to recover our costs within ten months after purchasing a batch of equipment on the market. I believe this is a fair profit.
92. If profit maximization is the goal, prices should be set higher. Within this price range, user demand is inelastic—meaning even if I raise the price by 50% or double it, the consumption of tokens changes very little.
93. With one of our models, we initially worried about excessive demand, so we set a high price, which made the team unhappy. Later, I lowered the price to a quarter of the original, and everyone became very happy.
The upper limit of B2B business should still be demand. Under the current generation of AGI and AI technologies, B2B demand should be limited. It will grow rapidly, but it is not an infinite phenomenon—ultimately, it is constrained by demand, not computing power.
I now feel it’s achievable to do both. If I can generate hundreds of millions of dollars in B2B revenue this year, combined with our C-end user base, that alone would establish a solid business foundation. If next year our B2B revenue grows further, the company could be very close to profitability—or even already profitable.
In the worst-case scenario, selling our API alone could sustain a publicly traded company. Even if there are no further technological advancements and our technology remains frozen as it is, we could fully focus on selling our API and delivering these services well—I believe that would be sufficient.
97. Looking at the current situation, I believe the most reasonable approach is to fully focus on building a general-purpose Agent, with lower priority given to other Agents, such as those for finance or healthcare. We should prioritize Coding first, because a Coding Agent can accomplish a lot, and there are many vertical Agents still to be developed. At this stage, we believe the most important focus should still be the Coding Agent.
98. I believe low cost is first and foremost a result. Our model has consistently evolved toward a lower-cost architecture, which aligns with our vision. We still have many algorithmic approaches that can further reduce costs.
Another reason costs are decreasing is that lower costs enable me to train larger models and afford more substantial models. With limited computing power, higher computational efficiency allows me to support larger models.
09 Open Source Strategy
100. I think we will open-source our code, and even our strongest model will likely be open-sourced, because I don’t see any clear advantages to keeping it closed. ByteDance’s models are closed-source—what benefits does that actually bring? I don’t see any.
101. Even if the model is open-sourced and you tell people everything, the barrier to entry is still very high. It’s also very difficult for others to actually use it. To use it effectively is hard; furthermore, making the cost of usage very low is extremely difficult— it’s not that easy.
102. Open source does not affect revenue. Open source, in my view, has no impact on our business model.
103. I’m not worried at all about others deploying our models and competing with us. In fact, we hope they do. We strive to support the open-source community as much as possible to help everyone deploy our models successfully.
104. When interacting with external parties, our stance is: we focus solely on the core AGI pathway. We are more than willing to assist and support anyone—even our competitors, including Alibaba, Zhipu, and Moonshot AI—to improve their work, because we gain nothing by withholding support, and we are open-source to begin with.
105. Is the open-source model we provide the same as the one we deploy ourselves? Yes, they are identical. We do not release an inferior model as open-source while using a better one for our own deployment—that’s not our practice. They are the same.
10 Data and Post-Training
106. The data should be almost half the size of the model. There’s also an issue with data labeling upfront. Regarding data labeling, this is tied to our capital investment structure; with our current capital allocation, we cannot afford the high costs of producing that much high-quality labeled data.
107. The cost of data labeling in the United States is not significantly different from that in China. There is no cost advantage in labeling data in China, especially for high-end data, making it difficult for us to invest in data labeling on the scale seen in the United States. This path is very challenging in China because data labeling is simply too expensive—whether we outsource it or do it in-house, it’s a significant burden.
108. Right now, we’re essentially taking a two-pronged approach. It’s not that we can’t label data at all, but some labeling has low cost while others are more expensive, so we start with the low-cost ones.
109. You could also say that right now, half of our company is engaged in labeling data—half of our core researchers, the most important people, are focused on labeling data. We’re concentrating entirely on data labeling. At this stage, solving AI’s challenges comes down to data labeling.
110. The bottleneck for high-quality data annotation, in my view, is time—it takes time. Because for OpenAI, for foreign companies, and for Anthropic, they started earlier and have more capital and more GPUs.
111. The hallucination issue in large models significantly affects user experience. There is a method to address hallucinations, but it is a complex topic. Hallucinations can be viewed as a solvable and improvable issue through better post-training.
11 Organization and Company Positioning
112. First of all, we have no model to imitate. Every step we take is based on reality, grounded in facts, and guided by actual conditions to determine the right course of action. Therefore, it is a product of its time—or a response to real-world circumstances—not the result of imitation.
113. We are clearly aiming for commercialization. In the end, we must be able to survive—we are a company, and the government will not give us a single penny.
114. We are fundamentally still a company, but we make deliberate choices about what to earn, when to earn, how much to earn, and how to earn it. Many companies achieve greatness because they pursue something beyond profit—and that pursuit doesn’t hinder their commercial success; instead, it enhances it.
115. For our partners, this funding round was carefully selected. First, I believe their interests are most aligned with ours—those who have the least hostility toward us, or who most genuinely want us to succeed. Not everyone wants us to succeed, because we have indeed impacted the interests of many others.
116.AI doesn't lack taste or intuition; what it lacks is the ability to learn continuously. AI's taste and intuition are not an issue. Have it write an article, and its taste and intuition, in my view, are perfectly fine.
117. We want to focus on just one thing. I think AI is a huge field, and I don’t need to do everything—I only want to do one piece. If we concentrate, I believe the business opportunities here are already large enough. If the AI era gives rise to many companies worth trillions, I believe we will be one of them.
118. We hope to support more people, but we simply don’t have that much capacity. We have the intention, and there is no conflict of interest—but whether or not we act on it is another matter. At the very least, there is no conflict of interest here; we aim for mutual success and collaboration.