Section Insights
Market Dynamics and Chip Diversity
What are the implications of Nvidia's recent news for the chip market?
The recent developments in the chip market, particularly with Nvidia's dominance, highlight a growing diversity of chips tailored for various AI tasks. This shift is driven by the increasing adoption of AI technologies and the emergence of alternative chips that cater to specific needs.
- The demand for chips is rising due to AI adoption.
- Diverse chips are entering the market, offering specialized capabilities.
- Market dynamics indicate a potential decrease in Nvidia's market share over time.
Future of Chip Market Share
Will Nvidia's market share decrease in the future?
Yes, while Nvidia currently holds a dominant market share in AI accelerators, the emergence of specialized chip companies suggests that this share will likely decrease as the market diversifies.
- Nvidia's dominance is currently significant but may not last.
- New chip companies are emerging with innovative technologies.
- The market for chips is expected to become more competitive.
Adoption of New Chip Technologies
How are new chip companies overcoming Nvidia's software lock-in?
New chip companies are rapidly developing their software to facilitate market adoption, aided by companies like Callosum that help integrate these technologies into existing systems, thus reducing the impact of Nvidia's CUDA lock-in.
- New chip companies are innovating quickly to compete.
- Callosum plays a crucial role in integrating diverse chip technologies.
- The software landscape is evolving to support multiple chip architectures.
Sovereign Capabilities and Chip Diversification
How are countries responding to the need for chip diversification?
Countries are increasingly focused on developing their own chip technologies to ensure sovereignty and leverage in the tech landscape. This shift is seen as a response to the need for independence from dominant players like Nvidia.
- Nations are prioritizing the development of their own chip technologies.
- Sovereign capabilities are becoming a key focus for countries.
- The tech landscape is moving towards a multipolar world.
Leadership Lessons from Nvidia's Journey
What leadership lessons can be learned from Jensen Huang's experience?
Jensen Huang's leadership at Nvidia exemplifies the importance of staying true to core values and maintaining consistency while navigating the challenges of a rapidly evolving tech landscape.
- Leadership requires adherence to core values.
- Consistency is crucial in maintaining a company's direction.
- The UK has potential for growth in alternative chip technologies.
Transcript
0:00 This is a momentous moment for your ecosystem. How are you digesting the news from Nvidia as you try to build a product, and an alternative that allows at least those dependent on chips to have alternate systems, and to be able to rely not just on one chip, but multiple chips. Yeah, great. So we've obviously had quite extraordinary news over the last kind of, you know, a few days, of course. And I think everyone's digesting the implications of this. I think it is important to take a step back and think about the fundamental market dynamics over the longer time period. It helps us understand what's really happening on the macro scale. What we're realizing is, is that there has been a huge adoption of AI technologies.
0:37 Of course, many people are understanding the different types of workflows and use cases for AI in new ways. and as a result, of course, there is a huge demand for chips in general, of course. And Nvidia have been building amazing chips and they've been doing so for a very, very long time. And what people are now realizing and we are realizing at Callosum inherently is, is that, actually, because there are so many different tasks and many models in the ecosystem.
1:00 There's actually now and also an increase in the number of alternative chips that are entering the market as well. And the reason for this is because those chips are good for different things. Some of them are very energy efficient. Some of them are very, very fast. And as a result, the ability to be able to map the different tasks in the AI systems of these different alternative chips is incredibly exciting. So whilst, of course there's a lot of short term dynamics that are happening right now in the market volatilities, this is this is quite normal.
1:23 Of course, is is very unique news that this happened. And really there's a macro picture here which is there are kind of a wide range of really amazing chips that are entering the market in which the market. I see the pie is growing for, for full compute in general. But that pie may be shared increasing over time, but it's very exciting and it's very simplistically what you offer is the ability to kind of hop seamlessly across different chip architectures and different chipsets for users and for those who want access, to compute. Do you think longer term, right now, Nvidia has about a 90% market share of AI accelerators.
1:52 Longer term, the picture is that market share shrinks. Is that your view? That's correct. I say I'd say that over, over the last kind of, period of time of AI in the last. In the most recent chapter of AI really we have had an Nvidia dominance. That's really because Nvidia has been, I've been building amazing chips. It's been the predominant one in our supply chains has been available to us. but now we have a growth of really amazing chip companies, for example, within Europe.
2:14 We've just had the other day, Fractile raising, $720 million in their Series C. We have companies like Olix in London building amazing photonic chips, raising hundreds of millions of dollars. And so what people are realizing is that there are specialized chips that are very good for particular tasks. And what does this mean? It means naturally, over time, the market can become increasingly shared. And that pie, as I say, can be shared across those chips. And it calls really for a type of company, like Callosum, which, as you describe, enables the software systems that allows us to break down AI systems into their parts and to run the best tasks on the best chips. And I think this is really, really good for everyone. It means that, people can have cheaper systems. They can have much more performance systems. it's great for the planet because it's more energy efficient. But ultimately, you're right.
3:02 In the market dynamics, we see that this will gradually shift to a much more shared picture on the chip space. Okay, so there's going to be more chips on the market, whether it's OpenAI's Jalapeño or what's coming out from Olix and Fractile AI. More chips, more diversity. The market share for Nvidia will shrink, but the demand for chips continues. What do you say to those who say, look, Nvidia has got this lock in with CUDA and an Olix or a Fractile or a Jalapeño. It's not gonna be able to compete because of that lock in that CUDA has. So the CUDA, CUDA mode so that so to be clear, with the software that really allowed an ecosystem to to begin, I guess at the at the kind of dawn of this kind of modern AI era was, it really helped developers be able to adopt Nvidia very quickly.
3:40 We've actually seen an acceleration of many of the the software life cycles have really accelerated now. So it's much quicker than ever before to be able to accelerate how you build your software. And so chip companies are obviously realizing this. They're realizing they can actually build their software very quickly. That enables that their chips to be adopted in the markets. And also you have, companies, in a company like Callosum, for example, which is also accelerating that adoption.
4:04 We work with many of these new chip companies, building great technologies to accelerate how their chips can be plugged into our systems, for example, which also gets them to market. And so all of these speak ultimately, to the fact that those those CUDA moats that we had before, there are other moats that exist. but that one seems to be declining. quite okay. Okay. The future is more chips and the ability to kind of hedge, but the here and now is that is that Nvidia remains and continues to have that dominance. What do you hear in terms of and your conversation. You're backed by the Sovereign AI Wealth Fund. What do you hear in your conversations about sovereign, sovereign capabilities and how dependent we are?
4:40 I was speaking to the UK AI minister. He's really focused on the chip space and the semiconductor space. Is this moment another moment that kind of catalyzes a view to kind of diversify? I'd say it's hard to say this exact moment because it's obviously quite stable. We're learning quite a lot. but what we are realizing, is this really kind of two things. We're moving in many ways to a more multipolar world. every company and country are thinking about how do they have leverage on their own futures?
5:08 How can they dictate the technology of their futures? And at the same time, really, the future is yet to be built. there are many, amazing new technologies. I think that that founders and companies are excited to be building. And so really, what does that mean? It means that everyone should really be concerned about what are the most ambitious, important technologies that you can be building today. And I see that this I guess one of the signals of this is that, countries are accelerating this.
5:31 So in the UK, as you described, Callosum has been backed by the UK Sovereign AI Fund. for this this is one of the reasons for this is because, of, of the leverage that, that we get that technologies can afford or nation states and being able to interoperate between different chip technologies. and so I think this is really a signal, that countries, are accelerating their sovereign efforts and I, I'd like to stress that really that is born from building the very, very best technologies that you can have to leverage on your own future.
5:58 Can I just briefly ask you before we end? Have you personally, as a founder, taken any lessons from the Jensen Huang saga in terms of leadership? Yeah, I have I think I think effectively, Jensen, if you kind of before this moment, Jensen is in an unbelievable job over the last 30 years in building Nvidia from a founder, to now, I think the way that it's very kind of hard to say exactly about what's been happening here is a very unique, example in any new case, the way that, I, the lessons that I guess that I take from this is obviously that when, when one is kind of, at the forefront of your, your power to, to to effectively, you know, make sure that you kind of stick to your values and then make sure you're consistent, okay.
6:37 Core to values. Make sure you're consistent. And you think the UK is well placed more generally in terms of building out alternative. Chip. Absolutely, absolutely. So we've as I've described already, the UK has a very interesting position in the world and in being able to interact globally with all of our partners. It's an exciting time. With all of that new chip technologies happening, there's a huge amount of investment coming to UK. We have unbelievable talent and super excited to be building.
Summary
- Nvidia currently holds about 90% market share in AI accelerators, but this is expected to shrink as new chip companies emerge.
- The rise of alternative chips is driven by diverse use cases in AI, with some chips being more energy-efficient or faster.
- Companies like Callosum aim to facilitate seamless integration across different chip architectures, enhancing performance and cost-effectiveness.
- The software ecosystem is evolving rapidly, allowing for quicker adoption of new chip technologies despite Nvidia's established CUDA framework.
- There is a growing recognition of the need for countries and companies to diversify their chip dependencies in a multipolar world.
- The UK is positioning itself as a significant player in the chip market, backed by initiatives like the Sovereign AI Fund.
- Founders are encouraged to focus on building ambitious technologies to secure their future leverage in the tech landscape.
- Leadership lessons from Nvidia's Jensen Huang emphasize the importance of consistency and adherence to core values in navigating industry challenges.
Questions Answered
What are the implications of Nvidia's recent news for the chip market?
The recent developments in the chip market, particularly with Nvidia's dominance, highlight a growing diversity of chips tailored for various AI tasks. This shift is driven by the increasing adoption of AI technologies and the emergence of alternative chips that cater to specific needs.
Will Nvidia's market share decrease in the future?
Yes, while Nvidia currently holds a dominant market share in AI accelerators, the emergence of specialized chip companies suggests that this share will likely decrease as the market diversifies.
How are new chip companies overcoming Nvidia's software lock-in?
New chip companies are rapidly developing their software to facilitate market adoption, aided by companies like Callosum that help integrate these technologies into existing systems, thus reducing the impact of Nvidia's CUDA lock-in.
How are countries responding to the need for chip diversification?
Countries are increasingly focused on developing their own chip technologies to ensure sovereignty and leverage in the tech landscape. This shift is seen as a response to the need for independence from dominant players like Nvidia.
What leadership lessons can be learned from Jensen Huang's experience?
Jensen Huang's leadership at Nvidia exemplifies the importance of staying true to core values and maintaining consistency while navigating the challenges of a rapidly evolving tech landscape.