68 – “Winning the AI Race”… America’s AI Action Plan

America’s AI action plan … “Winning the AI race” has just been announced. What is it all about? What are the implications? How will the rest of the world react? A deep dive into the announcement, approaches by EU and China, and overall implications of these action plans.

Navigation:

1. Intro (01:34)
2. Context of the White House AI Summit
3. Pillar I – Accelerating AI Innovation
4. Pillar II – Building American AI Infrastructure
5. Pillar III – Leading in International AI Diplomacy & Security
6. Comparing Approaches – U.S. Action Plan vs. EU AI Act vs. China’s Strategy
7. Implications and Synthesis
8. Conclusion

Our co-hosts:

• Bertrand Schmitt, Entrepreneur in Residence at Red River West, co-founder of App Annie / Data.ai, business angel, advisor to startups and VC funds, @bschmitt
• Nuno Goncalves Pedro, Investor, Managing Partner, Founder at Chamaeleon, @ngpedro

Our show: Tech DECIPHERED brings you the Entrepreneur and Investor views on Big Tech, VC and Start-up news, opinion pieces and research. We decipher their meaning, and add inside knowledge and context. Being nerds, we also discuss the latest gadgets and pop culture news

Subscribe To Our Podcast

Nuno G. Pedro

Welcome to episode 68 of Tech Deciphered. This episode will focus on America’s AI action plan, winning the AI race, which has just been announced a couple of weeks in by President Trump in the White House. Today, we’ll be discussing the pillars of this plan, from pillar I, the acceleration of AI innovation, to pillar II, building of American AI infrastructure, to pillar III, leading in international AI diplomacy and security.

We’ll also further contextualise it, as well as compare the approaches between the US Action plan, and what we see from the EU and China strategy at this point in time. We’ll finalise with implications and synthesis. Bertrand, is this a watershed moment for the industry? Is this the moment we were all waiting for in terms of clarity for AI in the US?

Bertrand Schmitt

Yeah, that’s a great question. I must say I’m quite excited. I’m not sure I can remember anything like it since basically John F. Kennedy announcing the race to go to the moon in the early ’60s. It feels, as you say, a watershed moment because suddenly you can see that there is a grand vision, a grand plan, that AI is not just important, but critical to the future success of America. It looks like the White House is putting all the ducks in order in order to make it happen. There is, like in the ’60s with JFK, a realisation that there is an adversary, there is a competitor, and you want to beat them to that race. Except this time it’s not Russia, it’s China. A lot of similarities, I would say.

Nuno G. Pedro

Yeah. It seems relatively comprehensive. Obviously, we’ll deep dive into it today across a variety of elements like regulation, investments, view in relation to exports and imports and the rest of the world. So, relatively comprehensive from what we can see. Obviously, we don’t know all the details. We know from the announcement that the plan has identified 90 federal policy actions across the three pillars. Obviously, we’ll see how these come into practice over the next few months, few years.

To your point, it is a defining moment. It feels a little bit like the space race of ’60s, et cetera. It’s probably warranted. We know that, obviously, AI platforms, AI services and products are changing the world as we speak. It’s pretty important to figure out what is the US response to it.

Also interesting to know that we normally don’t talk about the US too much in terms of industrial policy. The US seems to have a private sector that, in and of itself, actually stands up to the game, and in particular in tech and high-tech, normally fulfils or fills the gaps that are introduced by big generational shifts in terms of technology. But in this case, there seems to be an industrial policy. This seems to set the stage for that industrial policy and how it moves forward, which, as you said, we haven’t seen in quite a long time.

Bertrand Schmitt

Yes. At the same time, on one side, yes, there is some level of industrial policy, but I feel quite a big part is getting the government out of the way so that private companies can truly innovate. Because America, like many other countries, have accumulated over decades regulations, you could call it over-regulations of many sectors and industries. I would say part of it is showing the way, but part of it is really a lot about just removing obstacles along the way that were posed by decades of government regulations and acknowledging that if these regulations are not removed, it could truly impede America’s chances to win the AI race.

Nuno G. Pedro

How did we get here? Obviously, there was an executive order in January this year, 2025, on removing barriers to American leadership in AI, the directive being the US to become the world capital of AI, so to speak. We had this announcement at the White House in late July 2025, the White House AI summit, which set the stage for this, winning the AI race, America’s AI action plan piece. That’s how we got here in the first place. Shall we jump into the pillars and get into the meat of this?

Bertrand Schmitt

Yes. Maybe to finish on setting the context, what was interesting with this announcement, it was not done the usual way. It was more done in the context and format of a podcast co-hosted by Olin Podcast. Some of our listeners may know that David Sachs, who used to be part of the Olin podcast, is now the AI and crypto Czar of the White House.

Nuno G. Pedro

Yes. It was done with much pomp and circumstance in the good old days of empire-making in some ways. Let’s jump into the first pillar, accelerating AI innovation. It feels like this pillar, there are a couple of interesting elements to it, but one of the key elements to it, I’d say the overarching element that we can see is removing onerous regulations that would hinder the development of AI. So, having a very light-touch regulatory approach and proactive support for AI development all across the board in terms of R&D, also across the elements of data that are put at the table. The plan apparently does call for the identification of rules to eliminate and even pre-empt state AI laws.

Where the federal government would, in some ways, try to make sure that the state AI laws don’t come in and destroy a lot of these elements. We’ve discussed some of the stuff in the past; if our listeners remember, we talked about this, several onerous laws that were put forward in particular in California. It feels that the federal government is, in this case, sort of saying, actually we have to go in and be top down about this, and we want to remove onerous regulations across the board and make sure that they’re not put in place then by states.

Bertrand Schmitt

I’m very excited by all of this. Removing red tape and onerous regulations, I think, is critical to move forward, especially at such an early stage where we don’t really know where it’s going to go. As you said, California was preparing some very onerous regulations. Luckily, they were not signed into law by the governor, Gavin Newsom. But I think that was a very clear signal that states might decide to go crazy. We need to do something about it.

Personally, I think it’s good to have the states being able to define some of their own regulation. But I would make an exception in the digital domain where things are pretty different across much more easily the borders. There is no border to the Internet in some ways. I think when it goes digital, it’s a much bigger issue, especially at such an early stage. If you look at startup companies, they cannot deal with a plethora of state regulations depending on which state the user is in. This would be a nightmare, to be frank. Not just regulations, but the multitude of regulations state by state. Personally, I’m pretty excited about that.

The other piece of the puzzle is, of course, repealing the previous administration’s AI regulations that are just way too far-fetched, unnecessary, just going to hinder innovation. I’m very excited that this was removed, and I also believe they were a threat to free speech.

Nuno G. Pedro

There are elements here just to be a little bit even-handed, that we need to see how they play in action. One of them is this notion that this should not be acting as a Ministry of AI Truth, and that basically, there shouldn’t be ideological bias, so to speak.

There’s been some mentions in general about preventing so-called woke AI, et cetera, et cetera, in federal use. It could go the other way as well. We know that a lot of the data out there has biases because of the way it was built. I think that part on the data side and the ideological bias side, interesting thoughts, put at the table in the plan. We’ll need to see how it pans out. I still have at least a couple of concerns on that. The second element that I feel we have to be moderately optimistic about, we’ll need to see how it pans out as well, is almost this notion of, if we eliminate almost all the key onerous regulations out there, that in some ways, self-regulation and market forces will compensate for us. We also know that that’s not been historically true in many cases. We’ll also need to see how that pans out.

Obviously, just overregulating upfront, we’ve discussed it also in past episodes, like for example, the EU has done in the past, regulating upfront technologies without even understanding what the use cases are, is also not fruitful. It’s also not good. The US has had a long history of being lighter on regulatory environment, which has served it well in terms of technology advancements.

In this case, we’ll need to see how the regulatory environment actually evolves. I’m also moderately optimistic on that topic, but we’ll see how it actually evolves. It could be a big issue if we start seeing some elements of exaggeration in terms of how data is collected, how it’s put through to users, et cetera, et cetera. That would, in principle, require, at some point in time, regulatory intervention.

Bertrand Schmitt

I think what’s interesting in terms of AI not acting as a Ministry of Truth, first, I agree with that. I don’t think it’s good. What’s interesting, they’re not forcing anyone; they’re just saying, “Hey, if us, the federal government, are going to buy some AI products, your AI product needs to be free of biases.” They are not forcing companies to change their ways, but they are saying, “Hey, if you want to sell us products, you have to sell us neutral products. As us, the federal government, you, as a client, you basically do what you want.”

Nuno G. Pedro

I see a lot of companies right now and a lot of entrepreneurs saying, “Hey, this is potentially an opportunity for me to start a startup because…” Good luck with that. Good luck with understanding if data and algorithms are biased or not. How do you prove that they’re neutral? In principle, the concept is absolutely sound. We’ll see how it gets executed. Worst case, as I said, maybe there will be a bunch of startups actually in that space. We’ll see how that pans out.

Bertrand Schmitt

I certainly prefer startups in that direction versus startups focus on censorship.

Nuno G. Pedro

That’s a good way to look at it. Obviously, there’s been support across the board. You’d expect people like Jensen Huang, the CEO of Nvidia, to be supportive of it. Many other CEOs have been supportive of that. No shock there. I feel preaching to the choir would be the right expression to categorise a lot of the excitement that we’ve seen from CEOs across the board.

There is, in my perspective, an element of openness about the regulatory environment that would, in principle, be helpful for startups. We’ve seen the past, a couple of elements. We’ve discussed them as well in previous episodes, for example, in Healthcare, where there’s been like heavy-duty regulatory capture that really only benefits one player or a couple of players on top. So, the fact that the regulatory environment is lighter should, in principle, be beneficial for the creation of innovation and startups in the space, obviously, the larger players have huge advantages of scale, as we’ll discuss later in particular in pillar II, this notion of building of infrastructure at scale, they will have huge advantages as well.

But from my perspective as an investor, as a VC, I would say that this would in principle be positive, that there is less regulatory capture by nature because we’re saying is going to be a lighter touch, regulatory intervention that would be beneficial for the creation of startups and innovation at scale. We’ll see how it actually pans out. But should, in principle, be a positive movement.

Bertrand Schmitt

I agree with you. I think economies of scale from larger players, that’s a game. We are not going to hinder them. But what I felt was I was pretty worried around a year ago because it was clear that some actors were trying to do regulatory capture. There was a lot of talk about OpenAI behind the scene, trying to push a lot of their agenda through regulatory capture and added regulation. There was even a talk that open source AI, open weight AI will be forbidden, “for safety.” I’m actually very encouraged to see the direction it’s going.

There is actually a very clear encouragement on open source and open weight AI in the AI action plan. That gets me very excited that there will be less regulatory capture. I see also there is a lot around very practical stuff in that action plan, around trying to enable and push for more AI adoption. Organise industries so that they are more focused on adopting at scale, simplifying standards of adoption. There is also stuff around training, workers that might be impacted by AI or that might be made redundant by AI.

There is, of course, manufacturing. There is even talk about AI-enabled science, where there is a focus on building new data sets at scale that might benefit science. First, to be clear, it’s not just about LLMs, it’s really AI at large. It’s already bigger plans than just controlling chatbot, and what they say, it’s really all about putting in place the right infrastructure, scale, environment to push for AI in one direction.

Nuno G. Pedro

Yes, this is maybe a good time to move to precisely pillar II, which is the building of American AI infrastructure. Very central to the whole discussion. If I’ve read the documentation correctly, and if I’ve thought through the announcements correctly, there seems to be this notion of, in some ways, data centres, compute, power, electricity, energy, are strategic national assets. Because there are strategic national assets, they need to be treated as such. Obviously, their strategic national assets would be highways, telecom networks, et cetera. But there seems to be an emphasis in public-private efforts to expand that capacity relatively quickly.

The plan calls a lot for expedited construction of data centres and semiconductor fabs. Modernising the permitting processes and really even proposing potentially having nationwide permits to bypass state-by-state approvals, which we know is a flex by the federal government, but at the same time could be useful depending on how they’re actually deployed and put at the table. A lot of discussions happen around power and power grids and all the other elements, but very central this notion of we need to build data centres and semiconductor fabs as quickly as we can, and that should be our focus.

Bertrand Schmitt

We need the energy infrastructure to support it. So, power transmission, power generation, I think it’s well-thought-out. We talk mostly, I would say, on the software side in that first section, but now it’s more hardware. I’m glad it’s not just about chips. It’s really building the data centres that manage supercomputers at scale. It’s providing them energy. Let’s not forget that in the US, like in many other countries except China or India, power generation has plateaued, grown very, very slowly. What we are talking about, all these new data centres that are power hungry, it’s a dramatic shift in the power generation and transmission landscape because suddenly we need way more. We don’t want it in 10, 20 years; we want it tomorrow.

We cannot just keep doing the old ways, where this stuff was moving on a snail’s pace. Can’t say everybody was okay with it, but it was not so dramatically negative to the national interest. But here things need to move fast. We need this approval to set up factories to set up infrastructure. I’m glad there is a lot of expedited permitting, streamlining, removing of regulations to move fast.

One thing to keep in mind, the federal government is the biggest owner of land in the US, so they have some abilities to actually use this land, sell this land, do different things with this land without any specific state approval, for instance, because it’s federal land. There are a lot of things that could be done and that they seem intent to do, actually.

Nuno G. Pedro

There has been some debate in the past, for example, about as we’re onshoring again, for example, semiconductor fabs, if the US workers could be competitive, et cetera. If it’s anything to go by, the early indications from the TSMC factory in Arizona and Phoenix, extremely positive. We know it’s not the ultimate cutting edge that’s there currently, but if we measure yield, and just for those who don’t know what yield is, production yield in semiconductors is the percentage of functional chips produced from a wafer compared to the total number of potential chips on that wafer. Basically it’s an efficiency metric that measures how efficient we are at getting the most chips out of a wafer. So it’s normally a good metric to analyse efficiency.

We had had some public details already, and I’ve also had some conversations in private that have told us that, actually, we’ve seen yields of the Arizona factory that are in many cases superior to Taiwan factories. It seems like there is no big issue in terms of how these things are being deployed in production in the US. Actually, quite the opposite seems to be working well.

Again, the case for onshoring and bringing a lot of semiconductor production back into the US, which now mostly resides around Asia, has been quite positive. A lot of these elements I feel are extremely positive towards this pillar, that just this push that we’re seeing is a push that will yield, pun intended, great results over time.

Bertrand Schmitt

That’s great news, Nuno, thanks for sharing. It’s quite exciting to see what is being put in place and the execution of some of these plans. From what I’ve heard, there is a very big push on all this new manufacturing put in place in the US to be as frontier as possible in terms of technology. Use AI everywhere you can, use the latest production approach, use the latest tools and technologies to make it happen.

That’s very exciting because you don’t want low-end manufacturing coming to the US. That doesn’t make much sense. But the more advanced technological, the more the US worker will be able to really be a positive element of the equation. Talking about workers, there is a lot of training that is talked about in this document, in the action plan. It’s pretty good to see that they think about it at every level. Training for electricians, for HVAC technicians. There is a willingness to make sure the workforce is properly trained to be able to deliver on these investments. That means that it will be a lot of really good, solid opportunities for American workers to work on all of this.

If you remember, a lot of the talk in the past about AI was how workers are going to be impacted negatively by AI and jobs will be replaced and stuff. But actually it’s pretty clear that there are a lot of jobs that are going to be created in the pace of moving America to new AI industries. That’s very exciting from that perspective.

Nuno G. Pedro

Indeed. Obviously, on the energy part, it would be a huge omission from us if we didn’t talk about renewable energies and their role. As we’ve seen the past, President Trump and the administration seems to be doubling down on fossil energy. I mean, there isn’t like a specific element of saying that we’re not doing renewable at all. There is a revitalization of nuclear plants, which we know has been pushed by some of the large tech players, and we’ve mentioned it in previous episodes as well. There’s a little bit of an element here of big oil comes back into the table. We’ve seen, obviously, a lot of praises from people like the Chevron CEO.

The level of energy that needs to be produced. The question is, can we do it just on renewable with the current infrastructure that we have, and it seems like the answer has been historically no, and therefore it’s probably not an either/or at this moment in time. It’s more of a, “We need to throw everything and the kitchen sink at it and see how we can build energy production at scale.” As you also mentioned earlier, energy distribution is also flawed in the US in many elements.

That needs to be put up to par. But definitely there’s a little bit of a tightening relationship or a renewed relationship of big oil and fossil fuels back into the fold, which was not necessarily true in the previous administration.

Bertrand Schmitt

I have not seen any direct reference to that in the document, but what I have seen is a prioritization of reliable energy sources as well as embracing new energy generation at the technological frontier. Talking about nuclear fission, nuclear fusion, some enhanced geothermal. I don’t think there is anything specific pro-fossil fuel, for instance, but contrary to previous administration, there is certainly nothing against. Personally it feels like the right balance. We want to move forward, and we want to pick whatever is needed in order to deliver this plan.

Nuno G. Pedro

Why does this matter? I mean, the building of infrastructure. It matters because it’s the underpinning of the overall innovation that we need to see in the market. If we are talking about, for example, on my side as a venture capitalist, the building of startups that will innovate, that will come into the space, that will occupy specific areas. If you have the best infrastructure in the world at your disposal, the likelihood that you will be able to innovate faster and at scale and develop real cutting edge algorithms, platforms, models, underlying infrastructure in and of itself as well is much higher.

It creates an ecosystem of building that is top-notch, that is cutting edge for the world. That matters is, from the perspective of public private partnerships at this moment in time, if they are properly deployed. I think this is the central piece in some ways of what I mentioned earlier of the potential industry policy that we’re seeing from the US on AI. It will really create the underpinning, it will really create the infrastructure and the sandbox for startups to build on at scale and quickly. This matters a lot. If this is well done, we will see an ecosystem of startups that will emerge from this faster rather than slower. It does matter at scale.

Bertrand Schmitt

I totally agree with you. I think that we want one success for the tech companies, the small and big operating in the US, and we want the US overall national interest moving in the right direction regarding AI because the US cannot afford to be left behind, not to go to the same pace as its peer competitors like China. It’s really clear that there is some level of competition. You absolutely cannot miss it.

AI would be the future of many industries, but also certainly the future of the defence industry and the future of war. You cannot be left behind that just too much risk. I’m pretty excited that I think all of this is moving us forward in the right direction and maybe one last point, there is a lot of talk that in the US if nothing is done, we are maybe just one year away from reaching max power capacity to deliver AI solution. Meaning that we might be out of electricity as soon as next year in order to add a new data centre. Doing something right now is absolutely so much mission-critical. We cannot be left behind.

Let’s not forget we talk about China as a peer competitor. China increase in energy generation is just totally insane by any metric. There is no country that compare to what China is doing in terms of adding energy generation year after year. Everyone is behind China in that regard. It’s absolutely critical the US and other countries who don’t want to be left behind do something about power generation for their own citizen needs, one, directly, of course, and two, in order to stay ahead in the AI race.

Nuno G. Pedro

That’s maybe a great segue for pillar III, which is leading in international AI diplomacy and security, which is code for… I’m going to be facetious here. Let’s not allow China to get cutting edge ships and other things. Let’s not allow China to get an edge on us in some ways. Rightfully so, there’s a significant thinking and rethinking of export controls and where those export controls apply to not just to chips, but also to AI platforms and systems.

There’s an element of discussion around diplomacy and how to coordinate with so-called allies on export controls. Netherlands, obviously very important. Japan and other countries that will be very, very important as that coalition in some ways of allies against China. That’s basically the element that is being put here.

Then talking a lot about trade tools like secondary tariffs and talking about AI security at home. How do we ensure security around our systems and our infrastructure in particular and so there’s an element here where for me the takeaway is that we want to in some ways write a global play book for how the US keeps its technology out of adversaries or enemies or people that are seen as enemies or countries that are seen as enemies along the way. Also, how do we build the right AI standards for us to ensure competitive advantages honestly over the rest of the world?

Bertrand Schmitt

I think there is in that document a very clear-cut view of one, America needs to lead and be extremely successful in all elements of the AI value chain. Two, America wants to work with its allies to make sure they have access to some of the latest American AI products from chips to models. Three, we want to make sure that there is a blockade around China and China AI in some ways and yes, China can buy some products from the US, but it will be limited. It’s not the latest, potentially not the latest model, certainly not the latest chips.

There is a willingness to block China in some international bodies and standards as China was trying to push its way in some of these bodies and try to push its own interest, unsurprisingly in some of these bodies. An acknowledgement that we need to do something about it and I believe that’s quite important. Let’s not forget that China, and we’ll talk more about it later. But China has some strong view of its own AI national policy. It’s certainly not aligned with America or its allies.

Nuno G. Pedro

Yes, and we’ve mentioned a couple of the countries already… Obviously, the UK, there’s Japan, Netherlands, for some of you, might be why Netherlands? We’ve also mentioned this incredible company in the past called ASML, which is a Dutch company, and they are the global leader in development and manufacturing of lithography systems, which is crucial for microchips. That’s why Netherlands matters, because of this company, ASML, which is a significant player in that market.

Obviously, there’s a little bit of a thin rope for these countries, like are we aligned with the US or do we want to edge? It feels like this. For me, the stance that’s manifested by this announcement by the White House is you’re either with us or against us kind of thing. I feel there’s some undertones here of alignments, like, “Oh, they are our allies”, which means either you’re our ally or you’re not. We’ll see how that pans out.

Obviously this comes after tariff negotiations and a bunch of other conversations that are happening more broadly around the relationships between these countries and the US, but definitely it feels a little bit like you’re either on our side or you’re on China’s side. It’s a little bit where we’re seeing some of these conversations lead to.

Bertrand Schmitt

To be frank, ally typically connects to some type of military alliance. Let’s not forget that huge underlying piece of the puzzle. I think it’s pretty fair to say that if you want a military alliance with the US in order to be able to defend yourself better against some other countries, be it China, Russia, then you are going into a military alliance with the US, and obviously we have NATO in Europe, but there are also military alliance with Korea, with Japan, with many countries around the world.

I think what the US is also doing is enforcing it. “Hey guys, if you want that military alliance, and usually we know you want it, there are some economic implications. There are some policy implications in that situation. AI policy implication.” I think it’s quite reasonable to align the different pieces of the puzzle. It’s not a pick and choose, whatever you want. It’s not that there is a need for some real alignment, else it doesn’t really work any more. If you suddenly play the game on both sides for some piece of the puzzle. But you still rely on us for the military part. How does it really work?

Nuno G. Pedro

Yeah, maybe bringing it to one or two further elements that are pretty vital to this discussion. We’ve already talked about security and this notion of enforcing security and security around systems, around infrastructure, et cetera. But as Bertrand mentioned earlier in our conversation, basically there is also a push towards open source and open innovation and some might see that there’s almost like an element of duality in this. How can we be more secure and more siloed and more in control with a group of allies and at the same time enforce or push for open source initiatives to scale?

I personally think many others would probably see it as such, see this as a geopolitical move in and of itself. In some ways what we see in the world is that China seems to be quasi-leading the way, if not leading the way on open source initiatives and there’s a little bit of a concern that if China leads the way on open source initiatives, that it will have embedded its own Trojan horses into some of these initiatives and that the US needs to take a much more proactive and aggressive perspective on it over time.

We know it’s not totally true, it’s not like totally true that China has been doing all itself, but there’s definitely a lot of push by China on open source that the US seemingly wants to counter with some of the actions around this plan.

Bertrand Schmitt

I think you make an excellent point, Nuno. I cannot agree more. China has been playing smart. Like in a lot of markets, when you are not leading with your product, you try to go open source. That has been China approach to the AI race. I think they have been very smart doing that, and I’m glad that finally the US came to the realization that the answer is not just for me, it’s not forbidding open source or just promoting closed source, but it’s also to promote open source because some companies will absolutely want and need open source solution, some companies and governments, open source solution that they can run by themselves, control by themselves. It’s a mandatory part of the landscape, I believe, and it was the right thing to do.

Nuno G. Pedro

We get to compare approaches. We get to look at the US action plan versus China’s strategy, which we already talked a couple of elements of China’s strategy, and then also around European Union and the EU AI Act. These three big geopolitical masses seem to be doing their own thing. How does the EU compare to this, the AI Act? We’ve talked about it in the past, but how does it compare to it, Bertrand?

Bertrand Schmitt

I think it’s very interesting comparison actually, because I’m not sure we have seen a situation where the differences are that stark between US policy, EU policy, China policy. It’s a very interesting reflection of the economic, social and policy environment of each region. I’m really worried for the EU, just to be clear, I will start with that.

The EU AI act seems all about regulating as much as you can. There is this saying, “The US innovates, China copy, and the EU regulate.” I’m not sure it can be closer to the truth than this situation in AI. It’s pretty bad. I think it’s going to be a real problem for startups emerging from Europe. All these regulations would be too burdensome. As usual, the bigger guys will have an advantage because they can deal more easily with complex regulations.

We have seen already that many models and many solutions from OpenAI, Apple, Google and others are not available in Europe. Even some open source models, you cannot run them in Europe. I think it’s a crazy situation. It’s making very difficult to run an AI business from Europe, especially the underlying infrastructure. There’s a lot of risk of fines and stuff.

Personally I’m very worried. I think they are really putting the cart before the horses in that situation. Let’s have horses first, let’s make sure they run fast and at some point let’s see what we can do. But here it’s totally the opposite. Personally I’m worried.

Nuno G. Pedro

In a moment where we talk a lot about European competitiveness and the ability to build startups that scale, and AI seems to be, or is this seminal moment of evolution for the world on top of it, you’re putting regulations in a lot of these markets and so if I’m a local player and if I’m a player based in any of these markets in Europe, I mean, it actually hampers me even more. It’s a dual whammy. There’s the regulatory piece where it takes away the innovation from being deployed in some cases in Europe.

I’m not saying everything that is under the EU AI Act is wrong. There’re some elements around consumer protection that I think are valid, but to Bertrand’s point, putting the cart in front of the horses, so to speak, does create this dimension where it also hampers the ability for startups in Europe to scale and scale fast. In terms of access to data, in terms of access to elements that are important, and guess what? The US is seemingly taking away some of these elements and putting them in a world that allows for that innovation to actually happen.

China, as we know, doesn’t have tremendous restrictions on data as long as there’s political alignment, let’s just call it that. And so if you’re a EU startup, you’re immediately at a disadvantage. Immediately at the disadvantage. To start, which is obviously not a good place to be in, knowing that you have other elements of disadvantage like ability to raise funds at scale, access to talent at scale as well, et cetera. All in all, it’s not just that the regulatory environment can have a negative effect, it also has a negative effect on fundamental innovation from the ground up.

Bertrand Schmitt

You are totally right and the other big piece of the puzzle, energy production, energy distribution. The EU is in a catastrophic situation, let’s not mince our words. Many countries in EU have seen degradation of the energy production, many industries have been impacted. Energy prices in general in the EU have skyrocketed, putting many industries at, let’s be frank, total catastrophic risk for their own survival in the coming years. Forget about having additional spare capacity for AI. That part is also very scary and a big part of the puzzle, as we have seen with the White House AI Action plan. It’s regulation on top of existing regulations, with on top of it a crippled energy market compared to China and now the US.

Nuno G. Pedro

Then there’s China. I mean, we can discuss at length what we know and what we don’t know. What we do know, it’s very clear that they want to be a world leader by 2030 with tremendous investments across the board on everything, research, education, startups, infrastructure, energy deployment and distribution, which by the way, on the energy side, they’ve been doing it for several decades. It’s not a new thing.

We have large players already in the market, state-owned and tech giants like the Baidus, Alibabas, Tencents of the world. That’s the other guys, and the other guys have industrial policy, and they’re putting money behind it. I mean there’s no doubt and a lot of it is state developed, state sponsored, state funded and so you better bet that they’re going to build infrastructure themselves, and they already are, to create the underpinnings of that ecosystem. As I was reflecting earlier that the US is now trying to create, China’s been working at it for a while. Very, very interesting to see how this plan in the US will now compare in terms of speed and deployment versus China’s plans in developments at this moment in time.

Bertrand Schmitt

Regarding China, I think they are trying to do what’s best for their country. There is nothing to criticize from this perspective. It’s a rational plan from their perspective. As we discussed before, I mean China energy power plants and distribution plants are not new. They come at a perfect time. China is building 150 nuclear reactors for the coming decade. It’s already in motion. These are very safe force generation nuclear reactors. That will be able to underpin either additional factories’ production in general, but also of course AI.

The other piece of the puzzle is that definitely China is pushing AI realize it’s critical for their leadership. It’s also very reasonable, as you say. For sure, they are not as encumbered by regulations at other places, especially like the EU. At the same time, there are sometimes regulations to protect people in China. I don’t want to be too misleading here, but of course there are some dark sides. China is clearly pushing their own agenda, their own political agenda. They want generative AI to reflect “core socialist value”. That’s certainly not flying with me. That’s not something we want to support, we want to enable, we want to even I would say accept.

I hope there would be some pushback on products coming from China if they try to push this agenda and that’s really proof for me in some ways that we need a real alternative, and I’m glad the US is pushing its own more free solution.

Nuno G. Pedro

Yes, obviously the free flow of information that is not aligned with what it sees as a positive influence on its citizens will not be allowed. That is obviously a handicap. There’s also a handicap that we know China currently still has in terms of the production of high-end chips. I think on the other side you would be foolish to bet against China in terms of technological catch up. They will catch up, it’s just a matter of when, not if.

On the data side, although there’s these elements of what we in the West would call censorship, there’s also these elements of… There’s probably a lot more easy appropriation of data in general-

Bertrand Schmitt

Oh, yes.

Nuno G. Pedro

-that can be done by private and public companies, so to serve, that’s a balance. Last but not the least on the China factor and how I think through China one very clearly on talent, a lot of the world’s talent on AI we keep saying is in the US is like sort of right. It probably was educated in the US but a lot of it is actually Chinese. The second element that we can’t forget, which obviously we won’t discuss at length today, we have talked a little bit briefly ab