The editorial committee shares their insights

Open discussion with INPHO Venture Summit

logo inpho white

The editorial committee shares their insights

logo inpho white

Open discussion with INPHO Venture Summit

What future for Deep Tech ? Our editorial committee answers.

You were more than one hundred to have joined our editorial committee on February 28th.

Today, we are delighted to share with you the podcasts covering six topics that caught our attention.

Click on one of the links below to discover their insights:

Counteract a monopolized world in deeptech

With George UGRAS, Managing director at AV8 Ventures and Christian REITBERGER, Partner at Matterwave Ventures

Christian Reitberger expresses concerns about a potential future dominated by Nvidia, where all computing services are provided by a few major companies, creating a monopolized and homogenous landscape. He advocates for a more diverse and colorful world, where system integration incorporates various technologies and architectures.

George Ugras shares optimism, citing historical precedents where dominant market shares have been challenged by innovative newcomers.

CHRISTIAN REITBERGER

I Try to think in worlds that I would not want to happen. And a world that I do not want to happen is one where all of compute is provided by Nvidia.

All the uses of compute are provided by companies like open AI, stability, Mistral, olive Alpha, that need billions of Euros per quarter to sustain themselves, which they shell out completely to Nvidia.

We should live in a much more colorful, heterogeneous, and much more differentiated world, where system integration happens, incorporating all of these ideas from Neuromorphic designs to new memory architectures, to ultra-low power architectures, which needs pretty smart buyers and pretty smart system integrators. 

And here I am now thinking about the likes of an ATOS or Cap Gemini, the people who are advising users that think about the supercomputing needs, their data center needs, of how Heterogeneous Compute should actually be made to work. And we don’t want this. 

We don’t want a monopolized world, which is being sustained by 5, 10 billion VCs that fund these mega cash consumers. We want to a more colorful world, and I think that is where we can then come in and invite the local corporate localizing 500 kilometers around Bordeaux, the corporate ecosystem as well.

 

GEORGE UGRAS

Yeah, I’m very optimistic that future is not going to happen Christian. In 1965, I think IBM’s market share was 65 percent, I remember those numbers, 65% market share in 1965. And, yeah, pretty much done.

There is some kid about a mile and a half from my house at Stanford right now, who’s a 19-year-old, and was thinking about how to make that happen to Google. It’s good.

 

There are alternatives to Costly LLMs and GPUs

 With Eric Benhamou, Founder at Benhamou Global Ventures

Eric Benhamou discusses two main themes highlighted by George and Christian during the conference: infrastructure and applications. He acknowledges the capital intensity associated with AI generation, raising concerns about industry concentration. Benhamou emphasizes the need for a debate on open versus closed models in AI development. He expresses optimism about the potential of scaled AI applications, particularly in vertical deployments, to disrupt industries. Benhamou also discusses the challenges and opportunities in infrastructure, including the shift from cloud-centric to edge and on-premise computing. He underscores the importance of robust cybersecurity measures against emerging threats posed by Gen AI. Overall, Benhamou anticipates fruitful discussions at the conference, particularly regarding breakthroughs in AI deployments and productivity enhancements.

ERIC BENHAMOU

So, first, I, like the way George has framed the two major tracks that we could use at the conference, less infrastructure, versus applications. And also, I like, very much, the important problem that Christian has posed, which is, I’m struck by the immense levels of capital intensity associated with this generation of AI, it’s absolutely mind boggling, the more than $100 million to train GPT4, Altman now wants to raise $7 trillion to build the next generation GPUs. For calibration, I think 7 trillion is more than twice the GDP of France, so nobody wants to live in a world that has this level of capital concentration, industry concentration. 

 

So, I think there’s going to be a big debate in the next few years about open versus closed, proprietary models, versus, open models. If you put something in the open, you have to be pretty comfortable that it is safe. So there has to be regulations around it. If you keep it proprietary, simply because it said you are not sure it is safe. Then we have big societal problems there. But I think the application track of the conference should be very interesting, because after, after Gen AI has come on the scene in full force with GPG and so on.

The anticipation of applications has risen and the return on AI expectations are extremely high. They are certainly very high. In the minds of investors, the last wave of earnings calls, every analyst is asking CEOs, what’s your return on AI investments? So I think, after a period of experimentation, in 2023-24, we’re going to have to see some scaled applications that disrupts and transform industries.

We’re investing in a few, and I’m sure that many, many panelists here are investing in others, but I’m very, very excited about the vertical AI space. And we’re going to start to see some low hanging fruits materialized at scale.

 

Gemma has no excuse me on the panel today, but I know she. She raised a share of effect. Coding at scale, enhancing productivity of knowledge workers, like Software engineers are using it, using Gen AI. I suspect the most promising applications at scale will be those that don’t bypass to humans, but Smit, unable to human to operate much, much higher level.

 

The way we think about this is exactly the same level sets we describe autonomous vehicles, so you have autonomy level 1, 2, 3, 4. And, same thing, I’ve seen what happened with enterprises. Enterprises will raise their level of autonomy. In other words, process automation, such that, human tasks for the elevated to much higher strategic levels, enabling more routine tasks to be handled by AI.  We are going to see this with, for example, customer support applications. Nobody likes to deal with call center operators. They prefer it to deal with people who have much more knowledgeable. This doesn’t mean that call centers will go away.

It does mean that call center operators will use Gen AI to be more efficient to dealing with other humans. I expect that’s a lot of legal projects handled by the General Counsel; will be sped up as a result of the immense capabilities of summarization that exists in Gen AI today?

 

The question that many of us, investors are faced with is: what is the right level investments in terms of magnitude and where do you place your investments in this space? Um, I don’t think anyone of us here on this panel has the wherewithal to invest in next generation of LLMs. That’s way too expensive or next gen GPUs.

 

On the other hand, if you invest in applications, and data and middleware, and RAG enhanced applications (Retrieval Augmented Generation), then you’re very likely to be able to create a complete solution that can be deployed at scale. Now, this exists in a wide variety of different industries and different investors have different specialties and they’ll be able to get returns in different industries. But I’m struck by the fact that, within a couple of year, we’re going to see a few breakouts applications at scale truly, illustrating, the potential.

Regarding the infrastructure, track that George and others talked about, it’s pretty clear that we’re going to have so many different forms of compute. They’re going to have to be put together in a logical set, architectures.

 

I’m struck by the fact that, in the first 15 years of this century, we witnessed a massive migration to the Cloud. Workflows went to the cloud and now we’re starting to see a counter stream where many workloads move back. They moved to the edge or the move back on premise. Sometimes Gen AI privacy concerns cause you to bring your workloads and the data back on premise. But I think This creates a very, very complex architecture with some edge. components and some cloud components from, some on premise components space components and I think there’ll be a very interesting debates about the most powerful architectures from edge to cloud. This morning just before the panel are recharged from our electric vehicle on a charger. And I came to realize this is not the same as the gas pump. This is a very smart edge device.

 

That knows a great deal about my consumption, that recognized my vehicle, it knows who I am, and potentially can be hacked to the point of destroying my vehicle. So, all of this made me think, OK, these are characters who require a very fine level of thinking.

Not just because of the balance of capital deployment, but because some architectures will be very vulnerable to a new generation of cyber-attacks, and others will be strong.

 

So, I think we’ll have, within the infrastructure track, I think we have some interesting debates about what is in the minds of attackers, whether they are state level actors or just hackers, when they’re confronted with this massively augmented attack surface?

I think Gen AI offers the possibility of extremely creative new attacks.

The same sort of social engineering attacks that we witness the last few years can be augmented dramatically with Gen AI.

 

That can deep fake humans, whether, in text conversation, voice, even image.

And we have to figure out a way to protect ourselves from these deep fake attacks, particularly in a world that is geopolitically more unstable.

 

So, it seems to me, we have some very, very interesting topics that we can fix across these two tracks.

And my personal interests, as I mentioned, in the contributions, is that I think the next couple of years, there’ll be some breakthrough returns in vertical AI deployments.

 

I don’t have a crystal ball good enough, clear enough to see, to know which ones but I’m placing my bets right now.

And I think the potential that gen AI at scale offers in terms of productivity boost is tremendous. So much more focus on an industrial and enterprise deployments and consumer deployments.

Energy, Defense Challenges in France & Europe: Spotlight on

With Christian REITBERGER, Partner at Matterwave Ventures

Christian Reitberger highlights two distinct aspects of France: its strong emphasis on nuclear energy and its successful collaboration between major defense players and the startup ecosystem. Christian Reitberger discusses the significance of nuclear energy in France, contrasting it with the preferences of other European countries like Germany. He suggests a lively debate on the necessity of nuclear energy, including Small Modular Reactors (SMRs), to meet energy demands, particularly for industries and data centers. Christian Reitberger also praises France’s model of collaboration between defense corporations and startups, noting its resemblance to successful practices in the United States. He emphasizes the uniqueness of these dynamics and suggests showcasing them at a French conference to provide valuable insights uncommon elsewhere.

CHRISTIAN REITBERGER 

I think there are two special traits of France, that are less represented in other eastern European countries, if we think about, one of our key European challenges, it is the very high energy costs that we have across the continent.

We all would agree that we bet on renewable energies on energy conversion storage, et cetera, et cetera, but in France, there is still this enormously strong lobby on nuclear and the Germans, for example, don’t understand the French at all, and why they like it so much.

 

So, to give some spice to the discussion, one question could be for a dependable baseload energy supply, with the still rapidly ramping electricity needs of reshored industry, and all those data centers all filled with NVIDIA GPUs. Is nuclear required and, do SMRs play a role.

 

I’m seeing that because you could invite ORANO or EDF. One could invite a couple of the SMR players to have a very lively discussion that you would not have anywhere else, nuclear is part of the energy mix of the future.

Just what food for thought. And the other one is different. Where we’re France, again, France has built a very, very close and tight relationship between the 5-6 major defense players that you have, SAFRAN, THALES, et cetera. And your startup in VC community. The rest of Europe is only catching up to that. I know continuously being invited by defense organizations, contractors that want to talk about dual use technologies and investing in dual use technologies.

 

France, as I learned a couple of weeks ago with an organization, with, one of the defense players is advanced here. You have seemingly you have found the model how to establish these collaborations between this open innovation platform, between your Big Defense Corporates and the startup community. Something that the US has done in the last 30 years and trained it extremely well. And we probably should learn.

 

And that is something that I think you could definitely demonstrate at the French conference; I’d be hard pressed to find more than three German participants who could talk with credibility about this topic. Yeah This would be special. You would not find this anywhere else.

GenAI: Addressing Infrastructure Needs for LLMs

with T.M. Ravi, founder at The Hive

T.M. Ravi presents four key discussion topics: Gen AI, climate, health (specifically cancer), and defense. He emphasizes that Gen AI and infrastructure extend beyond GPUs to include middleware, new-gen operating systems, and a data layer for Large Language Models (LLMs). Ravi underscores the importance of integrating infrastructure with applications and improving operations, coding, R&D, and design. He highlights the CEO of Nvidia’s statement that 30% of new code at Google is Gen AI-generated, indicating significant advancements. Discussions also touch on vertical and horizontal applications, providing ample material for debate.

T.M RAVI

Maybe, at least, I cheated I have maybe four topics. one is kind of more fundamental, which is Gen AI, that everyone’s spoken about, and then three sort of fundamental issues of our time, Climate, health cancer and defence. 

 

And on the Gen AI side, when it comes to infrastructure, it’s not just GPUs. So, it’s kind of the middleware that is going to be an emergence of new gen OS. There’s a whole new data layer to make the data ready for LLM. 

 

So, so there’s a lot to discuss in the infrastructure side, but also kind of, like the, structuring the judge propose of infrastructure plus applications and use cases, You know, the use cases applications, just operations, how operations going to fundamentally kind of improve. You know, there was discussion of coding. And so just R&D and design substantially changing.

 

I do largely agree with the quote from the CEO of Nvidia. Already we are hearing that 30% of the code at Google, the new code is Gen AI generated.

And it’s there’s just a lot going on there. And then the kind of vertical and horizontal applications. So, I think I think we have a good, good lot of meat here to debate.

Frugal and Edge Computing

With Dieter KRAFT, Director ar TRUMPF Venture and François TISON, General Partner at 360 Capital Partner

Francois Tison challenges Nvidia’s viewpoint, advocating for less compute power to sustain the planet. He sees a global shift towards frugal computing and believes Europe’s engineering strengths position it well for reindustrialization.

Dieter Kraft highlights the trend towards more efficient computing, particularly on the edge, and its potential to reshape architectures. Both identify these topics as key for discussion at the INPHO event, emphasizing collaboration to address challenges and seize opportunities.

FRANCOIS TISON 

I think the comment of the Nvidia guys is extremely arrogant and goes. You always complete, completely contrary to where, where we’re going. I mean, there’s no way in Hell, you know.

 

The more you increase the level of abstraction, the more compute you need to achieve simple tasks, and we don’t need more compute to achieve simple tasks. We need less compute to achieve simple tasks, because the, you know, our planet cannot sustain more compute. So, I think I do think that, know, as I said in the beginning, we did a lot of CLEAN TECHS a lot.

 

In 2005, 2010, it was really, it was not the right time. I think now, we’re in an environment where both the public and the public at both ends, so, the consumers and the politicians that represent the consumers, are now extremely aware, and are pushing us and are pushing everyone to move into more frugal, way of consuming, and more frugal, ways of computing, cleaner, electricity, recycling, et cetera, et cetera. 

 

Net zero, but beyond net zero or just a more frugal economy. Generally speaking, and we’re talking about the reindustrialization of Europe. I think those two subjects are linked, because in order to be more frugal to recycle more, we’re going to need to promote with smarter, smarter solutions, more engineering precision, engineering, stuff that mixes, engine, mechanic, precision mechanics, and Chemistry to produce, energy. Store energy, etcetera, etcetera. 

And this is one of the traditional strengths of Europe. There are two places in the world where, where these technologies are really strong. It’s Europe, around the Alps, Austria, Switzerland, Germany, Italy, France, that area, and Japan.

 

But, but, Europe is one of the strong areas in the world in terms of Engineering on these, on these topics, so, I do think that reindustrilaisation and going towards smarter, more efficient solutions and a more frugal economy, those are things that play to Europe strengths, and they go together, and I think these would be certainly, one of the interesting topics to cover. AT INPHO

 

 

DIETER KRAFT

But think it’s, it’s indeed interlinked between, let’s say, more computing but also more efficient computing.

which leads to a trend I tend to observe, not really sure if it’s already a trend, but what we see is, in different European countries. Netherlands and Christian knows about that or Dresden, efficiency on the edge, AI with components.

 

Also, hardware investments, which are more or less enabling on the edge, computing with much more efficiency, than in the past.

And that totally changes the architecture, or might change the architecture, And might also, let us think about what can we do in order to lower down the power consumption of distributed computing?

 

That is something where we have a deeper look. And I think it’s potentially a point for discussion in the INPHO event because that is, for me, something, what is really differentiating, and where we need syndicates. Which can go together that way, might be capital intensive, but it might be huge in the impact.

Evolving Computing Hardware

With Mathieu Costes, Airbus Venture

Mathieu Costes highlights the evolving computing landscape, emphasizing Quantum Computing and emerging hardware technologies. He discusses the importance of these advancements for Venture Capital investments, particularly in early-stage ventures. Costes also underscores the need for software orchestration in hybrid architectures and the urgency for sustainability in data center operations.

MATHIEU COSTES

One thing, one theme that I think, resonates with what we discussed over the last few minutes, is the world of compute because, obviously Quantum Computing and the world of hardware is evolving. Who will win the race between ion trap, neutral Atom, and others? I don’t know, but things are progressing heavily.

 

If you think compute, we’ll go to the known world of CPUs that measure low consumption CPUs of tomorrow, GPUs with the Nvidia story that could enable AI and other stuff, QPU is quantum, The world of Neuromorphic chips may come in, and finally mature.

Then we see the world of photonics, that may be much faster to move a photon, to save time when latency and speed is critical and Think about a DNA based storage and chemistry-based computing that could offer unprecedented, paralyzed tasks. So, the world of compute, starting with the hardware, to me, is the key element that is relevant for Venture Capital, especially in early stage.

 

And on that one, you could even see the world of software to orchestrate between Solver, and the world of workload scheduler in this new hybrid architecture, will have a key role to play. This is a this is a global problem. Because obviously we know the world of AI, which generate a huge volume of workload. between that one, weather forecasts, scientific computing and whatever computing to be needed tomorrow. I think we’ve got a very promising field to share news and hopefully, too, deploy more capital into in the coming years.

 

I would even mentioned, sustainability, how to have datacenters, CPU, GPUs, QPUs, that will not absorb the full volume of green electricity produced. You see what happened in Ireland a couple of months ago, they stopped the growth and expansion of the datacenters. Say, they are consuming 18% of the electricity generated into the country, which is exactly the same level of the energy electricity consumed by the citizens. So things will have to change.

Join our actions

Blumorpho works accross the whole blue economy ecosystem. We are happy to have you involved in more actions.

logo inpho white

INPHO Venture Summit

A summit focused on investment, answering the question “What’s Next to Invest in Deep Tech?”

BLM Detection 6

Share & learn on our Platform

The first interactive platform that brings together innovation players around common interests.
Days
Hours
Minutes
The Inpho Venture Summit 2024 is live now!