A couple of years ago, I started seeing young startups raising early-stage venture capital funding and diving into the GenAI revolution. Investors are betting that GenAI will open the door to a new Industrial Revolution, where GenAI transcends from high-tech to becoming ubiquitous across agricultural, industrial, and service sectors.
What is becoming obvious is that this revolution carries a significant increase in capital investments, especially when compared to previous cycles. In a way, what’s happening now is not dissimilar to what we saw in the late 1990s, when thousands of servers were being racked up in dedicated colocation facilities, financed with VC money to create consumer experiences pursuing a blitz economics paradigm. Eventually, the boom was followed by the 2000 burst, and the lack of profits and huge capital expenditure led to only a few companies—those that had been very conservative in their capital spending and profitability—surviving and becoming the tech giants of the next decade, such as Yahoo, Amazon, and Google.
From a hardware advancement perspective, particularly in GPUs based on FLOPS and bandwidth, the current boom in GenAI has been one where a VC dollar invested today is equivalent to 50 cents a year from now. This necessitates investments to benefit from an early-to-market position. Unless a startup can quickly build a moat within 12 months, a competitor can appear in 12–18 months and wipe out the advantage with less risk and less investment.
On the other side of the spectrum, the investment in large models that Google, Meta, and others are making is on the order of tens of billions of dollars, now reaching hundreds of billions. For context, only the G7 countries have budgets greater than $100 billion—the equivalent of Microsoft’s and OpenAI’s plan for a GenAI-centric data center full of NVIDIA gear.
As a consequence, startups not only have to worry about a reduced time advantage of 12–18 months from their investments in GPU cycles (capital), but also about the disproportionate investments a handful of tech giants are making—investments otherwise inaccessible to most companies in the world. This is creating a valley of death in tech investments, where there is very little viability between the early stages up to $10 million checks and the $1 billion–plus rounds.
Fortunately, I believe there are two opportunities in this valley of death: open source and data.
Open Source as a Catalyst for Innovation
Open-source software has emerged as a powerful equalizer in the tech industry, offering startups and mid-sized companies access to cutting-edge technologies without the hefty price tag of proprietary solutions. By leveraging open-source platforms and tools, companies can accelerate development cycles, foster innovation, and collaborate with a global community of developers.
In the context of GenAI, open-source initiatives are particularly impactful. Meta’s strategic release of models like Llama 2 into the wild fosters a burgeoning ecosystem of developers and startups eager to tinker with and expand upon Meta’s groundwork and ultimately make it their own. The threat of someone closing off their access to the tech they depend on is gone, resulting in a willingness to invest and build—a breeding ground for innovation.
But this is not a decision born out of charity but cold strategic calculus—a strategic maneuver to enhance Meta’s operational efficiencies and revenue streams. The integration of external innovations can significantly reduce the astronomical costs associated with AI research and development while simultaneously boosting the technological capabilities of Meta’s platforms.
Moreover, by catalyzing a diverse array of applications, Meta nurtures a growing base of developers, each contributing to a vibrant ecosystem that will largely play out on Meta’s platforms. Furthermore, democratizing access to powerful AI tools lowers the barriers for startups, enabling a new wave of entrepreneurs to create attractive businesses that need customers. This increased demand for customers will directly translate to increased revenue for Meta as more businesses use their sophisticated tech to reach potential customers who are spending more and more time on Meta’s platforms.
Keeping Ahead of the Competition
Meta’s decision to open-source Llama 2 can also be seen as a defensive maneuver in the tech industry’s strategic game. By ensuring that a high-quality foundational AI model is freely available, Meta prevents potential over-reliance on competitive technologies and fosters an environment where global talent contributes to their platform, keeping it at the cutting edge of technology. This approach not only secures Meta’s independence but also sets a standard in the industry, potentially steering the direction of AI development towards an open, collaborative future.
Correcting Misconceptions About OpenAI
It’s important to note that while OpenAI started with open-source principles, it has shifted towards a more closed model in recent years. Unlike Meta’s open-source release of Llama 2, OpenAI’s GPT-4 and other advanced models are proprietary and not openly available for modification or redistribution. This distinction highlights the significance of Meta’s open-source strategy in democratizing access to advanced AI technologies.
Data as a Strategic Asset
While open source provides the tools, data serves as the fuel that powers GenAI systems. Proprietary data collected from unique sources can offer a competitive edge that is hard to replicate. Companies that harness their data effectively can train specialized models, improve product offerings, and create truly unique experiences for customers.
In our industry, the sheer volume of data generated by vehicles—ranging from sensor readings to user interactions—is invaluable. This data can be leveraged to enhance machine learning models, improve safety features, and optimize operational efficiency. By focusing on data collection and analysis, companies can develop proprietary insights that set them apart from competitors.
Moreover, strategic partnerships for data sharing can amplify these benefits. We’ve seen the deal between Wayve and Uber in this space, and I expect more to emerge in the next few months. We need to urgently play in this area to compete in the long run.
Navigating the Valley of Death
By capitalizing on open-source technologies and proprietary data, companies can effectively navigate the investment valley of death. This approach mitigates the need for colossal capital investments by maximizing existing resources and focusing on strategic advantages.
Open source reduces dependency on expensive, closed-source solutions, enabling rapid iteration and innovation. Proprietary data, on the other hand, offers unique insights and capabilities that can’t be easily duplicated by competitors or new market entrants.
For startups and mid-sized companies, this dual focus creates a sustainable path forward. It allows for continued growth and competitiveness without the necessity of securing exorbitant funding rounds that are increasingly scarce.
Conclusion
The GenAI revolution presents both formidable challenges and unprecedented opportunities. The current investment climate may seem dominated by tech behemoths with deep pockets, but the strategic use of open source and proprietary data offers a viable pathway for others, including us at Nexar, with our unique IP, exclusive dataset, and a network capable of continuously collecting data.
By embracing open source, companies can accelerate innovation, reduce costs, and foster collaborative growth. By leveraging proprietary data, they can develop unique solutions that provide a competitive edge. Together, these strategies enable companies to not just survive but thrive amid the seismic shifts brought about by GenAI.
In the automotive AV tech industry, where technology and data are paramount, these opportunities are particularly pronounced. By focusing our efforts on these areas, we position ourselves to lead in innovation, adapt swiftly to market changes, and ultimately drive the future of autonomous technology.