A simplified history of venture
The history of venture capital (VC) can be abstracted to specific developments that enabled a new generation of companies to drive markets and further technology development, creating disproportionate amounts of value in the process.
Incumbents have sometimes captured more of the value created by these developments due to structural & market advantages; other times, startups could get away with most of the new value. Note that this distinction is not necessarily related to VC returns in that era—a venture investor can make a killing if the value created is sufficiently high despite 70-90%+ of the overall value created going to incumbents. It does, however, change the overall odds of massive success for any given company. When we think of venture capital, the story goes more or less like this:
- The OGs: The inception of the semiconductor industry and the mass production of chips, on the shoulders of the first venture firms (ARDC, Venrock) and the Traitorous Eight (Fairchild Semiconductor, Intel, Amelco, Kleiner Perkins), laid the foundation for venture capital to become its own asset class and for companies like Microsoft and Apple to become the poster children of the first professional venture capital wave. It also marked the beginning of the application-infrastructure cycle[1] that has fueled many of the subsequent venture capital eras.
- Internet Companies: The arrival of the internet marked a significant shift. The first internet wave rewarded mainly infrastructure/platform players like Cisco, Yahoo, Netscape, and Microsoft. The frenzy was intense enough to eventually lead to one of the most well-known bubbles in history. Many captured phenomenal returns in this period—and many lost everything when the bubble burst. The underlying trend, the establishment of the internet and easy browser access, gave way to the second internet wave, providing startups in the application layer, such as Google, Amazon, and eBay, with the opportunity to capture considerable value.
- Shared Economy, Cloud, SaaS + Mobile: The subsequent era saw several waves emerge together, all leveraging or building on widespread internet access to empower users and enterprises to do more online. What initially started with staying in touch online (social media) and the massive success of Facebook despite incumbent attempts to win (remember Google+?) subsequently became a broader ‘shared economy’ trend that capitalized on mobile and included Airbnb, Uber, Whatsapp, Youtube, and more. The concurrent SaaS wave was marked by several massive startup successes as well (e.g., SalesForce) that also capitalized on mobile natively. Most of the value in the cloud went to the infrastructure layer, and incumbents made a killing there.
- First AI Wave: The first AI wave (mainly ML) has seen incumbents like Google, Facebook, and ByteDance capture most of the value so far, likely due to factors like proprietary data advantages, extensions fitting well into existing products, and the AI improvement not being significant enough to enable a startup to outcompete the incumbent (in many/most cases).
- The New AI Wave: This nascent wave suggests a potential shift to startups capturing a major chunk of the value created. The considerable technology improvements, massive model training on public data (with potential further refinement on proprietary data), and entirely new use cases enabled by the technology may have created the space for several massive AI-native winners. Nvidia and other chip-making companies have so far captured (and may continue to capture, at least while compute remains the bottleneck) a significant fraction of the value, and similar to the first wave, product extensions will also enable massive value capture by established companies like Microsoft. However, companies like OpenAI, Anthropic, or Midjourney have shown that winning against the incumbents is possible (e.g., Google invented the transformers that power most LLMs). The jury is still out regarding how much value will accrue to those building applications. Almost everyone expects models to commoditize, but perhaps the model companies will nevertheless succeed by vertically integrating, or incumbents will leverage their distribution power to win.
What About Hardware?
If you’ve been following along, you probably noticed that we did not mention many companies working on hardware. And the truth is that, for the most part, mainstream venture capital left behind its hardware-intensive days (chips, mostly) a long time ago. While there have always been exceptions (e.g., Nvidia was founded in 1993, the dot-com bubble saw tons of infrastructure investment, and some semiconductor technology venture investment has always existed), the two most significant departures from the overall trend signaled by Marc Andreessen’s software eating the world are the emergence of biotechnology venture capital (a success) and clean tech in the early 2000s (a failure). Let’s take a look at both.
- Biotech: Despite long timelines, heavy technical risk, and substantial capital needs, biotech has attracted significant venture investment since its early days. Some of the reasons include the potential for monopolistic returns from patent-protected breakthroughs, non-linear value creation from curing diseases, supportive regulatory environments with accelerated approval paths, venture investors with deep scientific expertise to assess technical risk, and syndication with pharma corporates now accustomed to valuing real options to share risk and provide an alternative exit avenue. High octane investments in a sophisticated supporting ecosystem.
- Cleantech in the 2000s: The cleantech VC boom and bust of the 2000s looms large in the current frontier tech VC landscape. Many of the same risk factors are still present today: long adoption cycles, high technical risk, uncertain market adoption timelines, capital intensity, policy risk, and more. But so are the massive, relatively stagnant markets that enticed investors back then. Macroeconomic conditions did no favors here, either (e.g., oil went from $40-60/barrel in 1985-2003 to >$100/barrel in 2005-2015; it was reasonable to think cleantech could compete with fossil at $120+/barrel prices!). Ultimately, it was a massive write-off for the industry, and many investors who were part of that still think it’s not a winning area today.
In theory, the allure of investing in deeply scientific hardware (”frontier tech”) is that companies should be much more defensible in success than software is. This is not always true, and while network effects loom large in the software world, in general, it’s easier to outcompete a software company, given low barriers to entry, fast pace of innovation, and limited marginal costs. It’s much harder to go head-to-head with a successful chemical producer, even if they were a startup a decade ago. What is the price to pay for that increased defensibility, however? Let’s take a look:
- Potentially longer timelines: While some ecosystems allow for fast scale-up (e.g., semiconductor technology, if compatible with existing fabs), many others don’t. Thus, it can take a long time for a frontier tech company to move from the laboratory to the kind of large-scale production and operations that enable the most successful exits. While the timeline of a standard closed-end fund should still be enough to succeed (10-12 years), frontier tech presents a generally higher timeline risk.
- Capital needs separated from PMF: One of the most common complaints heard about frontier tech is the “capital intensity” of these companies. We generally challenge that (Uber raised >$20B before its IPO!): the issue is not with the amount of capital required but with the timing. The biggest challenge frontier tech companies face is the fact that building technology from scratch is expensive and requires capital before product-market fit has been proven. This is their Achilles’ heel: early capital is not mainly dedicated to growing company sales (as it happens in software) but to developing and scaling up the technology. As one would expect, fewer investors have the risk appetite for it, and valuations tend to be lower to reflect this elevated product-market fit risk.
- Increased return hurdles: With rates close to zero for a long time, investors started chasing returns, and venture capital, including frontier tech, was an exceptional place to do so. With higher rates and low-risk ways to earn close to 4-5% a year (vs. 0% previously), the venture illiquidity and startup risk premia look less attractive. Thus, there is less interest in funding risky, early technologies when an allocator no longer feels challenged to attain target portfolio returns without the potential excess return of venture capital. This point exacerbates the impact of the two challenges above.
The Frontier Tech Investment Case
Knowing this, the question that comes now is the most dangerous question to ask. However, because the potential reward is large enough, one has to ask it anyway: is it different this time around? Can frontier tech follow the steps of biotech and become a venture success story? We certainly think so, and this is why:
- Truly massive opportunities: Frontier tech companies are tackling trillion-dollar markets and challenges, often solving problems that humanity has no choice but to solve. For the planet to thrive over the coming decades, many things need to change in the world economy, and frontier tech solutions will be significant drivers of this change, from our food systems (SuperMeat, Fork & Good, EVERY) to the materials we use (KEEL LABS, Modern Meadow) and the things we produce (AtmosZero, Electrified Thermal Solutions, Ammobia). These are challenging markets, but they have the potential for genuinely transformative impact and enormous value capture. Furthermore, because frontier tech is premised on differentiated IP rather than blitzscaling, valuations remain attractive at the earliest stages.
- New markets: Frontier technologies also enable entirely new industries & value chains. The explosion of the space economy, for example, is premised on reusable rockets and cheap launch costs (SpinLaunch), re-entry (Outpost), space mobility & communications (Skyloom), and the room to build an economy up there (Max Space). Foundational technologies have the potential not just to solve existing problems but to enable humanity to do more. Some, like photonics, already exist but continue to develop and enable new capabilities. Others, like quantum computing (Ephos) or nuclear fusion (Commonwealth Fusion Systems, Xcimer Energy, Thea Energy), offer a tantalizing view of a future where humanity is no longer constrained by energy availability and where some of our hardest problems no longer sit behind an unassailable computing barrier. Even others promise to create a new economy out of the ashes of the old one (Twelve, Modern Hydrogen).
- Ecosystem growth & talent flywheel: While the first wave of cleantech was a relatively more niche endeavor, with few investors and founders, the ecosystem has grown to encompass hundreds of venture capital firms and thousands of companies. This has created a positive feedback loop for the ecosystem, with a growing understanding of what it takes to succeed, founders and early employees with the battle scars to get it right, and a much wider founding talent catchment area. The best founders and scientists now know that this is a viable career path and increasingly choose to devote their careers to making change happen.
- Climate urgency & societal support: The unprecedented global focus on sustainability and climate change is one of the fundamental tailwinds for frontier tech solutions.
- Public initiatives, such as the increasing role of government agencies like ARPA-E or the growth of legislation like the Inflation Reduction Act, are helping to non-dilutively fund promising technologies at their earliest stages and are opening the door to companies leveraging the government as a first customer, for example, thanks to more easily transferable tax credits that can make early, risky projects bankable.
- Similarly, private companies are also stepping up, with an ever-growing number of incumbents willing to take technology risks and more creatively engage companies for earlier pilot projects, infrastructure access, and conditional offtake/purchase agreements. This is an essential part of resolving the funding-PMF challenge mentioned above, as these early customer commitments allow frontier tech companies to prove customer traction ahead of delivering products at scale.
- Maturing financing stack: To further tackle the PMF-capital challenge, frontier tech companies are learning to explore the entire gamut of financing options. Before, unsustainable valuations were needed to accommodate the dilution brought by equity financing of large capital expenditures. This was only possible due to a low-rate environment that pushed many pools to capital to find returns in increasingly risky deals, but it was a curse in disguise, as these deals further complicated the crucible moment where venture valuations must meet the real world (revenue, profit, growth). As new funding avenues open to frontier tech companies, such as venture debt and project financing, the ecosystem becomes more sophisticated and sustainable: it enables frontier tech companies to finance with company equity only what truly requires it. Further, a varied capital stack also favors companies that are built for capital-efficient growth from day one, preserving more company equity and better protecting the interests of venture investors.
Conclusion
Venture capital is entering a new era marked by a return to investing in foundational technologies that address significant global challenges. While this new wave presents unique challenges, it also offers unprecedented opportunities for transformative impact and venture-scale financial returns. The path forward for VC in frontier tech will require adaptability, patience, and a strategic approach to investment and support for portfolio companies, leveraging the lessons learned from previous eras to navigate the complexities of bringing groundbreaking technologies to market. We are privileged to be part of a shift that empowers the fusion of scientific brilliance, entrepreneurial ambition & resilience, and mission-aligned capital to build a better future. Starlight could not be more excited about what lies ahead for the frontier tech ecosystem.
Footnotes
[1] This cycle works as follows: some technology development enables certain applications; those applications eventually require further infrastructure development to keep growing; the new infrastructure provides the base for novel applications and use cases on which the cycle continues.
Return to article
