The venture capital landscape is in constant flux, but few shifts are as significant as the growing prominence of the secondary market, especially within the capital-intensive realm of deep tech. To unpack this evolving dynamic, we turn to Sriram Viswanathan, a founding managing partner of Celesta Capital, a firm synonymous with deep tech investment. Established in 2013 and based in San Francisco, Celesta Capital has carved out a unique niche by exclusively focusing on deep tech, driven by a foundational belief that truly transformative technological advancements are rooted in innovation at the infrastructure layer. With an impressive $1.1 billion in assets under management, Celesta Capital boasts a formidable track record, having made 110 investments to date and achieving 43 exits, underscoring an exceptional exit-to-investment ratio that speaks to their strategic acumen and deep industry insight.

Celesta’s portfolio of successful exits reads like a who’s who of cutting-edge technology, including chip design pioneer Credo, the innovative AI processor developer Habana Labs (later acquired by Intel), semiconductor firm Innovium, drone systems specialist IdeaForge, and digital cell biology trailblazer Berkeley Lights. These successes highlight Celesta’s ability to identify and nurture companies poised for significant impact. However, a broader trend is reshaping the venture ecosystem: startups are remaining private for extended periods, necessitating new avenues for liquidity. In this environment, the secondary market is rapidly emerging as an indispensable engine for venture capital, particularly for deep tech companies that often require substantial time and capital to bring their complex innovations to fruition. Viswanathan recently engaged in an insightful discussion with Crunchbase News, delving into the factors accelerating the secondary market’s growth, its specific implications for deep tech, and the broader opportunities he sees within the sector.

"Celesta has been doing deep tech from day one, and we were always investors in hardware and semiconductors and systems and the like," Viswanathan recounted to Crunchbase News, emphasizing the firm’s long-standing commitment to the sector. "And so in a way, we’re thrilled that what we do is literally at the core of what the opportunities are and what people want to do." This foundational expertise positions Celesta Capital at the heart of the current technological revolution, particularly as deep tech, with its emphasis on foundational innovation, gains unprecedented attention. The following Q&A, edited for brevity and clarity, captures Viswanathan’s profound insights into these critical market shifts.

Crunchbase News: Why do you think the secondary market has been so active, and what’s driving demand?

Viswanathan: The heightened activity in the secondary market is a pretty obvious outcome of fundamental economic principles at play in the venture capital world. When you have an abundance of capital aggressively pursuing a limited number of high-quality deals, investors naturally begin to seek alternative avenues for exposure and liquidity. We’re currently witnessing a pendulum swing; there are periods of capital abundance and periods of scarcity. Right now, thanks to the explosive growth in areas like artificial intelligence (AI) and the massive infrastructure buildout supporting data centers, an incredible amount of capital is flooding into the innovation space. AI, healthcare technology, and the expansion of data center capabilities are all experiencing exponential growth, creating a dynamic environment where investment opportunities are plentiful but truly exceptional companies remain a rarer commodity.

As a consequence, what we’re observing is a scarcity of truly outstanding companies that can absorb this deluge of capital efficiently. This drives some of the more discerning and capital-rich investors to look for greater exposure through the secondary market, simultaneously providing an exit route for earlier investors. Historically, early-stage venture capital firms primarily relied on mergers and acquisitions (M&A) or public listings (IPOs) as their primary avenues for liquidity and realizing returns. However, we’re now seeing a substantial volume of secondary transactions that increasingly rival M&A activity. This is, in my view, a very positive development for the ecosystem. It introduces a newer, more diverse mix of investors, often those with deeper pockets and different investment horizons, who can provide crucial liquidity for the foundational, earlier-stage investors. This mechanism allows the venture cycle to continue flowing, preventing capital from being tied up indefinitely in private companies and enabling earlier investors to recycle capital into new opportunities.

Let’s talk about the secondary market as it relates to deep tech specifically. How does this dynamic differ from broader venture trends?

Viswanathan: Traditionally, secondary markets were primarily driven by larger institutional investors or funds of funds looking for more mature, de-risked assets. Their focus was typically on companies with proven business models, established cash flows, and a clearer, shorter path to liquidity—often within a two- to three-year timeframe. These secondary players, usually larger asset groups, operate with different return expectations compared to earlier-stage venture investors, prioritizing stability and predictable returns over high-risk, high-reward early-stage bets.

However, when it comes to deep tech, particularly in the current climate, a distinct "feeding frenzy" is underway. This is largely because investors are keenly observing the monumental successes and valuations achieved in areas like AI training and are extrapolating that a similar, if not greater, opportunity exists in AI inference and other foundational deep tech applications. The market is witnessing a speculative phase where participants are less fixated on immediate revenue streams or robust cash flows. Instead, they are making a strategic bet on identifying the "winner" in what often appears to be a "winner-take-all" or "winner-take-most" market. We’ve seen this pattern manifest with large language models (LLMs), where a handful of dominant players have emerged rapidly. While it started with one or two, it’s highly improbable that we’ll see ten equally dominant players in the near future.

This speculative demand in deep tech, especially AI, marks a significant departure from the traditional secondary market investor’s focus on conventional financial metrics. These investors are making calculated gambles on technological leadership and market dominance rather than immediate profitability. While this intense demand is currently driving up valuations in AI, I anticipate that the secondary market’s demand for AI-specific investments will eventually moderate over time as the market matures and consolidates. Nevertheless, for now, the deep tech secondary market is characterized by this unique appetite for potential rather than just proven performance.

Do you think this hot market is going to continue in the near- and long-term, or are we headed for a correction?

Viswanathan: There’s always an inherent tendency for markets to revert to the mean, as one would naturally expect. Market dynamics often swing more dramatically than the underlying equilibrium would comfortably support, leading to periods of both exuberance and contraction. As a result, it’s almost inevitable that we will encounter some level of frothiness in the current market. Valuations, fueled by an abundance of capital and intense interest, tend to creep upwards, sometimes to unsustainable levels. This is where the "hype cycle" can truly set in, causing some investors to momentarily overlook the fundamental principles of building a meaningful business—one characterized by sustainable revenue, healthy margins, and positive free cash flow. We may very well be in such a period right now, particularly within certain segments of the AI infrastructure buildout.

However, it’s crucial to acknowledge the countervailing forces at play. On the positive side, we’re observing a dramatic acceleration in how quickly a deep tech company can develop and test its products. The time it takes to achieve initial revenue has shrunk considerably, and concurrently, the cost of development has also been significantly reduced, thanks to advancements in cloud computing, open-source tools, and more efficient hardware. This confluence of factors makes the concept of a "one-person, billion-dollar company" less of a futuristic fantasy and more of a growing reality, as the barriers to innovation have truly plummeted.

The flip side of this story, however, is a potential looming challenge: the race to build out infrastructure, especially in areas like data centers and AI hardware, is provisioning the supply side to potentially exceed demand in substantial cases. If the demand for these resources doesn’t catch up with the associated economic investments and buildout pace, we could face a significant "overbuild" scenario, leading to excess capacity. As it stands today, this oversupply could become a considerable problem. Deeper-pocketed investors, such as the hyperscalers (like Amazon Web Services, Google Cloud, Microsoft Azure) or the emergent "neocloud" providers, might be able to weather such a storm due to their immense financial reserves. But for many smaller players in the infrastructure space, this potential overcapacity could result in significant financial whiplash, making it difficult to achieve the necessary utilization and returns on their investments.

When it comes to deep tech, what other areas besides AI do you think are really interesting right now, and where do you see the most significant opportunities?

Viswanathan: While AI unquestionably holds the spotlight as the hottest deep tech sector currently, it’s vital to recognize that the deep tech universe is vast and encompasses many other areas attracting significant attention and offering profound opportunities. Beyond AI, biotech, for instance, is experiencing a renaissance, driven by technological advancements.

At Celesta, we broadly categorize these opportunities into three major "sandboxes" or areas of innovation:

The first sandbox encompasses hardware systems, intelligence infrastructure, and data center AI. This goes beyond just AI chips to include the entire physical hardware infrastructure that underpins modern computing. Think about advanced semiconductor design and manufacturing, next-generation networking technologies, specialized processors for specific AI workloads (e.g., edge AI), quantum computing research and development, novel materials science for more efficient devices, and sophisticated robotics that can operate in complex environments. This foundational layer is crucial because the advancements in software and applications are ultimately constrained by the capabilities of the underlying hardware. Innovations here can unlock entirely new paradigms for computation and data processing, impacting everything from enterprise data centers to personal devices and autonomous systems.

The second area is the software layer, extending far beyond large language models (LLMs) or frontier models. This includes the development of advanced AI tools, operating systems designed for AI workloads, and a myriad of applications and services built on top of these intelligent systems. We’re seeing an explosion of activity in key vertical sectors leveraging AI. For example, in healthcare, AI-assisted radiology is transforming diagnostics, and AI is accelerating drug discovery and personalized medicine. In retail, AI is revolutionizing personalization, supply chain optimization, and logistics. Financial services are benefiting from AI-powered fraud detection, algorithmic trading, and enhanced payment infrastructure. Supply chain management is becoming more efficient through predictive analytics and automation. These are not merely incremental improvements but profound shifts in how industries operate, driven by highly intelligent software systems that can analyze vast datasets, make predictions, and automate complex tasks. This segment focuses on translating raw AI power into tangible, industry-specific solutions that drive efficiency, innovation, and competitive advantage.

And the third critical area is how biology is being profoundly influenced and enhanced by deep tech. This involves the convergence of advanced engineering, computation, and biological sciences. We’re looking at cutting-edge robotics for laboratory automation and advanced surgical capabilities, revolutionizing medical procedures. Diagnostics are undergoing a massive transformation with miniaturization and the development of highly sensitive, rapid testing systems. The manufacturing of complex medical systems, such as advanced CT scan and X-ray machines, and sophisticated radiology equipment like ultrasounds, is being enhanced with integrated AI and cloud connectivity. Furthermore, areas like genomics, synthetic biology, bioinformatics, and the application of AI in understanding complex biological processes (e.g., protein folding, gene editing) are opening doors to new therapeutic approaches, disease prevention, and even sustainable bio-manufacturing. This intersection of deep tech and biology promises to redefine human health, agriculture, and environmental sustainability.

How do you think these prolonged hold periods for private companies are reshaping behavior for LPs, GPs, and founders?

Viswanathan: The trend of prolonged hold periods for private companies has indeed introduced significant shifts in behavior across the entire venture ecosystem. Everyone involved, from Limited Partners (LPs) to General Partners (GPs) and founders, is grappling with the implications. The central theme here is the growing demand for DPI, or distributed-to-paid-in capital. As a popular saying goes, "DPI is the new IRR" (Internal Rate of Return). This reflects a heightened impatience for actual cash distributions back to investors, rather than just paper gains. Consequently, secondary transactions have become an increasingly vital component of the market-clearing mechanism that must occur to provide this liquidity.

Given the remarkable performance of the public markets over the past seven to ten years—and the associated liquidity and returns that investors have enjoyed there—there’s a greater impatience among large institutional investors to see traditional venture capital timelines for exits and distributions shrink. The most effective way to achieve this accelerated liquidity without forcing premature public listings or fire-sale M&A is to leverage secondary markets. This translates into a greater desire from LPs for earlier distributions, more opportunities for GPs to provide liquidity to their investors or older funds, and a general increase in market impatience. In a way, this is a healthy development because, fundamentally, venture capital is about returning money to investors.

Now, the flip side of this scenario is that the public markets have experienced phenomenal growth. It’s certainly arguable whether that trajectory will continue indefinitely. As the public market’s attractiveness potentially moderates, we might see investors shift their focus back towards early-stage, illiquid asset classes like venture capital as a source of differentiated returns. However, for the time being, the strong performance of public markets is actually contributing to a healthy expectation for distributions from private market investments. This pressure ensures that even as companies stay private longer, there are still mechanisms, like secondaries, to provide necessary liquidity and keep the capital flowing.

What do deep tech companies uniquely need to focus on in this evolving environment to succeed and attract continued investment?

Viswanathan: In this evolving and sometimes frothy environment, deep tech companies, perhaps more than ever, need to get "back to basics." It’s simply not sustainable to continue building solely on the promise of an exciting core technology, even in a sector as hot as AI, if you cannot demonstrate fundamental business growth. This means deep tech companies must unequivocally show robust revenue growth, evidence of improving margin growth, and a clear path to scale. The core technology itself, no matter how groundbreaking, is necessary but ultimately not sufficient for long-term success and continued investment. It must be paired with strong technology adoption supported by a tangible and expanding revenue base.

Consider the recent history of the SaaS market as a cautionary tale and a valuable comparison. During peak market exuberance, companies were often valued at 20 times their Annual Recurring Revenue (ARR). Today, you’d be fortunate to achieve a valuation of 5 to 7 times ARR in a good market. This dramatic re-evaluation underscores the market’s shift from valuing pure potential to demanding proven financial performance. We are already beginning to see, and will continue to see, a similar dynamic play out in the AI sector. Investors are increasingly going to demand: "Show me consistent revenue growth. Show me tangible profitability growth."

This means deep tech founders need to be intensely focused not just on their scientific breakthroughs or engineering prowess, but equally on market fit, customer acquisition, monetization strategies, and operational efficiency. The ability to translate complex technology into products and services that generate significant and sustainable economic value will be paramount. Those deep tech companies that can demonstrate a clear trajectory from innovation to scalable, profitable business models will be the ones that continue to attract the deepest pockets and ultimately define the future of the industry. The market is maturing, and with that maturity comes a greater emphasis on fundamental business health alongside technological superiority.

The insights from Sriram Viswanathan paint a vivid picture of a deep tech investment landscape in transition. While the allure of groundbreaking innovation remains, the mechanisms for financing and exiting these ventures are undergoing profound changes. The secondary market, once a niche player, has transformed into a critical artery for liquidity, driven by an AI boom, extended private company lifecycles, and LP demands for earlier distributions. Deep tech companies, in turn, are being pushed to balance their revolutionary potential with a renewed focus on fundamental business metrics—revenue, margins, and scalability. This evolving ecosystem promises continued dynamism, challenging both investors and founders to adapt to a world where technological vision must be inextricably linked with financial discipline.