Tag Archives: technology

This past quarter 3,336 companies raised over $267 billion which eclipsed every prior full year, save two of them (2021 and 2025), underscoring the extraordinary enthusiasm among venture capitalists for anything AI. Notwithstanding the frothy valuations and capital intensity of AI platform companies, one of the pleasant knock-on effects might be the surge in new company formation. While AI adoption may also foretell dramatic job losses, it does strongly suggest that with the proliferation of these capabilities it will be easier to be an entrepreneur. 

New Business Applications

Source: Apollo

Just over the past twelve months the number of new business applications spiked from ~100,000 per month to ~140,000 per month (right hand axis, above). Notably, the onset of the pandemic and its short-lived recession six years ago pushed many to become entrepreneurial out of necessity. The forces are different now and the impact less dire and immediate, but the result may be similar.

Obviously, it is a complicated, confusing time, made more so by conflicting social media posts almost hourly from the Administration and the on-again, off-again war with Iran. While making short-term investment decisions is fraught with extreme risks now, there are important signals to watch. Unemployment rates are manageable and in line with a good economy. Inflation is only slightly elevated, even in light of the body blows delivered to the economy over the past year. Somewhat surprising, the NASDAQ Index is trading at levels that are below historic averages, which for many may be counterintuitive given frothy private market valuation multiples.

NASDAQ 100 Forward P/E Ratio

Source: Citadel Securities

The employment picture is more nuanced, exacerbated by the hostility directed at immigration. Currently the level of job vacancies is largely balanced with the level of unemployment. Early in the pandemic the number of unfilled jobs spiked labor inflation rates. The prospect of job losses, perhaps due to impact of AI-driven replacement capabilities, may serve to mute the impact on inflation. In fact, according to the Bureau of Labor Statistics the U.S. economy only added a net of 181,000 jobs in all of 2025. An analysis by Fortune estimated that the U.S. healthcare sector alone created 693,000 new jobs in 2025, and that were it not for healthcare, overall job growth would have been negative.

Vacancy to Unemployment Ratio

Source: Bureau of Labor Statistics, BofA Global Research

More specifically, the labor market for IT positions has turned decidedly negative. In March employment in the technology sector declined by 24% from March 2025, according to data from Challenger, Gray & Christmas, with an overall net IT job loss of 52,000 in 1Q26. Bureau of Labor Statistics data had IT unemployment at 3.9% at the end of 1Q26, up from 3.1% in 1Q25. Bank of America data highlights the rotation from payroll costs to increased expenditures on technology among U.S. small businesses – technology replacing labor.

 Payments to Payroll vs. Technology Services

Source: Bank of America (percent change, 3-month rolling average)

The level of venture capital investment this past quarter was historic by any measure. The five largest investments totaled $195.6 billion of the $267.2 billion (73% of the total) and yet was only 0.1% of the deals closed. While one might credibly argue that the $122 billion OpenAI round is not a venture capital deal, undeniably the AI companies simply dominated this past quarter, capturing nearly 89% of all invested capital and represented 43% of all companies. Notwithstanding a level of “AI-washing” with how companies now describe themselves, Pitchbook estimated that in 1Q26 AI companies accounted for 45% of the $9.5 trillion value held by venture capital funds.

This level of activity somewhat obfuscates the continued decline in the absolute number of investments made at each stage, although the proportion of larger and later stage rounds continued to increase. This rotation highlights the concentration of capital behind fewer companies as well as the dramatic increase in late-stage company valuations. Across all sectors late-stage rounds averaged $47.6 million ($44.0 billion invested across 925 deals) in 1Q26 as compared to $24.8 million in 4Q25. 

Quarterly Venture Capital Investment Activity

Source: Pitchbook

This is not just a U.S. phenomenon although the U.S. market carries much of the water. Globally, 1Q26 venture capital investment activity touched $300 billion in nearly 6,000 companies, according to Crunchbase data. Of this amount, $242 billion was invested in AI companies.

Global Venture Capital Investment Activity

Source: Crunchbase

An obvious implication of this level of concentrated investment activity is the restocking of the unicorn herd. In 1Q26 another 66 unicorns were minted bringing the total to 912 companies valued at $5.8 trillion based on the post-money valuations of the last round according to Pitchbook. While many choose to stay private and independent longer, many other scaled companies continue to struggle to exit.

This shines a bright light on a cruel irony: rarely has such a new technology generated so much investor interest and created so many valuable companies and yet exit activity remains frustratingly elusive. Notwithstanding that 1Q26 looks like a blow-out quarter for exits, the $347 billion includes the $250 billion combination of SpaceX with xAI, which was not really a liquidity event. Setting that transaction aside, the $97 billion of remaining exit activity continues the modest recovery seen over the last several quarters in total exits. The average transaction value (ex-xAI) in 1Q26 was nearly $300 million, which is appreciably greater than the $230 million in 4Q25.

Quarterly Exit Activity

Source: Pitchbook

Analysts are calling for the SpaceX IPO this year to be valued over $1.5 trillion (~55x 2026 revenues) – crazy for a company that was last valued at $350 billion just over a year ago. This event could well be followed by IPOs at OpenAI and Anthropic, among several other very significant private venture-backed companies. Much could still derail any or all of these liquidity events but there is cautious optimism that investors will start to see dramatic distributions unlocked later this year. Obviously, it will be unevenly felt at first but one should expect to see that fundraising prospects for the long tail of venture capital firms should improve.

Secondary transactions continue to come of age as a possible path to liquidity. Last year secondaries totaled $240 billion across all private asset classes according to Jeffries, which was a six-fold increase from 2015. To provide some context, the broader private capital industry is estimated to be $24 trillion.

Fundraising in 1Q26 totaled $47.8 billion across 172 funds (average fund size $278 million), which compares nicely to the $66.7 billion across 626 funds (average fund size of $107 million) in all of 2025. What this masks is again the extreme concentration of success for just several firms in 1Q26. In fact, just six funds accounted for $36.4 billion or more than three-quarters of the capital raised. Exclusive of these top six funds, the average fund size was $69 million. Of all the funds raised in 1Q26, only 29 were raised by first-time fund managers. Shockingly, the median fund size in 1Q26 was an extremely modest $15.3 million.

As a point of comparison, “hot money” institutional investors in this risk-off environment moved nearly $11 billion just in March alone from junk bond funds into investment grade bonds, a 24-fold increase from February according to JP Morgan.

In the current climate investment returns have been hard to come by. HFR, a leading hedge fund data provider, tallied industry performance in March to be -3.1%, the worst performance for hedge funds since March 2020 (admittedly, April 2026 may show some recovery in performance given the “TACO” phenomenon). For 1Q26, not surprisingly the best performing asset was the New York Mercantile Exchange (NYMEX) Heating Oil Index, which surged 96.3%; the top four performing assets were all oil related. Cocoa was in last place, generating a -45.6% return. Cambridge Associates recently released its most recent venture capital performance data (4Q25), which were 7.8% and 11.6% for early-stage and late-stage, respectively, which does not look too bad in comparison.

Investment activity in the healthcare technology sector continues to be both resilient and encouraging. According to Rock Health, $4.0 billion was invested in 110 companies in 1Q26, suggesting an annual pace of ~$16 billion in approximately 450 companies – or just 13% of the last OpenAI round. The average deal size of $36.7 million is nearly twice that of just two years ago. Although 1Q26 activity was slightly down from 4Q25 ($4.2 billion), it was up smartly from the $3.0 billion invested in 1Q25.

Digital Health Investment Activity

Source: Rock Health

An internal Flare Capital analysis of 1Q26 AI healthcare technology investments based on both Pitchbook and Rock Health data highlights an interesting market development. It certainly appears that early-stage round sizes are materially larger and being done at elevated valuations. Dilution in Seed and Series A rounds coalesced between 18 – 22% (which is obviously not how one should value a company but is an oft-cited benchmark nonetheless). In 1Q26, our data showed that for Seed rounds the average and median pre-money valuations were $20.7 million and $17.1 million, respectively. For Series A and B rounds average and median pre-money valuations spiked significantly to $96.4 million and $58.0 million, and $402.6 million and $459.8 million, respectively.

Arguably, the value creation curve historically has been more asymptotic for healthcare technology companies as it took a few years at the outset to establish product/market fit to confirm that the products or services lowered costs and/or increased revenues and/or improved outcomes. That dynamic seems to have changed as companies are raising more capital, sooner. The curve now appears more linear, up and to the right as product development timelines have compressed dramatically.

In 2023 healthcare became the fastest growing component in the personal consumption expenditure price index (PCE) and is now nearly 17% of the index. As of late 2025, healthcare accounted for more than 50% of the growth in PCE. Coupled with the growth in the healthcare labor force and aging U.S. population, healthcare costs continue to drive overall inflation. It is estimated that by 2030 there will be more Americans older than 65 than younger than 18 for the first time in the country’s history. Fortunately, Goldman Sachs analysts recently concluded that a 70-year-old today as the equivalent cognitive capacity of a 53-year-old in 2000.

All of this points to the extraordinary opportunity technology has to re-architect clinical and administrative workflows throughout healthcare, especially as AI can make the system less expensive, more effective. A recent study by Johns Hopkins estimates that 10% of Medicare Advantage enrollees will disenroll this year as insurers modify plan offerings due to costs, exacerbating issues of access. Already 14% of all ACA members have missed first premium payments in 1Q26. Notwithstanding that many healthcare providers are seeing improved operating margins through deployment of agentic AI technologies, anxiety is high over pending deep cuts to reimbursement rates, particularly for the Medicaid segment in 2027.

This is happening against the backdrop of the fiscal condition of the U.S. economy. Recent Congressional Budget Office estimates the federal deficit will be $1.9 trillion in 2026 or 5.8% of GDP, increasing to 6.7% in 2036. The trailing 50-year average was only 3.8%. The publicly held federal debt is projected to be 120% of GDP, up from 101% in 2026, eclipsing the all-time high of 106% in 1946. This deficit needs to be financed, which will run the risk of crowding out other investment opportunities and/or absorbing significant capital, likely to the detriment of other possible uses.

As Percent of U.S. GDP

Source: Congressional Budget Office

We hope you will join us for our next quarterly Expert Roundtable Series on June 9, 2026 at 12:00pm ET (register here). To stay connected, subscribe here for the latest updates, news, and insights from Flare Capital Partners.

Over the past 100 years there have been 29,078 publicly listed U.S. stocks, of which 4% of them accounted for nearly all the economic wealth created this past century. Less than 100 companies were responsible for half of those gains. Approximately 52% of those companies had negative cumulative returns according to research from Arizona State University.

Those same researchers then studied the 63,785 global stocks between 1990 and 2020 and showed that 55% of U.S. and 57% of non-U.S. stocks generated returns below that of a one-month Treasury bill. In that same period just five companies accounted for a shade over 10% of the $75.7 trillion of the net equity value created. Half of that wealth creation was by 0.25% of the companies; all of it by just 2.4% of the companies during that period.

Issues swirling around the concentration of capital, of wealth, of resources are quite acute now. According to Bloomberg, the top ten U.S. technology founders saw their collective net worth increase by $550 billion to $2.5 trillion in 2025, largely on the coattails of the AI phenomenon. The top ten private companies are now worth $2 trillion.

Globally, hedge funds had their best year in 2025, increasing assets by $628 billion to over $5 trillion, in spite of only generating a 12.8% return. This was rather middling when compared to dozens of other asset classes (NASDAQ was up 20.4%, S&P 500 up 16.4%). Silver was the top performing asset this past year, increasing over 142%; three of the top four assets happen to be metals. Sadly, the poorest investment in 2025 was orange juice, which dropped nearly 59%.

The venture capital industry is navigating the noisy distractions coming from D.C., while coming to terms with the impact of AI during a confusing but reasonably good economic climate. In December inflation was 2.7% and the unemployment rate ticked down to 4.4%. Additionally, labor productivity increased 4.9% in 3Q25 while output jumped 5.4% with hours worked increasing by just only 0.5%. As a point of comparison, annual improvements in labor productivity in the 2010s averaged 1.1%.

Not surprisingly, most of the recent venture capital activity was centered on AI as each sector of the economy scrambles for solutions to drive additional productivity gains. Of the $339.4 billion invested in 16,709 companies in 2025, $222.1 billion was attributed to 5,793 AI companies according to Pitchbook (65% of the capital, 35% of the companies).

Globally, $469 billion was invested according to CB Insights, 48% of which was in AI companies, and goes on to caution that many of the marque venture funds have become highly “concentrated AI funds, not diversified portfolios.” Obviously, AI round sizes were larger than for non-AI companies: $38.3 million versus $10.8 million, respectively, underscoring profound investor exuberance.

S&P 500 Companies’ Real Revenue per Worker

Source: Bank of America US Equity & Quant Strategy, FactSet, Bloomberg

The investment activity over the course of the year saw a notable increase in 4Q25, returning back to the quarterly pace of euphoric 2021. Notwithstanding that the total number of investments continued to modestly decline, the average deal size spiked to $27.1 million, likely reflecting several mega-AI financings. In fact, eight of the eleven rounds greater than $1 billion this past quarter were for AI companies; those eleven companies raised $37.4 billion or 41% of 4Q25’s totals. Anthropic’s $15 billion “venture” investment punctuated this past quarter’s activity.

Venture Capital Investment by Quarter

Source: Pitchbook

Underneath a relatively serene surface are profound challenges to the venture capital model. First and foremost is the significant lack of liquidity. Arguably, liquidity serves as a proxy for the relevance of much of what VCs do: does the inability to sell portfolio companies point to a general apathy or indifference towards those companies? Obviously, some founders may choose to simply stay independent, or the board may determine that the runway for significant additional value-creation is long so best to stay the course. More often than not, though, that is not the case.

An analysis by Pitchbook highlights the extreme state of affairs. Since 2022 net cash flows from limited partners totaled $197 billion through 2Q25. While this does not yet reflect the significant exit activity in 2H25, the venture capital model is predicated on the ability to recycle capital. This does not mean to imply significant impairment but rather it underscores the challenges over the past few years to find attractive buyers coupled with a largely non-existent IPO market. Having said that, though, there are some concerns that companies funded several years ago may be at severe risk of obsolescence given the rapid pace of change.

Net Cash Flow to Limited Partners ($B)

Source: Pitchbook (as of 2Q25)

Notwithstanding the lack of predictable and material distributions back to limited partners, exit activity in 2025 started to recover with $297.6 billion across an estimated 1,635 transactions (average size of $182 million). The year ended with an upswing with 4Q25 exits totaling $93.6 billion across 422 transactions (average size of $222 million) and was the strongest quarter for exits since 4Q21 ($221 billion, 565 transactions, average size of $391 million). IPO activity in 2025 continued to be somewhat underwhelming with only 48 IPOs of venture-backed companies, although the total value was $116.7 billion, both ahead of 2024 amounts of 44 companies and $41.4 billion, respectively.

Another industry concern is that fewer firms are raising much larger funds, exacerbating a “bar belling” of the industry. Fundraising in 2025 continued the anemic downward trend started in 2022 with only $66.1 billion raised by 537 funds (average fund size of $123 million; median fund size was only $26 million). More than $23 billion of the total raised in 2025 (35%) was by funds greater than $1 billion in size. Furthermore, only $6.6 billion was raised by 92 first-time fund managers (average fund size of $72 million), reflecting further concentration among larger established branded firms.

With just under $300 billion of “dry powder,” the venture capital industry certainly has the wherewithal to power through economic turmoil. But even this investment capacity underscores the concentration of capital among fewer investment managers. Nearly 60% of the “dry powder” is held by funds greater than $500 million, and yet those funds represent just under 7% of the number of funds raised over the past four years. Similarly, only ten private equity funds raised 46% of all capital in 2025 according to the Financial Times. A knock-on implication of this concentration might well be fewer start-ups being funded as larger funds tend to make larger, later stage investments.

Venture Capital “Dry Powder” By Fund Size

Source: Pitchbook (as of 1Q25)

The venture capital industry has been referred to as a game of grand slams – wish it were that straightforward. According to MLB.com, in the 2025 season there were 120 grand slams sprinkled across 163,685 at bats (0.07% of each at bat). While the average is markedly more attractive in venture capital, failure is unfortunately a reality. Data from S&P Global Market Intelligence shows that bankruptcies in 2025 were at a 15-year high with 785 filings. A review of failure rates by stage reveals two interesting observations: failure rates decrease as companies scale and the top decile of companies ranked by valuation at each round fail less often.

Failure Rates by Stage

Source: Pitchbook

While these data might suggest a reasonably predictable progression up and to the right, occasionally interim financings are more “complicated.” If milestones are not hit, but existing investors remain engaged and enthusiastic, it is not uncommon for the subsequent round to be at a lower valuation and/or with economic terms to attract (i.e., induce) additional capital. This is a dynamic that is closely watched to gauge the health of the venture ecosystem. Fortunately, the percentage of down and flat rounds has improved in 2025 with less than 20% as compared to mid-20% in 2023, but not as good as the rocking 2021 – 2022 timeframe when that was in the mid-teens.

Percentage of Down and Flat Rounds

Source: Pitchbook

It is estimated that today there is $261 trillion of investable assets globally. In fact, since the 1990s as a percentage of global GDP, assets have increased from 75% to approximately 200% (~120% equities, ~80% bonds/other). According to the National Venture Capital Association, at the end of 2024 there were 3,111 firms managing $1.25 trillion of venture assets in the U.S., a relatively modest portion of the global market. Recently released data  from Cambridge Associates, a leading institutional advisory firm, presented venture capital returns data as of June 2025 that highlight the recent “air pocket” these firms are navigating. As of 2Q25, the one-, three-, five-, ten-, and 15-year returns were 11.4%, 0.1%, 15.0%, 13.1%, and 15.1%, respectively.  

Limited partners need to stay convinced that commitments to venture capital funds will drive enhanced returns while also diversifying overall portfolios. The conundrum now is whether the incremental volatility is justified in light of significantly reduced liquidity. A comparison of various asset class 15-year returns by Pitchbook highlights the dilemma confronting investors today.

Returns by Asset Class

Source: Pitchbook (as of December 2024)  

Fortunately for investors seeking direct exposure to the AI craze, venture capital remains the most direct, efficient way to do that, as it has for every other transformative wave of innovation. As these new AI capabilities demonstrate attributed enduring economic value creation expect to see marked increase in liquidity (arguably there are early signals in 4Q25 data). In addition to the improved productivity metrics, another indicator that technology may be profoundly replacing labor is evident in the share of economic output that is captured by workers’ wages, which has never been as low as it is now. What was once in the mid-60% range has dropped now to below 54%, suggesting we have become meaningfully less reliant on humans.

That may be the more troubling development…

Share of U.S. GDP for Labor

Source: Bloomberg

We hope you will join us for our next quarterly Expert Roundtable Series on March 3, 2026 at 12:00pm ET (register here). To stay connected, subscribe here for the latest updates, news, and insights from Flare Capital Partners.

The private credit market shuddered this past week with the disclosure of significant portfolio issues tied to the fraud and bankruptcies at First Brands and Tricolor Auto Group. Fears of further credit concerns saw 74 regional banks lose over $100 billion in market value as chatter of a possible contagion emerged, according to an analysis by Bloomberg.

Given the magnitude of capital required for the announced AI infrastructure buildout, as much as $320 billion in 2025, robust and accessible capital markets will be essential. J.P. Morgan now estimates that there is $1.2 trillion of debt linked to AI issuers, accounting for 14% of the high-grade debt market, making AI the largest sector.

Against the backdrop of the government shutdown, the state of the labor force has several flashing warning signs that provide some cautious insights into where the U.S. economy might be heading. For the first time since the outset of the pandemic, the number of unemployed workers is now greater than the number of open positions (“V/U ratio”). Over the past 25 years, a decline in this ratio has been correlated to a recession.

Number of Unemployed Workers

Source: Bank of America, Haver Analytics

Additionally, declines in temporary service jobs often indicate the onset of a recession. With approximately 2.5 million temporary service workers, these tend to be the first people let go. There are now hints of a significant correction looming as the level of temporary employment has been dropping for over a year, according to an analysis by Charles Schwab. The absence of government labor data further heightens investor anxiety.

Changes in Temporary Services Labor Force

Source: Charles Schwab

With the government shut down, which Bank of America estimates is costing the U.S. economy $15 billion per week, investors have focused greater scrutiny on other economic indicators. Goldman Sachs has developed its index of Current Economic Activity which is a composite of several foundational pillars of the U.S. economy. Notably, this index has declined over the past several months, with particular weakness in the Manufacturing sector. “Liberation Day” this past April was a particularly bad month, which further underscores lingering investor concerns about the haphazard application of the U.S. tariff strategy.

Indicator of Current Economic Activity

Source: Goldman Sachs Research

But valuations of financial assets and commodities have been on a tear. The best performing asset in 3Q25 was Silver, which increased over 29%, followed by Coffee, up a hyper-caffeinated 22.2%. Surprisingly, Cocoa declined 27.9% which industry analysts attributed in large measure to a decline in chocolate consumption (GLP-1 impact?). The NASDAQ Index increased by 11.2% while the S&P 500 Index increased by a more pedestrian 7.8% in the quarter. Globally, equity values are now approximately 120% of worldwide GDP, which is converging on an all-time high.

Global Stock Market Capitalization as Percent of Global GDP

Source: Goldman Sachs

Obviously, the artificial intelligence phenomenon is driving profound investor interest in all-things AI. But that excitement is not just contained to AI. Speculation euphoria is driving other sectors as well, perhaps fueled by the enabling capabilities of AI. The crypto sector is now valued at over $4 trillion. Gold has increased 59% just this year alone. According to the American Gaming Association (“AGA”), there was $150 billion bet on spots in 2024, with the AGA indicating that it will be “significantly higher” in 2025.

Business Insider recently estimated that $320 billion will be spent on AI infrastructure this year, which is an outsized contributor to overall domestic GDP growth. Bridgewater calculated that 35% – 45% of GDP growth year-to-date is just due to the technology sector’s investment in AI infrastructure. Absent this investment, GDP growth would likely be running at less than 1.0% in 2025.

Contribution to U.S. GDP Growth

Source: Bridgewater

A Harvard study paints an even more sober or daunting picture, concluding that through the first six months of 2025, 92% of GDP growth was attributed to AI data center development, highlighting the economy’s reliance, and therefor exposure, to the technology infrastructure build-out today.

Another word of caution: much of this infrastructure may have a relatively short shelf life. ChatGPT is less than three years old and now NVIDIA is releasing its next generation of AI chips (Blackwell Ultra and Vera Rubin platforms) which effectively obsoletes chips from several years ago. Chip depreciation is estimated to be as much as half the cost of running data centers (power is only between 2% – 6%), so the path to value capture may be elusive for some participants.

This investment – so far – appears to be justified. Nearly 18 months ago Google reported that 9.7 trillion tokens (a unit of data processed) per month were processed across its products. This past May that number spiked to 480 trillion, an increase of 50x in just twelve months. This increased to 980 trillion just three months later and is now running at 1.3 quadrillion tokens per month.

An analysis from McKinsey & Co from earlier this year offers an often-cited word of caution though. Notwithstanding that U.S. companies are estimated to spend $62 billion on AI solutions this year, 80% of surveyed companies reported “no significant bottom-line impact” with 42% of respondents terminating their AI initiatives.

The rapidly emerging dilemma is now how best to finance all these AI infrastructure investments. Setting aside the scrutiny on “circular deals” with the leading companies investing in each other to purchase one another’s’ products, which has opened a non-trivial systemic risk debate, investors have piled into various technology sector funds (below) and AI entrepreneurs have been embraced by venture capitalists.

Technology Sector Fund Flows

Source: Bank of America Global Economics

This rising AI tide is most definitely lifting all venture capital boats. According to Pitchbook, there was $80.9 billion invested in 3,175 companies in 3Q25. Year-to-date 64% of all venture capital investments were in AI companies, yet this represented less than 38% of all companies, reflecting the extraordinary round sizes of AI financings. There were nine $1.0+ billion AI rounds in 3Q25 alone. Across every stage the median deal size increased year-to-date over 2024 by more than 10%. Pre-seed, Seed, Series A, Series B, Series C, Series D+ median round sizes year-to-date are $0.5 million, $3.8 million, $14.0 million, $32.4 million, $56.5 million, and $100.0 million, respectively.

Quarterly Venture Capital Investment Activity

Source: Pitchbook

Even more dramatically was the increase in median pre-money valuations year-to-date over 2024 levels. For instance, Series C valuations increased from $225.0 million to $307.0 million, and Series D+ spiked from $630.9 million to $838.8 million. Notably, the average Series D+ valuations skyrocketed from $1.8 billion in 2024 to just over $4.0 billion year-to-date, reflecting a handful of extraordinarily large AI infrastructure investments.

This has led to a crowded pasture of unicorns. According to CB Insights, there are now 498 AI unicorns that are valued collectively at $2.7 trillion, 100 of which were founded just since 2023. It is estimated that there are more than 1,300 AI start-ups valued at more than $100 million each. Across all sectors, Pitchbook estimates that there are 828 unicorns with an average post-money valuation of over $3.7 billion. Year-to-date more than $142 billion was invested in 284 unicorns, representing 57% of capital invested but only 2.7% of companies.

Unicorns Created by Quarter

Source: Pitchbook

The question confronting most venture capital investors turns on when the industry will see predictable liquidity (i.e., exits). There are glimmers of hope here. Already year-to-date exit activity is running more than 33% greater than all of 2024 with $204.9 billion across an estimated 1,135 exit transactions. While the average exit value is of modest utility, it is considerably greater year-to-date at more than $180 million when compared to $121 million for all of 2024. Given the concentration of invested capital in a relatively select few companies, one might also expect that a relatively small number of companies will generate the bulk of the investment returns.

To be more specific, a recent analysis by the Financial Times found that just ten unprofitable AI companies saw an increase of over $1 trillion in aggregate valuation over the past twelve months. Those companies include Anthropic, Anysphere, Databricks, Figure AI, OpenAI, Perplexity, Scale AI, Safe Superintelligence, Thinking Machines Lab, and xAI.

Exits drives fundraising. To refill venture capital coffers, limited partners need to see robust distributions. Year-to-date 376 funds were raised that totaled $45.7 billion (average fund size of $122 million), which candidly is quite anemic when compared to 2024 (832 funds and $85.7 billion raised), although 3Q25 was stronger than 1H25 with $19.1 billion raised over 138 funds (average fund size of $138 million). Of the 376 funds raised this year, only 68 were from new firms (average fund size of $71 million), underscoring the challenges to break into the venture capital industry.

The profile of great venture capital funds tends to have returns generated by a small number of portfolio companies, which is mirrored in the overall venture capital industry with 3,111 firms managing 7,969 funds; that is, top decile fund performance is significantly better than the average fund. Carta, a leading private markets analytics firm, recently published an analysis of 1Q25 venture capital fund performance which highlighted this performance distribution by vintage. Notably, the relatively poor performance of the 2021 vintage reflects the dramatic inflow of venture capital that year. This also underscores growing concern about “capital absorption” when too much is invested, too quickly.

Venture Capital Performance by Vintage Year  

Source: Carta

American “exceptionalism” has been bandied about for generations but has been recently called into question. Over the past 15 years, the S&P 500 Index has performed significantly better than the basket of all other countries, and yet approximately 80% of the performance can be attributed to a strong U.S. dollar and valuation multiple expansion. Business fundamentals are a relatively small contributor to the overall performance, and in fact, analysts at GMO conclude that over the past five years the difference in performance between the U.S. and the rest of the world nearly vanishes.  Notably, the bulk of superior performance is reflected in a small number of companies and was generated prior to 2015.

Performance of S&P 500 vs. MSCI World Index (ex-U.S.)

Source: GMO, Compustat, S&P Global, MSCI

Undoubtedly, these new AI capabilities are profoundly exciting and will be unexpectedly disruptive. The question will quickly focus on whether attractive returns can be generated by these investments. Investors will be listening carefully to reports from Corporate America over the next several quarters, given how much capital will be tied up in this infrastructure. We better hope it all works…

And please join us for our next quarterly Expert Roundtable Series webinar on December 9, 2025, at noon ET. And subscribe here to follow other news and insights from Flare Capital Partners.