In the series WeCrashed, there is a scene where WeWork founder Adam Neumann demands a generous credit facility from his banker. Based on his actual bank balance, the banker offers a $10K limit.
Adam, being Adam, says, “I think I’m going to need a bit more... $50 million.” “Google me,” he adds.
In the next scene, he is sitting with Jamie Dimon, who notes his UHNWI status: “In your case the challenge is that your business has received billions... but your primary challenge, Adam, is liquidity.”
Adam nods and then delivers one of the most iconic lines in the series, in a voice that has stuck with me ever since: “Make me liquid, Jamie. Make me liquid.”
I think about that line a lot when I look at venture right now. For four years, our industry has been saying exactly the same thing to the liquidity gods. Plenty of TVPI. Plenty of paper marks and impressive decks. Far less DPI.
In 2026, the gods finally answered.
The catch is that they answered a very short list.
Click the image above to open the DocSend file, a 180-page deep dive into venture and startup dynamics.
The quarter in one word = Liquidity
In the Q1 edition, I wrote that concentration is now the whole game. Concentrated exits, concentrated fundraising, concentrated LP capital, AI winning on every dimension.
That did not change in Q2. It intensified.
But something else moved to the front of the queue, and it is the thing the entire asset class has been waiting on since 2022. So this edition leads with liquidity, because that is where the most staggering change happened.
The headline numbers:
🔹 Venture exits reached $2+ trillion in H1 2026, more than the past decade combined
🔹 U.S. H1 deal value already exceeds the full-year 2025 figure by nearly 30%
🔹 Deals over $50M now make up an unprecedentedly high share of total deal value, roughly nine of every ten dollars
🔹 AI now makes up 86% of the total startup fundraising deal value in the U.S.
Read those four lines quickly, and it sounds like the best year venture has ever had.
Read them slowly and the picture changes.
🎙️ Special thanks to TheOnePoint Podcast's guests
All focused on the new ”Positive Energy” series I’m recording, covering AI + our physical world.
Anurag Kamal (ElectricFish), Topic - The Gas Station is Becoming a Power Plant (🎧 Listen here)
Arvin Ganesan (Fourth Power), Topic - The Battery That Runs at 2,400°C (🎧 Listen here)
Dong-Su Kim (LG Technology Ventures), Topic - The Everything Transition (🎧 Listen here)
Jan Lozek (Future Energy Ventures), Topic - Energy Transition (🎧 Listen here)
Juergen Mayerhofer (enspired), Topic - Monetizing Flexibility, Fully Automated (🎧 Listen here)
Sanjay Aggarwal (F-Prime), Topic - Robotics Boom (🎧 Listen here)
Scott Wharton (Tandem PV), Topic - Solar’s Next Layer (🎧 Listen here)
Sean Kelly (Amperon), Topic - Forecasting the Grid, AI-Native (🎧 Listen here)
Names listed alphabetically.
The list, not the market
Venture exits reached $2.2 trillion in H1, more than the past decade combined. But the value is extraordinarily concentrated: 16 IPOs valued at $1 billion or more accounted for 81.2% of total exit value. SpaceX alone drove much of that (check out the special section on SpaceX IPO), and it will probably impact a broader surface of investors because of secondary activity over the decade+, but we can’t say the same for the other two prominent startups, Anthropic and OpenAI, which are IPOing soon. LPs’ DPI in 2026 is essentially a function of whether they owned one name. So a handful of liquidity events are generating the headlines (SpaceX and Cerebras IPO in Q2, Google’s Wiz acquisition in Q1, SpaceX’s acquisition of xAI and Cursor in H1) and driving the numbers, and a handful of VCs-LPs sit in the benefit zone. If you are a VC or an LP outside it, you can keep chanting “Make me liquid” and wait for your turn.
The two marquee listings tell it best. Cerebras (checkout special section on Cerebras IPO) priced at $185, opened at $350 on a book reportedly 20x oversubscribed, and later traded below its offer. Its cap table records the humble history: Series A at $0.85, a Series F markdown of 47%, a CFIUS block and later withdrawal, and a refiling only once revenue hit $510M and the company printed $237.8M of net income. The window didn’t open for Cerebras - it rebuilt itself until it could force one.
SpaceX (checkout the special section on SpaceX IPO) priced at a fixed $135 rather than a range, raised ~$75B, closed day one at a $2.1T cap, peaked on day three at $2.64T, and by mid-July traded below its offer price at ~$1.63T. On exit value alone it eclipses every VC-backed IPO by 17x and was the largest IPO of all time. But roughly $1T of market value erased in five weeks, with short interest near 30% of float, all ahead of 1+ billion shares unlocking in August. Every structural mechanism, tiered lockup ladder, index fast-entry, Musk’s 366-day hold, existed to manage one topic: $2T listing has no natural marginal buyer. For anyone underwriting the coming AI IPO wave, the offer price is not the mark that matters.
And then there is this different side of exits. Two smaller stories worth watching sit at opposite ends of the same question: can pre-revenue hard tech hold a public listing? In ClimateTech, ten companies raised a record $11.62bn in 2026, led by Fervo’s $1.89bn IPO on $138K of revenue, underwritten not on sales but on contracted offtake and the AI power build-out. In Quantum, $5.7B of exits arrived almost entirely through SPACs, the same route that took IonQ, Rigetti and D-Wave public in 2021 before steep drawdowns. The distinction is the discipline. Companies with binding contracts and physical construction are clearing the IPO process; companies with a thesis are taking the SPAC route. One reopens a durable window for hard tech, the other reruns a playbook the market has already repriced once.
Also, the window opening is not the same as the window paying. Twelve of the last twenty notable IPOs trade below issue: Gemini -84.8%, BitGo -71.2%, Via -60.6%, Klarna -49.4%, Figma -45.2%. The winners cluster tightly in compute infrastructure and a narrow band of fintech and digital health: CoreWeave +148.9%, Hinge Health +159.4%, Circle +102.0%.
M&A re-levelled on price, not throughput numbers. Global venture-backed M&A hit ~$115B in Q2, the highest since Q1 2021, but company count has been flat at 570-620 since 2024. In the US, disclosed acquisition value of ~$120B YTD sits above the 2021 peak on 55-80 deals versus ~205 in 2021. This matters more than the IPO spike, because M&A is the route available to the thousands of companies that will never list, and a flat count means the startup mid-market backlog is not clearing. Under-$10M companies fell from 59% to 42% of acquisitions while $100M+ raisers doubled from 4% to 8-10%. But the bigger point is most M&A startup deals (80%) have raised less than $50M.
Secondaries are still the release valve. The top five names have taken a rising share of secondary market value, climbing from ~44% (Q1 2023) to ~57% (Q4 2025) before easing to ~45% in Q1 2026, and the top twenty consistently account for roughly 80-88% of all value traded. That is a market with real depth in about a dozen companies and almost none beyond them. Direct transfers collapsed from ~65% of secondary deals (Q3 2025) to ~29% (Q1 2026), while single-layer SPVs rose to ~45% and multi-layer SPVs jumped from ~4% to ~23%, a near-sixfold increase in one quarter. I covered special section on “SPV Saga” and all it relates to in the current market including - restrictions, fraud, education, and much more. Median trade price is back to 0% against last round, but the 25th percentile still clears at -32% and 2021-vintage names at -60%. Meanwhile fund-led secondaries grew 10x to $1,029M and from 9% to 44% of sanctioned volume. That is GPs, not employees, manufacturing distributions from unexited positions.
Project this liquidity of rather open exit environment forward and the arithmetic stays humbling: $165B of insider liquidity from $4.7T of listings is a 3.5% conversion rate, with only $14B landing in 2026 and the peak quarter not arriving until Q3 2027. The interesting question is where that capital goes next, and the honest answer is that a large share comes back into this asset class.
The bottom line. Liquidity is returning to venture. Though it is returning to a list. The 2026 exit boom is a handful of names clearing enormous value while listing counts run at a fifth of 2021, mid-market M&A stays flat, and secondaries deepen only for the assets which are investors’ favorites. Across 25 years, a few companies a year reach the threshold where venture math works. That number did not change in H1. Only the size of the ones that did.
So chant it if you like. Just know that the gods are answering a very short list.
Make me liquid!
Three things I changed in this edition
This is the eleventh edition of the Big Book, and I wanted to push the format.
1. Special sections and deep-dives.
Some moments this quarter deserved more than a chart and a caption. So this edition carries dedicated breakdowns on the SpaceX IPO, the Cerebras IPO, the Secondaries SPV Saga, and the ClimateTech pre-revenue IPO dilemma.
2. A lot more India.
Until now, this report has been largely a U.S. and Europe story. The calls from the ecosystem to look beyond those two have been getting louder, and honestly, the India data this quarter earns the space.
3. More nuggets and ecosystem voices.
I have expanded the nuggets sections considerably, bringing in more original and insightful thinking on the topics we tend to overlook in venture but that quietly matter a great deal. New lines of thought, from people worth hearing.
Enjoy reading the 180-page bundle of insights, thoughts, data, and more.
Rohit Yadav
Author, The Big Book of VC