We’re back with another episode of 20VC x SaaStr, and this week we’re diving into the deals and market dynamics that are reshaping venture capital as we know it:
- Cursor just raised $2.3 billion at a $29.3 billion valuation with less than 100 or so employees. Stripe is doing tender offers at all-time highs and showing no signs of going public.
- Oracle’s credit default swaps are screaming warnings about AI infrastructure risk.
- And the entire venture industry is making a concentrated bet on just 4-5 companies that represent 40% of all deployed capital.
Harry Stebbings, Rory O’Driscoll, Jason Lemkin and special guest Tomasz Tunguz of Theory Ventures break down what’s really happening in venture right now—and why the next few years could either validate the biggest bull case in history or expose the most correlated risk we’ve ever seen.
Key Takeaways:
The Cursor Phenomenon: A New Paradigm for Software
Cursor raised $2.3B at a $29.3B valuation with only ~30 employees, representing a fundamental shift in what’s possible in software. The product has achieved near-100% penetration among developers, moving from “productivity boost” to “default and necessary.” With 100-200M developers globally willing to pay $400-500/month, the TAM could approach $1 trillion annually. Revenue multiples become less relevant when considering 30x fewer employees than traditional SaaS and minimal ESOP dilution. The big question isn’t growth—it’s whether they can push into enterprise and overcome the 50% gross account retention typical of vibe coding companies.
The Moat Question: When Does the Fat Congeal?
As agentic coding improvements asymptote, switching costs will increase dramatically through memory, customization, and enterprise ELAs. Market share will likely consolidate to Cursor (40-60%), Microsoft/GitHub (20%), and Anthropic’s Claude Code (20%). However, the pace of innovation remains so rapid that new entrants could still disrupt—we may not have reached the “congealing” point yet. The key risk: if you can copy prompts between systems (as Jason demonstrated with Salesforce Agent Force), moats are lower than they appear.
The Coming Price War Nobody’s Talking About
Current AI pricing represents BLS-style deflation (more tokens for same price), not destructive deflation. But true price wars could emerge if #3-5 players decide to compete on price to gain share. The difference from traditional SaaS: portability of prompts and abstraction layers like Iceberg in data could enable rapid switching. Warning sign: if mid-market or enterprise can easily abstract their AI layer, pricing power evaporates. The counterpoint: integration count remains the #1 predictor of retention—if you’re just “a database of prompts,” you’re vulnerable.
Why the Best Companies May Never IPO
The “access premium” has inverted the traditional illiquidity discount—private companies now command 20-30% premiums versus public comparables. Transaction costs tell the story: $1M for a late-stage private round versus $25-30M (6-7% of raise) for an IPO. With quarterly tender offers and secondary markets, the top 30-50 names have created a “new public market” that’s still private. Stripe at all-time highs of $41/share proves the model works—why pay $30M in transaction costs when private capital is cheaper? IPOs will increasingly be for “A-minus” companies (names 50-150) that can’t sustain quarterly tender offers at scale.
The Retail Tsunami Is Coming
Venture capital hit $184B in 2024, matching 2021’s peak—but half went to just 4 companies (OpenAI, Anthropic, xAI, SpaceX). For the rest of the market, deployment was at 2020 levels, suggesting a bifurcated environment. Goldman buying Industry Ventures signals Wall Street’s bet on retail flowing into private markets through ETFs and fund-of-funds structures. The industry will likely reach $500B by 2030, driven by retail demand and concentration in mega-names. Critical risk: long duration between capital deployment and returns means retail could flood in before discovering actual performance.
The Oracle Credit Default Swap Warning
Oracle’s credit default swaps tripled versus peers, signaling market concern about AI infrastructure debt loads. The company’s entire market cap gain from the OpenAI deal has been unwound. This represents early-stage risk repricing as capex for data centers moves from $500B to $800B+ annually. The market is starting to perceive increasing risk in these massive contracts, even as hyperscalers remain cash-flow positive.
The 900 Unicorn Problem
There are now 900 unicorns that will likely never IPO or achieve meaningful exits. The solution: a middle-market secondaries business trading names 50-200 at market-clearing prices. PE will continue picking off public software companies (took 12% private in 2022 alone). VCs will increasingly “dollar cost average” out of positions across multiple private rounds rather than waiting for IPO liquidity. This creates a PE-style model but with partial sales across rounds instead of full ownership transfers.
The Late-Stage Business Paradox
“The late-stage business is either the best business in the world or the worst business in the world, and there’s nothing you can do to determine which it is.” When prices go up, putting in $100M and watching it go to $200M with no effort feels incredible. The key to success: being a ruthless trader, not a “ride your winners” long-term holder. Critical difference from public markets: liquidity only exists on the upside—when things go wrong, you’re stuck. 15% of Q1 2024’s newly minted unicorns had already raised step-ups by Q3, showing unprecedented velocity.
Why Venture’s Concentration Risk Is Unprecedented
Nvidia’s customer concentration is 10x worse than Lucent’s during the dot-com era. Two customers represent 40%+ of revenues; four customers are 50%+. The entire venture returns question now reduces to: will the 4-5 companies with 40% of deployed capital succeed? If OpenAI/Anthropic/xAI deliver, pooled returns can swamp 40+ failed unicorns. This creates a “singularity bet”—everyone’s returns are correlated to the same small set of outcomes.
The Legal Tech Battle: Distribution vs. Product
GC.ai raised at $550M post (Rory led), competing against Harvey at $8B and Lorra at $2B. The market is fracturing between corporate law firms (Harvey) versus in-house legal teams (GC.ai) versus SMB (Lorra). Customer love and adoption matter more than VC kingmaking—corporate buyers don’t care who invested. Entry price becomes crucial when TAM is unclear; execution matters most when TAM is massive. Cash efficiency is the new safety valve—companies that can’t burn through their last round have optionality if markets turn.
Quotable Moments
On Cursor’s Valuation: “Entry price counts when TAM is unclear. Winning is the only thing that counts when TAM is huge. If you’re not seeing massive TAM expansion, there’s just no point in even playing as VCs.” —Tomasz Tunguz
“This is the way we code now. I don’t know anybody not using cursor. We’re going to approach 100% penetration per developer at $5,000-6,000 per year. Do the math: 100-200 million developers times $5,000. That’s a trillion dollars.” —Jason Lemkin
“Can I see a 3x from here? You don’t have a lot of ESOP dilution because total employee count is 30. You don’t have the capex dilution you see within foundation models. Does it go public? You have massive revenue growth, increasing margin—go bull market bet, classic bull market bet.” —Rory O’Driscoll
On Product Stickiness and Moats: “As the models improve in performance dramatically, people switch. But as improvements in coding start to asymptote, I’m going to stay where I am because there’s memory and it remembers how I program. Agentic coding is no longer on this extremely steep improvement path.” —Tomasz Tunguz
“I took the prompt from another AI agent we trained for months and gave it to Salesforce Agent Force. We iterated for about a day and it worked just as well. Don’t overestimate your moats today. They’re there, but they’re lower than you think.” —Jason Lemkin
“I think we’re at a point where if there’s some wobble, the magnitude of the correction will be fast and brutal. Everyone knows we’re at the red line—we’re going 1,000 mph in a car designed to go 900.” —Tomasz Tunguz
On the Future of IPOs: “Why would you pay $25-30 million in transaction costs to raise a round of capital? It’s like getting a million dollar mortgage and having to pay $150,000 in legal fees. There used to be an illiquidity discount. Now there’s an access premium.” —Tomasz Tunguz
“The late-stage business is either the best business in the world or the worst business in the world, and there’s nothing you can do to determine which it is. When prices go up, putting in $100 million and having it go to $200 million with no effort—that feels as good as life gets.” —Rory O’Driscoll
“I think venture moves in the PE direction. You buy and hold for 3-5 years, package it up for the next person. You find your Nth fund returner, your Nth decacorn, and you dollar cost average your way out over multiple rounds rather than waiting 15-20 years for liquidity.” —Tomasz Tunguz
On Market Dynamics: “If the concentration works, it’s all going to be fine and the industry will keep chugging. The question really resolves to: will the four or five companies that constitute 40% of deployed capital be good? If OpenAI, Anthropic, xAI deliver the returns everyone hopes, you’ve taken half the risk off the table.” —Rory O’Driscoll
“We’re building more data centers than offices. Data centers are the new cities. Why shouldn’t we have three to four 30-40% corrections along the way? We’ve seen it before in our investing lifetimes.” —Jason Lemkin
“This is where we find ourselves. The bet is on and the bet is singular and utterly correlated. But holy shit, the fear of public markets and impending doom has not reached early stage yet. YC batch companies are raising at $50 million post standard.” —Harry Stebbings
On the Reality of Returns: “In the end, the only thing that matters is returns. You may take longer to figure it out, but when people lose money, they figure it out. The runway at which things can continue is very long, but eventually the only thing that matters is returns.” —Rory O’Driscoll
“All this is great until you’ve had to go in a room and look people in the eye and say you’ve lost them money. You’ll take these big retail funds, you’ll have fun putting all the money out, and then you’ll realize you’ve locked in retail investors to a subpar return for a decade. That will not be fun.” —Rory O’Driscoll
On Investment Philosophy: “I remember looking at churn in SaaS companies and the number one predictor of retention was the number of integrations. If it’s easy to rip it out, you will rip it out if it’s cheaper. If it’s hard to rip it out, you won’t bother.” —Rory O’Driscoll
“People rarely sell stocks because they think they’re going to go up. In the 10 seconds it took to run that decision by Peter Thiel, he said ‘yeah, you should sell that stock.’ It’s not that he’s unloading his position—it’s not like when he was dumping Facebook. But it tells you something.” —Jason Lemkin
Want more tactical, no-BS insights on scaling B2B SaaS? Subscribe to SaaStr and join 500,000+ founders, executives, and VCs getting our weekly newsletter.
