20VC: Jensen Huang Investigation: What Happens Now | Sequoia Leave London, Close Scout Program and End Evergreen Fund | NASDAQ Loses $2.5TRN in 24 hours | Anthropic targets $1.4TRN IPO valuation

Sequoia Capital

Across the shows: 1 speaker in other episodes, 0 bullish and 1 bearish calls, past 30 days

• Rory O'Driscoll · Scale Venture PartnersNeutral

  Sequoia's Evergreen fund was logical but badly timed, illiquid for LPs, and misaligned with partner economics.

  Rory says holding a diversified portfolio of high-quality VC-backed IPOs was intellectually correct and could compound above the S&P. The fund had three problems: it launched before a drawdown, so a year later stocks fell 30%; LPs have different liquidity needs and want to make their own choices; and partners who have cleared carry can already hold their own stock, so forcing LPs to hold looks like trying too hard.

  “The first problem is they picked a horrible time to do it because they implemented in [2022], which is when stocks actually turned down.”at 5:30

• Jason Lemkin · SaaStrNeutral

  Ending Sequoia's Evergreen fund makes sense; the structure was too complex.

  Jason asks why Sequoia wound down a structure that would have held the dominant brand's winners like Nvidia and Google. He learns LPs could not leave and that the structure encouraged holding long after IPO. He concludes that trying to force LPs to hold was too complex, so getting rid of it to keep things simple makes sense, and Sequoia's partners can still make money.

  “It's just trying too hard, too much complexity. So I think getting rid of it was just, let's keep it simple here, people.”at 7:25

• Rory O'Driscoll · Scale Venture PartnersNeutral

  Sequoia's scout program gave a big advantage for little effort, so closing it surprises him.

  Rory says he never understood the scout program, because it was relatively little lift for a very significant advantage. He was surprised to see it ended, though he gives Sequoia credit for reinventing itself and focusing its resources.

  “I always did think the scout program was relatively little lift for a very significant advantage.”at 11:11

• Jason Lemkin · SaaStrNeutral

  Sequoia's scout program worked better as a secret; closing it reflects managing it was too much.

  Jason says the scout program was cloned by others and worked better when it was secret, when Sequoia was years ahead. He reads its end as the program becoming too flat and lacking energy. Running a scout network with a small partnership and six London partners in different time zones and incentives was too much, and he thinks simpler, conviction-led deployment is the better model.

  “Well, it did get cloned... I think it worked better when it was a secret. They were years ahead of everybody else.”at 11:32

Revolut

• Jason Lemkin · SaaStrBullish

  Revolut looks unassailable with attractive margins and could double or triple.

  Revolut appears unassailable, and unlike the other deals it has more attractive margins. It is going from strength to strength and is the simplest of the three to see doubling or tripling. Lemkin frames it as the lowest-risk way to compound among the options in this batch.

  “And unlike the others, the margins are more attractive, right? And it seems to be going from strength to strength. It's the simplest one I could see doubling or tripling.”at 35:11

• Rory O'Driscoll · Scale Venture PartnersBullish

  Revolut is set to become Europe's biggest bank by taking incumbents' young customers.

  Revolut is mostly money movement and does little lending today, but it is taking business from European retail banks and winning their young customers. Those customers, aged 19 to 25, will want mortgages over the next ten to twenty years, so Revolut is well positioned to become Europe's biggest bank. He describes it as obliterating European incumbents.

  “over the next ten, twenty years, all those customers who are 19 to 25 today are just gonna get older. They're gonna want mortgages. ... They're well positioned to be the biggest bank in Europe.”at 39:33

• Harry Stebbings · 20VCBullish

  Revolut will take unprecedented European market share and could be a 4x from here, key-man risk aside.

  Stebbings expects Revolut to gain market share faster than any historical incumbent, with a product suite that keeps broadening and a deep customer base. He calls it the best asset by far, led by the best founder of his generation, and sees a four-times return from the current price. The downside is key-man risk on the founder. He would compound it for ten years.

  “We will see market share for Revolut unlike we've any... Ever seen for any historical incumbent... This is a four x from here with the downside that we have key man risk on Nick.”at 39:51

Anthropic

Across the shows: 117 speakers in other episodes, 159 bullish and 93 bearish calls, past 30 days

• Rory O'Driscoll · Scale Venture PartnersBullish

  Anthropic's modest $70–78 IPO range is smart; starting low beats starting high.

  Rory calls the recommended $70–78 price relatively low and clever. It is up 20–30% per share from the $54 last private round, but the $1.4T versus $1T headlines come from counting all options in private valuations versus only vested options publicly. Pricing low lets the company generate demand, then raise the range or declare victory at the high end. He says it is far better to start low and go up than to start high and come back.

  “they put a recommended price of 70 to $78 a share, which is relatively low... But it is far better in an IPO to start low and go up than to start high and come back.”at 24:27

• Rory O'Driscoll · Scale Venture PartnersBullish

  Public listing gives Anthropic easy capital; expects a $50B convertible within six to nine months.

  Rory expects Anthropic's CFO to file a convertible within six to nine months and raise a large sum, since public companies can access public capital at scale. He notes Anthropic would have about $200B of cash, more than any other US company, and that if growth continues it will need more capital but can get it. Public status gives it a clear advantage over OpenAI, which is still private and unprofitable.

  “I'm willing to bet six, nine months from now, that CFO will have a convert on file. He'll raise a company of 50 bill.”at 28:03

Zipline International

• Jason Lemkin · SaaStrBullish

  Zipline is a likely winner with better unit economics than Waymo, though the price seems rich.

  Lemkin had dismissed Zipline for years as drone delivery in Africa, but now sees autonomous delivery alongside Waymo-style robotaxis as a real market and Zipline as poised to win. He says the valuation seems rich, but he prefers it to investing in Waymo because its unit economics are better, and he thinks it is clearly a winner worth betting on.

  “It seems rich, but I like it better than investing in Waymo. The unit economics are better. And I think it's clearly a winner.”at 34:28

• Rory O'Driscoll · Scale Venture PartnersUncertain

  Zipline is an amazing company, but at $21B it needs to become infrastructure, given tiny per-delivery revenue.

  The main business is now US drone delivery, with a Walmart contract, a large Dallas operation, roughly 3 million US deliveries and more markets rolling out. The catch is that each delivery likely earns only about $5–10, so it needs enormous volume. He guesses the company is valued on being a next infrastructure play rather than a revenue multiple. The story is plausible, but at $21B the price is hard to justify.

  “You probably only get [$5-10] a delivery. Right? You've got to do a lot of deliveries to make this work... my guess is you don't get there on a revenue multiple, you get there on the, this is the next infrastructure play.”at 37:31

fal.ai

• Jason Lemkin · SaaStrBullish

  fal is the best pick of the three: strong trends, with a $40B worst-case flip.

  Lemkin would pick fal from the three deals because the underlying trends look strongest for the next two to three years. He says some investors at Altimeter think it could be flipped at $40B even in the worst case, which would return more than their preference. He admits the 6% gross margins could be a problem and that fal is perhaps the least durable of the three, but still prefers it.

  “And if you make me pick, I'm gonna pick [fal] because the underlying trends for the next two to three years. And I know some of the guys at Altimeter and others, think we can flip this at 40,000,000,000 worst case.”at 35:43

• Rory O'Driscoll · Scale Venture PartnersBearish

  At $16B fal looks near the end of its run, as third-party inference economics could shift at scale.

  The third-party inference market is only about $5B and growing fast, but its long-term structure is uncertain. O'Driscoll says his own firm misjudged these inference providers early on. The companies were great at $1–3B, but at $16B the upside may be mostly behind them, and the industry's attractive dynamics could change at scale. fal's differentiated video position helps, but generic inference players are exposed.

  “...at 16,000,000,000, you might be closer to the end than the beginning.”at 38:59

Ada Support

• Rory O'Driscoll · Scale Venture PartnersBearish

  Ada was a strong pre-LLM chatbot but has likely lost momentum against AI-native rivals like Sierra.

  Ada was a very good chatbot in the pre-LLM generation of customer service software. Its problem is that it has to compete with AI-native players such as Sierra and Decagon, and even if it has replatformed to an LLM-first technology, it has probably lost momentum. That makes it hard to compete at scale, so being bought by a larger company is a sensible outcome.

  “I think if you're ADA, you are a very good pre LLM chatbot, services chatbot. So the problem is you're up against the fact that you had a prior generation of technology, maybe you've replatformed to an LLM first technology, but you kind of probably lost the wind out of your sails.”at 45:28

• Jason Lemkin · SaaStrBearish

  Ada is a non-AI company that can't move fast enough; the Dialpad deal is not a real exit.

  Lemkin is disappointed that non-AI companies like Ada still cannot change how they run their companies, shipping at the same pace as in 2021–2023. AI-native startups ship far faster. He says the Dialpad acquisition in stock isn't really an exit without liquidity, though it is probably the best option for Ada's investors.

  “as we're rolling into 2027 guys like Ada, I'm just so disappointed that these non AI companies can't get their acts together. They can't run their companies differently than 2021, 2022, 2023.”at 46:16

Figma

• Rory O'Driscoll · Scale Venture PartnersBullish

  Figma's 40%-plus growth and strong free cash flow mean the trends are good; the hit reflects AI spending.

  O'Driscoll describes Figma as a 40%-plus revenue grower with very strong free cash flow. He says the trends are good, and the stock's drop comes from heavier AI spending that weighs on gross margins. He treats the decline as the familiar price of being a company that stays private or independent rather than being acquired, and says the stock is worth roughly half its 2022 level.

  “...are dealing with a 40% plus or minus revenue grower, very strong free cash flow... So the trends are good, and the whole hit was because they're spending more on AI, so that's going to have more gross margin hit.”at 54:30

• Jason Lemkin · SaaStrBearish

  Figma's seat-based model will likely stay in the doldrums as agents reduce reliance on seats.

  Lemkin believes the most affected companies are seat-based software that markets don't think will benefit from agents. Figma trades at around 7–8x and has fallen below $10B. Its growth is strong, but he thinks users are increasingly agents rather than humans. Unless Figma radically changes, it will stay in the doldrums. He suggests it would need to go all in on agentic design, at a cost to gross margins, to avoid becoming like Adobe.

  “I truly believe if you look at all the most impacted companies... they are seat based ones that the markets do not believe will benefit from agents and AI. And I believe Figma, unless it radically changes, it's going to be in the doldrums”at 56:26

NVIDIA

Across the shows: 114 speakers in other episodes, 98 bullish and 69 bearish calls, past 30 days

• Rory O'Driscoll · Scale Venture PartnersUncertain

  Nvidia's probe is a messy compliance inquiry; a clean outcome cannot be assumed.

  Rory says the board inquiry is about compliance and security breaches, possibly including how the government handled the matter and whether Nvidia lobbied to back off an enforcement test. He thinks national-security questions and government awareness make it hard to go away, and there is a real chance the findings are problematic. The board is long-tenured, but professionals facing congressional and DOJ second-guessing will not close their eyes. He also argues Jensen Huang's push to sell chips to anyone who will buy them is what will backfire.

  “I don't think you can assume a pass. You can hope a pass, but you can't assume it.”at 17:42

OpenAI

Across the shows: 42 speakers in other episodes, 22 bullish and 33 bearish calls, past 30 days

• Rory O'Driscoll · Scale Venture PartnersBearish

  OpenAI is in a tough spot: still private, unprofitable, and raising while Anthropic goes public.

  Rory says OpenAI is growing again and doing great things, but it is still unprofitable, still private, and still trying to raise about $11B plus retail. Anthropic will have a public float, access to public capital at scale, and a quick path to a convertible, so he would much rather be in Anthropic's position. He notes OpenAI's reported growth is stunning, but the funding position is tough.

  “Meanwhile, OpenAI is smaller, starting to reaccelerate, doing great stuff, but still unprofitable, still private, and still trying to raise $11,000,000,000 plus retail. Tough place to be.”at 28:03

Fireworks AI

Across the shows: 4 speakers in other episodes, 5 bullish and 0 bearish calls, past 30 days

• Jason Lemkin · SaaStrBearish

  Fireworks' back-to-back breaches signal long-term concerns, though attacks are industry-wide.

  Fireworks was breached twice in 30 days. Lemkin says attackers, using bot swarms, are trying to steal Anthropic and OpenAI API keys, and the same has hit Railway and competitors. A stolen default key could run up huge bills in minutes, and the attacks will get worse. He notes Fireworks is young and its security teams immature, but says it is still better than most SaaS firms. Providers can't be abandoned because capacity is scarce.

  “fireworks and others, I have long term concerns about these businesses. Fireworks was hacked twice this month. It's brutal,”at 35:30

CrowdStrike Holdings

Across the shows: 6 speakers in other episodes, 8 bullish and 1 bearish calls, past 30 days

• Rory O'Driscoll · Scale Venture PartnersBullish

  CrowdStrike will keep buying agent-security startups; a classic incumbent acquirer.

  O'Driscoll says agent security is the number one issue for end-to-end security platforms, so incumbents should buy the products that address it. CrowdStrike is a consummate acquirer, and buying $400M-ish agent-security products is a sensible, obvious move. The space typically has three or four companies, and the trick is buying the hot one before rivals do. He calls this classic incumbent-buys-product M&A.

  “And that's the CrowdStrike Palo Alto. They're all doing kind of small, 400,000,000 plus or minus perfectly great outcomes products and agent security, which totally makes sense.”at 44:02

Palo Alto Networks

Across the shows: 3 speakers in other episodes, 2 bullish and 1 bearish calls, past 30 days

• Rory O'Driscoll · Scale Venture PartnersBullish

  Palo Alto Networks will keep buying agent-security startups; a classic incumbent acquirer.

  O'Driscoll says agent security is the number one issue for end-to-end security platforms, so incumbents should buy the products that address it. Palo Alto is a consummate acquirer, and buying $400M-ish agent-security products is a sensible, obvious move. The space typically has three or four companies, and the trick is buying the hot one before rivals do. He calls this classic incumbent-buys-product M&A.

  “And that's the CrowdStrike Palo Alto. They're all doing kind of small, 400,000,000 plus or minus perfectly great outcomes products and agent security, which totally makes sense.”at 44:02

Cloudflare

Across the shows: 5 speakers in other episodes, 5 bullish and 0 bearish calls, past 30 days

• Jason Lemkin · SaaStrBullish

  Cloudflare's growth jump to 39% is driven by agent traffic and spend; a public beneficiary.

  Cloudflare's growth accelerated from the low 20s to 39% at almost $4B in revenue, and Lemkin attributes this to agents. He expects agents to become the majority of activity in his space, so Cloudflare is a public beneficiary. He says agent-driven spend is also visible at Twilio, and that for 2027 the only thing he wants to discuss is agentic beneficiaries.

  “this is what's fueling Cloudflare from 20 some odd percent to 39% growth at almost 4,000,000,000 in revenue. It's agents.”at 51:48

AI pricing & unit economics

• Rory O'Driscoll · Scale Venture Partners

  AI labs' competition will push token prices down; demand elasticity is the open question.

  Rory describes a wildly competitive two-way race between OpenAI and Anthropic on these products, with basis of competition being compute, capital and token pricing. He expects the labs to push down on prices and says that is already visible. The key question is whether a 10–20% price cut lifts usage enough to raise total revenue or lowers it.

  “I think these guys are going to push down on pricing. I think you're already seeing some of that.”at 29:57

• Jason Lemkin · SaaStr

  Compute and GPU prices are rising, not deflating, threatening always-on agent economics.

  Jason says the cost to run agent products like Lovable or Replit used to be $4–5, but now is up to $20–30 a month. Always-on agents like Muse cost about $20 a month to run all day, and compute is constrained with GPU prices going up and no deflation in GPUs. He doubts it is sustainable and believes persistent agents are needed for the products to be interesting, while expecting Meta can carry the cost.

  “[Compute] going up. GPU prices are going up. There is no deflation, right, in GPUs. So I don't know if it's sustainable.”at 32:33

Wealth, inequality & billionaires

• Rory O'Driscoll · Scale Venture Partners

  Voters rejected California's wealth tax on its merits, a clear no to such propositions.

  O'Driscoll says voters are generally suspicious of propositions, and Californians saw the wealth tax as a stupid way to make legislation. Tech founders and others made a convincing case that the state would lose wealthy residents and their income tax. A follow-up measure banning a repeat also passed, showing voters wanted a clear signal. He notes that Sergey Brin and allies outspent the unions ten to one, which matters for money-related propositions.

  “I think Californians recognize it's an incredibly stupid way to make legislation.”at 1:00:04

• Jason Lemkin · SaaStr

  Passing the California wealth tax would have driven top founders and talent out of the state.

  Lemkin says he spoke to two quiet billionaire founders, not CEOs, who were already planning to leave California before the vote. He sees the tax as a threat because the people building companies, not passive wealth holders, would leave. He thinks passing it would have signalled that California is toxic to the most successful entrepreneurs, with a domino effect on the centre of AI and wealth. He cites Anthropic's IPO wealth, which would be at stake.

  “I spoke to two quiet billionaire founders, but not CEOs who were both planning to leave. ... it would have sent a clear message that California is toxic to the most successful entrepreneurs.”at 1:01:34

US–China AI & chip race

• Rory O'Driscoll · Scale Venture Partners

  China loses as GPU sales and offshore hosting for China become harder to sustain.

  Rory argues the Nvidia controversy will make it harder for China to get GPUs, both through direct sales and by hosting them across the border in neighbouring countries. He concludes the whole arrangement gets shut down, so China loses big.

  “China loses big time here. China's gonna lose. It's gonna be harder to not only sell GPUs to China. It's gonna be harder to just host them across the border.”at 0:00

Market sentiment, positioning & volatility

• Rory O'Driscoll · Scale Venture Partners

  Last week's Nasdaq selloff was reactive, driven by forced deleveraging of aggressive funds.

  Rory says stocks popped on Wednesday after the election outlook, then fell Thursday on data-centre angst, which he sees as reactive rather than a sign the AI trade is cracking. He thinks aggressive hedge-fund pods that were over their skis got blown up, since risk managers cut them at drawdown thresholds, and that is all that happened.

  “So a lot of this is fairly reactive. It's some version of, oh my god, we can't build data centers.”at 21:48

AI: overall outlook

• Rory O'Driscoll · Scale Venture Partners

  The AI trade won't crack on supply constraints; only weaker demand would break it.

  Rory argues the AI trade does not crack because of supply-side problems such as data centres not being built or Nvidia losing share to AMD. If demand is there, evidenced by OpenAI and Anthropic ARR, everything else will be roughly fine. Some companies like CoreWeave may get ahead of themselves, but the broader trade holds.

  “the AI trade doesn't crack on the supply side, doesn't crack... If the demand is there, everything else will be roughly fine.”at 21:48

AI infrastructure & data centers

• Jason Lemkin · SaaStr

  Markets are right to be nervous given the heavy capital and debt needs of AI buildout.

  Jason says it makes sense for markets to be nervous about AI because the buildout rests on many assumptions, requires a lot of capital to be deployed, and needs a lot of debt. He ties this to the Nasdaq selloff.

  “I think it makes sense for the markets to be nervous about this. There's a lot of assumptions. There's a lot of capital to be deployed. There's a lot of debt required.”at 23:21

Frontier AI labs & model race

• Jason Lemkin · SaaStr

  OpenAI and Anthropic will keep leapfrogging; switching between their models is trivial.

  Jason says running a simple eval between models is trivial and developers have no loyalty, so the labs keep trading the lead. Both are growing, and only they have the compute available in the US, which keeps the race close. He cannot predict the revenue ratio, though he expects both to have far more revenue in twelve months.

  “it is truly trivial. So the entire time we've done this show, it's been ping pong, right? Ping pong between these guys. They both grow and it's gonna continue to be ping pong.”at 31:09

AI & jobs

• Rory O'Driscoll · Scale Venture Partners

  AI is the excuse for layoffs; the real driver is demand growth making marginal headcount unnecessary.

  O'Driscoll says AI has made companies notice inefficiency that was always there, since many firms overstaffed in 2021. AI leverage plays some role, but the bigger driver is strong secular demand. When demand is growing that fast, a company doesn't need the marginal headcount it hired earlier. It can staff the machine and take the business as it comes, which makes efficiency look compelling.

  “There is AI leverage, but I think the real truth is, yeah, especially when you've got that kind of growth, if you've got a giant secular demand pull, you don't need to do all the marginal things that you were doing in 2021 to make revenue happen.”at 50:35