20VC with Harry Stebbings

Sequoia Leaves London, Close Scout Program & End Evergreen Fund | NASDAQ Loses $2.5T in 24 hours: summary

YouTube summary10 sectionsWatch on YouTube ↗

This is an AI-generated summary of the YouTube video "Sequoia Leaves London, Close Scout Program & End Evergreen Fund | NASDAQ Loses $2.5T in 24 hours" (20VC with Harry Stebbings), made with Samuraize and published by Samuraize. It condenses the YouTube video into 10 titled sections you can read in a couple of minutes, each linking to the moment in the video it covers.

1
Filed under📈 Business0 comments🍱 Add to trayReport
Study this

Sequoia Leaves London, Close Scout Program & End Evergreen Fund | NASDAQ Loses $2.5T in 24 hours

20VC with Harry Stebbings

切

Sequoia’s structural reset 1:00

Sequoia is cutting back in several ways. It is leaving London, ending the scout program, and combining early and growth investing into one team. The talk turns first to the evergreen fund, because that was the part that felt most striking. The idea was to keep great public tech stocks for decades, but the structure was awkward, forced LPs to stay in, and landed badly when markets fell after 2022. The point was simple enough, but it was too complicated to live with.

切

Focus over extra machinery 6:01

The change is read as a move toward simplicity. London can still be useful, especially for founders and for deals in Europe, but the view here is that Sequoia can cover that ground from the United States if it stays close and moves fast. The scout program also seems to have lost its edge once others copied it. The wider lesson is that institutions can last even as partners move on, and that a tight partnership may work better than more programs and more layers.

切

Jensen and the inquiry 13:00

The talk stays on Jensen Huang and the investigation around him. It is framed as a compliance review, not a probe into a relationship. The concern is whether there were security breaches, government handling issues, or lobbying around a test on hyperspeed. Even so, the view is that you should not lose sight of how important he is to Nvidia, AI, and the US.

切

Market shock and Anthropic 20:00

The discussion then turns to the NASDAQ 100 losing $2.5 trillion in a day. The drop is described as reactive, tied to fears about data centers, chips, and how much capital the AI buildout will need. Anthropic’s IPO is read as a test case. The pricing is said to be careful, with a lower range and a smaller public valuation than the private one.

切

Public capital advantage 26:32

Anthropic is cast as the better place to be than OpenAI because it is about to go public, gain float, and tap public markets at scale. Once public, it can use converts or secondary offerings to add tens of billions to the balance sheet and meet demand more easily. OpenAI is still private, unprofitable, and trying to raise more money. The move to quarterly reporting is also welcome because it will finally reveal real numbers.

切

Compute and pricing 30:00

The discussion then turns to cost. The agent products are expensive and compute heavy, and the market may be finding out whether token pricing can fall without hurting revenue. They note that a 10 or 20 percent price cut could either lift usage enough to grow revenue or not. The same pressure shapes the broader market, where GPUs are still getting more costly even if compute should get cheaper over time.

切

Security Breaches and API Keys 40:31

Fireworks is used as the example of a wider security problem. The breach matters because attackers are not just stealing data. They are hunting API keys for inference, especially Anthropic and OpenAI keys, and then spending the account almost at once. A stolen key can burn through $100,000 in minutes. The speaker says this is now the most lucrative crime in the field, and that it is only getting worse as bot swarms become harder to stop.

切

Acquisition Pressure and Efficiency 45:01

The talk then shifts to M&A. CrowdStrike and Palo Alto are framed as natural buyers in agent security, while Meta and OpenAI are buying nearby products in the software development stack. Dialpad buying ADA is treated as a reasonable soft landing for a slower company, not a triumph. The same complaint runs through the rest of the discussion: older companies are moving too slowly, while AI native teams ship much faster and force the rest of the market to change or fall behind.

切

Seat models under pressure 54:30

Agents are said to care little about interface polish, so seat-based software faces more doubt. Figma is used as the example. Even with strong revenue growth and free cash flow, the stock drop, antitrust worry, and lower valuation can create a malaise inside the company. The view here is that the market now believes these tools may be bypassed as agent-driven work becomes normal.

切

Wealth tax and exits 1:02:31

California’s wealth-tax measure failed, and a follow-up ban on trying again passed. That result is read as a sign that voters can reject policies they see as harmful to builders. The speakers say the bigger risk was driving founders and investors out of California. They point to Anthropic’s huge employee windfall as the sort of wealth creation the state should want to keep.

AI-generated summary. It can be wrong or incomplete - check anything that matters against the original.

Summarize your own YouTube video

Paste a YouTube link, article, PDF, ebook or slide deck and get a summary like this in seconds. Free to try, no sign-up needed.

⚔️ Try the YouTube summarizer

Discussion

Sign in to join the discussion. Sign in

More from the Bento Box

Browse the Bento Box →

With your OK, Microsoft Clarity and Google Analytics set cookies (data goes to the US) so we can see what breaks. Cookie details