Overview of 20VC: Sequoia, NVIDIA, AI Markets, and Anthropic’s IPO
Harry Stebbings, Rory O’Driscoll, and Jason Lemkin discuss Sequoia’s restructuring, scrutiny facing NVIDIA CEO Jensen Huang, a sharp NASDAQ selloff, and the prospects for Anthropic’s IPO. They also weigh AI infrastructure investments, rising security threats, how established software companies can adapt to AI agents, and the potential effects of California’s proposed wealth tax.
Sequoia’s Restructuring and Evergreen Fund
Sequoia is leaving London, ending its scout program, combining early stage and growth investing into one team, and winding down its evergreen fund.
The panel views the changes as a move toward focus and simpler operations. They argue that a London office can help with early stage investing, but Sequoia’s brand and reach may let it invest in leading European companies from the US. They see the scout program as less valuable now that similar programs are common and more work to manage across a smaller partnership.
The evergreen fund’s goal—holding strong companies beyond IPOs—made sense in theory, since much of a company’s long term growth can come after it goes public. In practice, the structure limited LPs’ choices, complicated liquidity, and was launched during a market downturn. The panel says partners can choose to hold their own shares, while LPs should be able to decide for themselves.
NVIDIA, Jensen Huang, and China
The panel discusses reports of an investigation related to Jensen Huang and an alleged relationship with a suspected Chinese spy. They distinguish the personal allegations from the more consequential questions: whether security or compliance rules were breached, and whether US restrictions on chips for China were circumvented by hosting GPUs in nearby countries.
The panel says losing Huang would matter to NVIDIA and the AI industry, while cautioning that the investigation may be difficult to dismiss because of its national security implications and expected scrutiny from Congress and the Justice Department. They argue that NVIDIA’s push to sell chips broadly reflects its commercial interest in protecting market share, but predict that the controversy could make it harder to sell or host GPUs for Chinese customers.
NASDAQ Selloff and AI Market Outlook
The panel attributes the NASDAQ’s reported $2.5 trillion one day decline largely to market volatility and investor positioning, rather than clear evidence that demand for AI has weakened. O’Driscoll argues that the AI trade ultimately depends on customer demand, which should show up in revenue for companies such as Anthropic and OpenAI.
Lemkin is more cautious: AI companies need enormous amounts of capital and compute, and investors may be tiring of repeated large funding rounds. The guests agree that the market will be watching whether AI usage and revenue can justify that spending.
Anthropic’s IPO and OpenAI Competition
The guests see Anthropic’s proposed IPO pricing as a deliberate attempt to start at a level that can attract demand and potentially rise during the offering. They explain that media estimates near $1.4 trillion may include unvested options, while a lower figure reflects the public market’s treatment of those shares.
They disagree on how the stock might trade. O’Driscoll expects strong interest and sees public listing as a major advantage: it could give Anthropic substantial cash and access to future public market financing. Lemkin worries that investor appetite for large AI offerings may be waning and expects the stock could trade flat. Both note that Anthropic and OpenAI will need significant future funding.
OpenAI’s reported 70% annualized revenue growth in the third quarter and rising enterprise sales point to a competitive two company race. The panel expects compute availability, capital, and token pricing to shape the contest. They also debate whether compute heavy coding agents can sustain their costs until computing becomes cheaper.
AI Investment Picks: Zipline, Fal.ai, and Revolut
The panel considers three investments: Zipline at a reported $21.5 billion valuation, Fal.ai at $16 billion, and Revolut at $115 billion in a secondary sale.
- Zipline: Its US drone delivery business and Walmart relationship offer significant potential, but the panel questions whether delivery economics can support its valuation.
- Fal.ai: It is benefiting from fast growth in third party AI inference, though the panel worries that competition and changing compute economics could make the business less durable.
- Revolut: The guests see it as a potentially enduring financial platform with room to expand across Europe, though its valuation is high and its future still depends partly on founder Nick Storonsky.
Lemkin is most drawn to Fal.ai for near term growth, while Stebbings favors Revolut for its potential to compound over time.
AI Security, M&A, and Company Adaptation
The panel discusses recent attacks targeting AI infrastructure providers, including Fireworks AI. They say attackers are increasingly stealing API keys to consume or resell access to expensive inference services. Their advice is to treat API keys and usage limits as high risk: stolen credentials can generate large bills quickly. They expect security firms to benefit as companies face more automated attacks.
Acquisitions by CrowdStrike and Palo Alto Networks are presented as examples of established security companies buying capabilities for emerging threats such as AI agent security. The guests also discuss the Dialpad–Ada deal, arguing that acquisitions can give older software businesses a path forward, but may offer limited liquidity to investors.
A broader concern is that many established companies are moving too slowly to redesign products and operations around AI. The panel highlights Cloudflare’s growth, Freshworks’ efficiency, and Palantir’s revenue per employee as examples of how AI and strong demand can support leaner operations. They expect AI agents to pressure seat based software models and call for companies such as Figma to invest more deeply in agentic design tools.
California Wealth Tax and Talent
The panel views the defeat of California’s proposed wealth tax as a signal that voters were concerned about driving entrepreneurs and investment out of the state. They say founders and employees who build companies contribute economic activity, not just passive wealth, and argue that retaining them matters to California’s technology ecosystem. The discussion also notes that campaign spending may have influenced the result.