• For private AI investing, consider a small initial position and add only when revenue quality and business results are verified; treat Anthropic as a high-risk private investment, not a broadly suitable 20% portfolio allocation.
• Before choosing private AI startups over public alternatives such as NVIDIA (NVDA), account for illiquidity and dilution; the discussion suggested private investments should target roughly 10 percentage points of additional return to justify those risks.
• Don’t rely on a startup acquisition as downside protection, and scrutinize claims based on projected or annualized revenue rather than live, durable sales.
• For any investment that has grown substantially, consider taking partial gains to manage concentration risk; the discussion gave no specific public-stock price targets or near-term buy/sell calls.

Anthropic (Private)

• Menlo identified Anthropic as a core AI investment and said it had stayed with Anthropic rather than investing in OpenAI.
• Menlo’s Anthropic position reached 20% of the fund after additional investments. The guest presented this as an exceptional case, made after the company had produced strong results—not as a standard position size.
• The guest said Menlo owned less than 2% of Anthropic, using it to illustrate that a very large potential outcome can matter more than a high ownership percentage.
• The discussion emphasized increasing a position as measurable evidence of an outlier emerges, rather than committing a large share of a fund before that evidence is available.

Takeaways

• For private AI investing, the discussion favors an initial position followed by selective follow-on investment when business results support greater conviction.
• Low ownership does not automatically make a large opportunity unattractive, but returns may depend heavily on the company becoming an exceptional outlier.
• Menlo’s 20% exposure is an example from one fund, not a general portfolio allocation recommendation.

OpenAI (Private)

• Menlo said it did not invest in OpenAI, citing its preference to make a committed investment in Anthropic rather than back both companies.
• OpenAI was also discussed as an example of the large AI companies that may eventually access public markets and compete for investor capital.

Takeaways

• The conversation frames choosing among leading AI companies as a question of investment focus and commitment, not just whether the sector is attractive.
• No valuation, return forecast, or specific recommendation on OpenAI was given.

NVIDIA (NVDA)

• NVIDIA was described as a major AI infrastructure company that venture-backed businesses may depend on or indirectly “pay a tax” to through their use of computing resources.
• The guest argued that private-market investments must compete with the possibility of investing in large public companies such as NVIDIA through a low-cost index fund.
• NVIDIA’s reported acquisition activity was discussed as part of a broader pattern of strategic competition and AI-related dealmaking.

Takeaways

• When evaluating a private AI investment, consider how dependent its economics are on major infrastructure providers such as NVIDIA.
• The guest’s benchmark for private investing was to seek returns that justify the added risk and illiquidity compared with public-market alternatives. He said venture should aim to exceed those alternatives by about 1,000 basis points.

Advanced Micro Devices (AMD)

• AMD was cited as an example of a large public company making an AI-related acquisition.
• The guest suggested that acquisitions can prompt competitive responses from other large technology companies.

Takeaways

• AI dealmaking may be influenced by strategic competition among large technology companies, but the discussion did not establish a specific investment case or price target for AMD.

Meta Platforms (META)

• Meta’s hiring and investment activity in AI was discussed in the context of competition for talent and technology.
• The guest praised CEO Mark Zuckerberg’s capital allocation, citing Meta’s past acquisitions of Instagram and WhatsApp as examples of investments he viewed as successful.
• The discussion also noted that Meta’s large market value can make sizable acquisitions a relatively small portion of the company.

Takeaways

• The guest emphasized capital allocation as an important quality to assess in a company’s leadership, alongside product vision.
• The examples were qualitative; no valuation analysis or recommendation to buy Meta was provided.

Apple (AAPL)

• Apple was mentioned as a possible strategic buyer of an AI company in a hypothetical discussion about potential downside protection for private investments.
• The guest cautioned that investors should not assume a large incumbent will acquire a startup if it struggles; a potential buyer may prefer to hire the team instead of acquiring the company.

Takeaways

• Treat a possible acquisition as uncertain, not as a reliable floor on an investment’s value.
• The transcript gave no specific view on Apple’s stock or likelihood of making a particular acquisition.

Snap Inc. (SNAP)

• Snap CEO Evan Spiegel was praised as a product visionary, but the guest said shareholders had not been rewarded over the preceding seven to eight years.
• The guest contrasted product strength with capital allocation, arguing that product vision alone does not guarantee attractive shareholder returns.

Takeaways

• Assess whether product success is translating into financial results and shareholder value—not just whether a company has a compelling product.
• The discussion did not provide a price target or a current buy-or-sell recommendation.

Salesforce (CRM)

• Salesforce was cited as a major venture success: the guest referred to it as the standout return in a portfolio discussed on the show.
• The example illustrated the risk of selling an exceptional investment too early, while the guest’s own experience with another stock showed the risk of holding too long.

Takeaways

• The discussion supports avoiding an all-or-nothing approach to exits: consider taking some gains while retaining exposure to a potential long-term winner.
• The right decision depends on an investor’s circumstances and portfolio; no specific Salesforce selling level was mentioned.

Avanex (AVNX; referred to as “Avonex” in the transcript)

• The guest recalled buying $5,000 of the stock after its IPO. At one point, the position was worth about $200,000, but he did not sell when his fiancée suggested taking some gains for a house down payment.
• The stock later fell by 90%, and he said he eventually sold for roughly $8,000 to $9,000.
• He described the experience as a lesson in taking some chips off the table when a position has become a large part of a personal portfolio.

Takeaways

• A large unrealized gain can still disappear. Consider whether a position has grown too large relative to personal financial needs and risk tolerance.
• The guest stressed that there is no universal rule for when to sell; personal circumstances matter.

AI Startups and Private AI Investing

• The guest described the current AI funding environment as unusually fast-moving, with some companies seeking rounds of $100 million or more, and some labs seeking billions.
• Seed rounds have grown: the guest cited application companies raising $10 million to $20 million, compared with older seed rounds of roughly $3 million to $5 million.
• Menlo said it sometimes invests at seed to secure a place in a company’s future financing, while expecting to invest more if the company later demonstrates strong results.
• Companies mentioned in the discussion included Cursor, SpaceX, Hugging Face, OpenRouter, Town, Higgsfield, Lagora, and Instinct. They were used in examples about rapid growth, fundraising, acquisitions, or market dynamics; the transcript did not give a full investment thesis for each.
• The guest warned that startup revenue figures can be presented in ways that overstate underlying performance—for example, annualized projections from a short period, contracted revenue that is not yet live, or other metrics that may be gamed.
• He argued that investors should focus on whether founders are building durable businesses and long-term value, rather than pursuing fundraising markups.
• The guest said venture investors may face about 60% dilution from an initial seed investment to an exit, from later financing and option-pool expansion. He also noted that fast-growing companies may need less financing and experience less dilution.
• He cautioned against assuming a struggling startup will be acquired for a high price. Potential acquirers may instead hire the team, and historical acquisition patterns can change when market conditions turn.
• A hypothetical downside case in the conversation referenced a $1.5 billion acquisition value, but the guest warned that such outcomes should not be treated as dependable downside protection.

Takeaways

• Verify revenue quality: distinguish live, collected or recurring revenue from contracted, projected, or extrapolated figures.
• For early-stage investments, treat initial checks as option-like bets and reserve the ability to invest more after business evidence improves.
• Account for future dilution when estimating ownership at exit, and consider how long the company may need to remain private.
• Do not rely on a future acquisition as a safety net. The guest specifically warned that acquirers may not protect a startup’s investors.
• The transcript offered no specific price targets or public-market recommendations for the private companies named.

Venture Capital Funds and Portfolio Strategy

• The guest said seed investing is more difficult when rounds are large and major funds are willing to invest early at high valuations to secure access to later rounds.
• He described venture portfolios as a set of option-like investments: make enough initial investments to have multiple chances at an outlier, then increase exposure to companies with stronger evidence.
• He said ownership still matters, while noting that a smaller stake in an exceptionally valuable company may outperform a larger stake in a modest outcome.
• Menlo said it would generally avoid investing more than 20% of a fund in one company, while acknowledging it had reached that level in Anthropic.
• The guest highlighted the importance of returns in cash, or DPI, as well as headline valuation measures such as TVPI. He said quick growth and exits can support stronger IRR and reduce dilution.
• For public-market competition, he said venture investments should aim to beat public alternatives by about 1,000 basis points to justify their additional risk and illiquidity.
• He warned that fast fundraising and fund deployment can increase exposure to a single market period. He cited a Menlo fund invested over a 10-month period between 2000 and 2001 as an unsuccessful example and said vintage diversification matters.
• The guest said a major market break often begins with debt defaults and leverage rather than equity losses alone, because debt holders expect repayment.
• When asked about outside funds, he named Bessemer as a growth fund he respected and E14 as an interesting seed group focused on AI companies. He described Benchmark as particularly difficult to compete against. These were personal views, not performance guarantees.

Takeaways

• Evaluate venture funds on realized cash returns, investment discipline, and the partners founders respect—not only on reported valuations or past headline performance.
• Look at portfolio construction, position sizing, and fund vintage exposure together; rapid deployment may be necessary in a fast-growing market but can concentrate timing risk.
• For private investments, compare expected returns with public alternatives and account for illiquidity, dilution, and the possibility of losing the entire investment.

SaaS and Private Equity

• The guest said some venture-backed SaaS companies raised money at high valuations during the 2021 market and may now face difficult outcomes as AI changes the software landscape.
• He argued that private equity firms may have more ability to restructure companies because they often have majority control, while some venture-backed companies may lack a single owner able to make difficult decisions.
• For some weaker SaaS companies from that period, he described getting investors’ capital back as a favorable outcome and a total loss as a possible downside.

Takeaways

• Assess how AI could affect a software company’s business and whether its owners have the ability to respond.
• Do not assume that a high-quality or established software business will maintain its previous valuation; ownership structure and the ability to restructure may matter.

Coinbase (COIN)

• Coinbase and CEO Brian Armstrong were discussed as an example of a company setting explicit boundaries around political activity at work.
• The guest praised Armstrong’s willingness to state the company’s principles, while noting the stance had drawn criticism and led some employees to leave.

Takeaways

• The discussion was about leadership and corporate culture, not Coinbase’s financial performance or valuation; it did not provide a stock recommendation.