Anthropic’s IPO Details, Building ‘The Sims,’ & the World Bank 10/27/26

Anthropic

Across the shows: 63 speakers in other episodes, 49 bullish and 50 bearish calls, past 30 days

• Leslie Picker · CNBCUncertain

  Anthropic's compute commitments may be the biggest claim on its large cash pile.

  Anthropic's S-1 shows about $112B in cash plus $19B in short-term investments as of June 30, more than Meta, Microsoft, Nvidia or Amazon. Against that sit $587B of long-term compute obligations, leases and fixed commitments running one to twenty years, plus a separate $245B tied to new and amended agreements. Picker concludes that compute, not cash, may be the largest claim on Anthropic's balance sheet. Some of those commitments are non-cancelable, so they will need future equity and debt raises and a path to profitability.

  “...compute obligations, leases and fixed commitments that span anywhere from one to twenty years. [Anthropic] said these amounted to $587,000,000,000 as of June 30 plus a separate $245,000,000,000 tied to new and amended agreements. So if cash is king, compute may be the biggest claim on the crown.”at 4:20

• Leslie Picker · CNBCUncertain

  Anthropic's lockup schedule and index inclusion may offset each other, leaving the stock's trading unclear.

  The S-1 lists roughly two dozen lockup dates, from 27 days after the offering to about 365 days, with some tied to earnings. Picker expects a stair-step unlock similar to SpaceX's. Lockups usually pressure a stock as shares unlock. Fast-track index inclusion works the other way by growing the float and forcing index buying. Bankers line up lockups with likely index inclusion to balance the two, so Picker says it is unclear whether this will work for Anthropic as it did for SpaceX.

  “This idea that lockups tend to provide or produce some pressure on the stock, but then you also have index inclusion, that works the other way, whereas the float gets bigger,”at 6:00

• Leslie Picker · CNBCUncertain

  A $1.5–2T Anthropic IPO valuation needs growth sustained over a long period.

  Anthropic's revenue growth slowed from about 60% to about 50%, though still very strong. Reports suggest a $1.5 to $2 trillion IPO valuation. Picker argues that to command that price, Anthropic must show growth can continue over a long period. This matters because its future non-cancelable compute obligations will eventually require equity and debt raises and a pathway to sustainable profitability.

  “in order to command, say, a 1.5 to $2,000,000,000,000 valuation as reports are out there suggesting, you're going to need to be able to show that that growth can continue over a long period of time.”at 9:00

AI: overall outlook

• Ajay Banga · World Bank Group

  Small, targeted AI applications are a real productivity enhancer for developing economies.

  Banga contrasts 'big AI' built on large language models, which need compute, power, data, people and capital that are scarce in the developing world, with 'small AI' for agriculture, health care and education. The World Bank is building and launching these in El Salvador, India and other countries. Upper-middle-income countries are adopting faster, and Africa needs reusable applications plus local data and smaller data centers. He says small AI creates jobs that would not otherwise exist.

  “small AI as compared to big AI, because what's being talked about in the developed world is all about large language models. ... But small AI applications for agriculture, applications for health care, applications for education and skilling ... That, Andrew, that's a real productivity enhancer.”at 16:48

• Bing Gordon · Kleiner Perkins

  Long-term positive on AI, though edge-case 'money plays' will break it, fixable with data and fast reversion.

  Gordon says he is long-term positive on AI. He compares it to Madden's 'money plays', exploits such as the punter running back for a touchdown. Those edge conditions break AI systems the way the 'Hugging Face' bug did in Madden. You can never test every edge case, so the answer is to rely on data at scale and the ability to revert quickly. The bullish case rests on that engineering discipline, not on a specific financial metric.

  “I'm long term positive. ... So AI always has money plays. The hugging face was a money play. It was an edge condition that wasn't prepared for. There's always edge conditions you can't prepare for.”at 34:21

AI infrastructure & data centers

• Ajay Banga · World Bank Group

  AI will be transformative and profitable, but investors are worried about circular financing and bumps ahead.

  Banga says investors are worried about circular financing in AI infrastructure and about unknown future business models for the large investments. He still believes AI will be transformative and profitable over time. He expects bumps along the way and says no one knows when the next one will come. Emerging markets are different: their capital goes to roads, power, digitization and health and agriculture infrastructure, not AI-linked deals.

  “In general, I still believe that AI is gonna be transformative and profitable, but everybody sees bumps in the road from here to there, and no one knows when the next bump will come.”at 20:28

Emerging markets

• Ajay Banga · World Bank Group

  Emerging-market sovereign spreads have tightened, but higher rates raise EM debt-servicing costs.

  Overall rates have risen, but emerging-market spreads have narrowed over the past couple of years. Banga credits that with helping EMs through the period, since they learned from the previous fifteen years and managed their balance sheets relatively well. He warns, however, that even with tight spreads, higher overall rates mean governments spend more on debt servicing. He notes that World Bank private capital mobilization goes mostly to projects, not to governments.

  “Rates overall have gone up, but there has been a reduction in emerging market spreads over the last couple of years. ... Even if the spreads are tight, the reality is if overall rates go up, they are spending more on debt servicing than earlier.”at 20:46

Natural gas & LNG

• Ajay Banga · World Bank Group

  Developing countries with gas should be allowed to use it, since affordable baseload power is a human development right.

  Banga argues that gas is a natural transition fuel in the developed world, so developing countries with domestic gas should be allowed to use it to manage their own energy. He calls affordable baseload power a human development issue. This is his reason for shifting the World Bank's stance on mitigation versus adaptation, where he now prioritizes resilient infrastructure and adaptation.

  “If you have gas available in a country, and gas is a natural transition fuel in the developed world, then why would you not allow them to get access to that gas to be able to manage their own energy? Because the right to affordable baseload power is kind of a human development issue.”at 22:28

Critical minerals & rare earths

• Ajay Banga · World Bank Group

  World Bank should finance five times its current critical minerals amount in coming years.

  Banga says the World Bank is not yet financing enough critical minerals. Minerals are needed for urbanization (iron ore, bauxite), agriculture (potash), and energy (copper wire to connect 300 million people in Africa to baseload power). His plan is to get 20 countries to adopt regulatory frameworks for safe mining, and to add more processing in-country to create local jobs. He gives the example that breaking a magnet in Zambia is hard, but processing there is possible.

  “I'm not yet financing enough. I think we should be doing much more. We should be doing five times the amount in the coming years.”at 24:11

Venture capital & startups

• Bing Gordon · Kleiner Perkins

  Tech cycles are faster and bigger than ever, with trillion-dollar companies now possible.

  Gordon cites a shift in scale. In 2000 about five tech startups reached $1 billion in value, and in 2010 investors doubted more than five new unicorns would appear. Now there are trillion-dollar companies. He argues technology has reduced entropy, so scale that used to be hard, such as 30 million cable households in 2008, now comes easily. He takes this as evidence that cycles are both faster and bigger.

  “I would say that the first thing is the cycles are faster than ever. On the next is they're bigger than ever.”at 36:01