Neostellar: Focus on Anthropic IPO Overshadows $200 Billion in AI Venture Funding

Willie Lee, a principal at investment firm Neostellar, thinks talk of a possible Anthropic IPO is drawing too much attention. In an interview on Bloomberg's The Close, hosted by Scarlet Fu and Hema Parmar, he warned that a high-profile event can hide the scale of AI investment and the speed of AI adoption. Lee's main argument is a number. By his account, nearly $200 billion in venture capital was invested in the US in the first half of the year. That is more than 74% of the estimated total for the entire previous year.

What exactly Lee said

The occasion for the interview was the pre-IPO investor day of Anthropic, the developer of the Claude family of models. Lee is not talking about the company itself but about where the market focuses. Any potential offering of this size becomes a barometer of sentiment for the whole industry. The news cycle bends around it: valuation, timing, the makeup of investors. Everything else falls into the background.

In Lee's view, this distorts the perspective. Money is flowing into AI projects on a broad front, and adoption of the technology is happening in more places than just the largest foundation model developers. Bloomberg's account does not specify which segments he meant: infrastructure, applied services or something else. So conclusions about the structure of the flow cannot be drawn from this quote.

The half-year arithmetic

The figure looks impressive on its own, and more so when broken down.

• Volume: nearly $200 billion in venture capital invested in the US in the first six months of the year.
• Share of last year: more than 74% of the estimated total for the entire previous period.
• Our estimate: if $200 billion is a little over 74%, the estimate for last year falls in the range of $260–270 billion.
• Pace: doubling the half-year gives roughly $400 billion on an annualized basis. This is our calculation, not Lee's forecast, and it assumes the second half of the year does not slow down.
• Context: Anthropic's pre-IPO investor day is taking place against the backdrop of these numbers, not in a vacuum.

One caveat: the source refers to venture capital and to an estimated total for last year. Such estimates differ between analytics firms, and the total may be revised. Treat the comparison as a rough guide, not precise accounting.

Why one company does not equal the market

The IPO story is convenient for headlines. There is a specific firm the public understands, and a clear date. An investment boom works differently: thousands of rounds, dozens of sub-sectors, different payback horizons. If you keep only one deal in mind, it is easy to mistake its outcome for a verdict on the whole sector. A successful offering will read as confirmation of the entire trend, a failed one as a signal to retreat. Lee suggests not putting that much weight on a single ticker.

The thesis is easy to test, though. If the money really is spread widely, the market will survive swings around any single security. If a significant share of capital is concentrated in a few of the largest rounds, dependence on sentiment around the leaders will be higher than it seems. The account of the interview does not show which scenario is closer to reality.

How to read this signal

For startup founders, the takeaway is simple: track the dynamics of the whole venture market, not news about Anthropic. The amount of available capital and the competition for it are set by the broad picture. Investors should look separately at the structure of investment: who gets the bulk of the $200 billion and how much is left for applied companies. Executives planning AI adoption should look at their own metrics: cost per task, quality of results, payback. Procurement decisions should not depend on a vendor's share price. Watch for new data as well: if third-quarter reports confirm the first-half pace, Lee's thesis will gain serious support.