LBXpro
Insights
Newsletter

Anthropic goes public. What will secondary money make?

Victoria Arinstein4 min read
Office building in San Francisco

Anthropic starts trading today, at $70–78 a share, about $1.33–1.4 trillion.

Sam Parr worked out on X that Anthropic will make about 1,000 employees worth more than $100 million each at a $2 trillion market cap. That would be roughly 1 in 12 Americans with $100 million in investable assets.

And then there are the people who bought Anthropic on the secondary market before 2026. They listed $740 million of bids in 2024 and 2025. If each of those bids was one person, that’s about 120 buyers. With the median bid around $2 million, 20 of them would have passed $100 million at IPO, next to Sam Parr’s 1,000 employees. Let’s talk about those guys.

A textbook case of private investing

Anthropic's implied valuation went from $27 billion in January 2024 to $1.34 trillion on 15 November 2026. ARR went from $87 million to $81 billion.

Anthropic implied valuation and ARR, Jan 2024–Nov 2026: valuation $27B to $1.34T, ARR $87M to $81B

Source: LBX Pro implied valuation (secondary price × implied shares) and reported ARR, Jan 2024 – 15 Nov 2026.

Secondary buyers almost always paid above the last round. The median premium since January 2025 is +25%.

Anthropic implied valuation vs primary rounds, 2024–2026: secondary price above most round valuations

Premium = secondary price ÷ last round price − 1. Source: LBX Pro, Jan 2024 – 15 Nov 2026.

So far Anthropic is the textbook case for secondary investing: the earlier you bought, the better you did. But most of the money came late. Bids listed on LBX Pro grew from $168 million in 2024 to $572 million in 2025 and $1.31 billion so far in 2026. Almost two-thirds of all bids since 2024 came this year.

Anthropic bids and asks listed on LBX Pro by month, 2024–2026: bids $168M, $572M, $1.31B per year

LBX Pro listed orders (not trades), capped at $25M per order, repeat listings counted once. Through 15 Nov 2026.

Every month before 2026 was a good time to buy. Bought at its month’s price, a share from 2024 is worth 11–23 times as much at the bottom of the IPO range. November 2025, at about $17 a share, was up 390% six months later. The biggest month for bids came much later: June 2026, at an average bid of about $56. 41% of all bid dollars since 2024 are up 10% or less today.

Anthropic secondary price, bids and 6-month returns by entry month, 2024–2026: best entry Nov 2025, +390%

Bids are listed orders, not trades. Prices are after the 1:10 split. Source: LBX Pro.

Everyone lowballed Anthropic, including Anthropic

Forecasting AI revenue is a great way to be wrong in public. Since 2024, nearly every forecast for Anthropic’s revenue, including its own, has been too low, and it has been overtaken within months.

Past Anthropic revenue forecasts vs results, 2023–2026: $850M, $26B and $70B forecasts all beaten

Sources: TechCrunch, The Star, SiliconANGLE.

The forecasts now are much larger, and they disagree:

Current Anthropic revenue forecasts, 2026: $52–53B in 2026, $270–280B in 2028, over $600B by 2030

Forecasts as reported by Oninvest (citing the Financial Times), Reuters via BusinessDay and Euronext, not LBX Pro estimates.

The end of the lock-up or the SAR: which comes first?

There is more to forecast than revenue, but so far markets are not sensitive to it. On 5 November METR, an independent group that evaluates AI models, published its assessment of AI research at Anthropic. Progress over the past year ran at about twice its earlier trend, and METR guessed that around 80% of that speed-up comes from AI itself. Anthropic redacted much of the evidence METR found most informative.

Five days later the AI Futures Project, the team behind the AI 2027 scenario, led by former OpenAI researcher Daniel Kokotajlo, updated its forecast:

“Looks like the intelligence explosion will happen in 2027 after all”

They now expect a superhuman AI researcher (SAR) sometime in 2027: an AI that does AI research better than a team of the best human researchers. Their central estimates run from February to August 2027. They give a 35–65% chance that AI beats top human experts at virtually all tasks before 2028.

Why they think it is dangerous: companies are not on track to make such a system safe; letting it improve itself would put us in an extremely bad position; and China could probably steal how it was trained. They say it is urgent to pause or substantially slow frontier AI development. One of the authors calls METR’s report perhaps the biggest single piece of evidence on near-term economic growth, and asks why markets did not react at all.

The researchers don’t connect a SAR to revenue. Maybe they should make a guess; it would get them more attention.

And we are wondering what comes first: the end of the lock-up for most private investors, around May 2027, or the SAR.

Sources: METR, 5 Nov 2026, and AI Futures Project, “Q3.5 Timelines Update”, 10 Nov 2026: their forecasts, not LBX Pro’s. Lock-up: S-1 terms as reported, to confirm in the final prospectus.

Victoria Arinstein

Product & Intelligence, LBXpro

More insights

All insights
LBXpro

Stay in the loop

Get private market insights and platform updates delivered to your inbox.

Privacy Policy.

Office

Building A1, Dubai Digital Park, Dubai Silicon Oasis, Dubai, United Arab Emirates

Nothing on this site and the investment platform is intended as an offer to purchase or sell securities or a solicitation or recommendation of our securities transaction. Any financial information presented on the site and the investment platform are opinions, were prepared without taking into account your objectives, financial situation or needs. Investment results are not guarantees of future results.

Leo

AI analyst

Ask anything

Companies, valuations, IPO plans, market activity