I’ve been watching portfolio moves for years, and every so often one lands that makes you stop scrolling. When Artemis co-founder Jon Ma casually mentioned he was about to unload his Bitcoin and Robinhood shares because roughly half his net worth now sits in Anthropic, it felt like one of those moments. The AI company had just closed a $93.5 billion initial public offering. Markets were still digesting the debut, and here was a well-known figure in the crypto space openly rotating out of digital assets into pure AI exposure.
Why One Investor Is Walking Away From Bitcoin Right Now
Ma’s post on X laid it out without much drama. He had already shifted about 50 percent of his net worth into Anthropic stock, ticker ANT on Nasdaq. His portfolio snapshot showed Anthropic dominating at 52.1 percent, with Nvidia trailing at 11.4 percent. The rest of the chart included smaller positions in names like Taiwan Semiconductor, Nubank, and Datadog, plus some fund holdings. Alphabet sat at a modest 2.7 percent. He planned to keep only five stocks after the sales: Anthropic, Alphabet, Meta Platforms, Nebius Group, and Nvidia.
That kind of concentration is not for the faint of heart. Most people I know still preach diversification like it’s gospel. Ma is betting the opposite. He sees Anthropic as the clearest winner across cybersecurity, coding, law, healthcare, drug discovery, and finance. In his words, the company will eventually become a $10 trillion business. He also expects markets to rotate hard out of cash and general bearishness toward AI application companies.
I’ve found that when smart people start speaking this confidently about one name, it pays to look closer at the numbers behind the confidence. Anthropic’s IPO details help explain why the timing felt right for him.
The Size And Structure Of Anthropic’s Public Debut
The roadshow originally pointed to 825 million shares priced between $70 and $78. The final prospectus expanded that base offering to 980 million shares at $83 each. That jump of roughly 18.8 percent came entirely from Anthropic itself. Existing stockholders kept their selling amount fixed at 23.65 million shares. The company supplied every extra share in the base deal, adding about $12.87 billion in gross proceeds before fees.
Underwriters then exercised their full option for another 147 million shares. Total shares sold reached 1.127 billion. Anthropic walked away with roughly $91.6 billion in gross proceeds from the IPO before costs. A separate $26 billion private placement closed alongside the public deal. Google and Nvidia sat among the seven buyers in that private round, paying the same $83 per share for Class N stock. The Nasdaq listing covers the Class A shares that retail and institutional investors can trade.
Those numbers are hard to ignore. Raising that much capital in one go changes the runway for any company, especially one competing in foundation models. Ma’s decision to concentrate so heavily suggests he believes the capital will translate into lasting advantage.
How The Stock Traded After The Opening Bell
The stock opened around $112 on Thursday according to available market data. By Friday’s close it settled near $88.30, still about 6.4 percent above the IPO price but roughly 21 percent below the opening print. Ma’s own screenshot captured a price of $88.91 around the same time. Retail investors poured more than $750 million into the name on debut day, the highest first-day turnover some tracking firms have recorded since 2012. Net retail buying ranked high, behind only a couple of other high-profile listings in recent memory.
Stabilization mechanics mattered too. The lead bank covered the additional shares through the underwriters’ option. Once that option was fully exercised, some analysts noted that a traditional source of post-IPO buying support had disappeared. The stock gave back most of its opening premium within a day, which is common for large offerings when early excitement meets real supply.
Still, the fact that the shares held above the IPO price after such a large float suggests genuine demand. Ma appears to have viewed the post-debut pullback as less important than the multi-year thesis.
Ma’s Longer-Term Revenue Expectations
A few days before the IPO closed, Ma updated his internal model. He now projects $800 billion in annual recurring revenue for Anthropic by 2030. That figure is lower than an earlier $1 trillion estimate, adjusted after external research firms revised their AI capital spending forecasts downward. Even at the reduced number, the model assumes Anthropic captures half of enterprise spending on tokens, the basic units models process when handling requests.
Those are ambitious assumptions. Capturing half of any large market is rare. Yet the applications Ma lists feel concrete. Software development teams already lean heavily on coding assistants. Cybersecurity firms experiment with models that spot anomalies faster than traditional tools. Healthcare and drug discovery groups test the same models for literature review and molecule suggestion. Legal and finance teams explore document analysis and risk modeling. If even a fraction of those use cases scale, the revenue path becomes more believable.
In my experience, investors who size positions this large usually have spent months stress-testing the downside. Ma did not present the move as a short-term trade. He framed it as a conviction bet on the company he believes will lead AI applications for years.
The Broader Context Of Bitcoin Selling
Bitcoin had already slipped below key levels in the days leading up to the IPO. Some market observers suggested that investors needing cash for Anthropic allocations were liquidating more liquid assets, including crypto. Ma’s comments lined up with that theory, though he focused more on expected returns from AI stocks than on any forced selling pressure.
The timing still raises questions. Bitcoin has long been treated by a certain group of investors as digital gold or a high-beta risk asset. Selling it to fund an AI position marks a clear preference for growth stories tied to enterprise software and model usage rather than monetary narratives. Robinhood shares faced the same fate in Ma’s plan. Both positions, once core enough to appear in his public chart, were headed for the exit so he could keep the portfolio lean.
Perhaps the most interesting aspect is how cleanly the decision separates two narratives that sometimes overlap. Crypto and AI often share the same speculative capital pools. Here an investor is choosing one over the other in a decisive way.
What A Concentrated Portfolio Looks Like In Practice
Keeping only five names creates real risk. Any company-specific setback at Anthropic would hit Ma’s net worth hard. At the same time, the remaining holdings offer some balance. Nvidia remains the dominant supplier of training chips. Alphabet and Meta continue to push their own models and cloud offerings. Nebius Group brings additional infrastructure exposure. The mix still tilts heavily toward AI infrastructure and applications rather than broad market coverage.
I’ve watched similar concentrated bets play out over the past decade. Some deliver outsized gains. Others force painful rebalancing when the thesis slips. Ma’s public chart and clear language suggest he accepts that trade-off. He is not pretending the portfolio is balanced. He is stating the opposite: this is the highest-conviction slice of the market he can find.
- Anthropic as the core application-layer bet
- Nvidia for continued hardware demand
- Alphabet and Meta for platform and advertising strength tied to AI
- Nebius for additional compute and infrastructure optionality
That short list leaves little room for traditional diversification. It also leaves little room for hesitation. Once the Bitcoin and Robinhood sales complete, the portfolio becomes a pure expression of one view of the next decade.
How Other Investors Might Interpret The Move
Not everyone will follow the same path. Many long-term Bitcoin holders view price dips as accumulation opportunities rather than exit signals. Others maintain strict rules against concentrating more than a certain percentage in any single name. Ma’s approach sits at the aggressive end of the spectrum.
Still, the public nature of the post creates a useful data point. When a co-founder of a company focused on portfolio tracking decides to move this aggressively, it signals where he believes the highest expected returns now live. Whether that view proves correct will take years to measure. In the near term, it adds another voice to the idea that capital is rotating toward AI application winners.
Retail flow into the IPO already showed strong interest. Institutional participation in both the public and private placements confirmed demand at the $83 price. The subsequent trading range between the IPO price and the opening spike gives later buyers a clearer window than the chaotic first hours.
Risks That Come With This Level Of Conviction
Concentration risk sits at the top of the list. A single regulatory setback, a major model failure, or slower enterprise adoption could pressure the stock for months. Competition remains intense. Other large players continue to pour capital into competing models and infrastructure. Capital spending forecasts have already been revised once, which is why Ma lowered his revenue target. Further revisions remain possible.
Liquidity in the new stock will improve over time, yet early days can still bring sharp moves in either direction. The full exercise of the overallotment option removed one source of buying support. Future lock-up expirations will introduce additional supply. None of these factors appeared to deter Ma, but they remain real considerations for anyone thinking about similar sizing.
I’ve always believed that the size of a position should match the strength of the research and the investor’s ability to withstand volatility. Ma’s public comments suggest he has done that homework. The rest of us still need to decide whether the same thesis fits our own risk tolerance and time horizon.
The Bigger Picture Of Market Rotation
Ma expects a broad shift out of cash and defensive positioning into AI application companies. That narrative has been circulating for months. Earnings reports from major cloud providers and chip makers have repeatedly highlighted rising demand for training and inference capacity. Enterprise software firms are embedding model capabilities into existing products. The question is no longer whether AI will matter, but which companies will capture the largest share of spending.
Anthropic’s ability to raise nearly $92 billion in one offering gives it substantial resources to compete on talent, compute, and go-to-market. The private placement participation from Google and Nvidia adds strategic relationships that could prove valuable. Whether those advantages translate into the market share Ma anticipates is the core of the bet.
For crypto markets, the episode serves as a reminder that capital is mobile. When a high-profile IPO appears with strong narrative force, some investors will reallocate. Bitcoin’s recent softness around the IPO period may partly reflect that dynamic, even if other factors also played roles.
Practical Takeaways For Everyday Investors
Most people should not mirror a 50 percent position in a newly public company. Position sizing has to match individual circumstances. What Ma’s move does illustrate is the value of clear conviction and the willingness to act when the numbers line up with the thesis.
Reviewing existing holdings against current opportunities remains useful. Asking whether capital locked in older positions could serve a higher expected return elsewhere is a healthy exercise. The answers will differ for every portfolio. The process of asking the question is what matters.
- Clarify the core thesis behind any large position
- Quantify expected returns against realistic downside scenarios
- Decide in advance how much concentration feels acceptable
- Monitor new information that could change the original view
- Keep enough liquidity for future opportunities or unexpected needs
Those steps sound basic, yet they prevent many common mistakes. Ma appears to have followed a version of them before publishing his plan.
Looking Ahead After The Listing
Anthropic now trades as a public company with substantial capital and high expectations. The stock’s early performance will be watched closely as a proxy for broader AI sentiment. Ma’s decision to increase exposure and reduce other holdings adds a visible data point to that conversation.
Whether Anthropic reaches anything close to a $10 trillion valuation remains an open question measured in years, not months. The intermediate milestones of revenue growth, enterprise adoption, and competitive positioning will determine the path. For now, one prominent investor has placed a very large share of his personal capital on the affirmative side of that question.
The rest of the market will keep watching the price action, the next earnings updates, and any further commentary from large holders. In the meantime, the simple fact that a crypto-adjacent founder chose to exit Bitcoin in favor of an AI pure-play tells its own story about where certain pockets of capital are flowing in late 2026.
Markets rarely move in straight lines. Conviction bets of this size rarely look comfortable in the short term. Yet the investors who size them carefully and hold through the noise sometimes capture returns that diversified portfolios struggle to match. Ma has made his choice public. The coming quarters will show how the underlying business performs against the ambitious model he has laid out.
For anyone tracking the intersection of crypto and AI capital, the episode offers a clear case study. Liquidity, narrative strength, and expected growth rates continue to pull money across asset classes. When those forces align as strongly as they did around this IPO, some portfolios will look very different afterward. Ma’s is already well on its way.
The decision to sell Bitcoin after years of crypto involvement is never casual. It reflects a calculated judgment that the risk-adjusted opportunity in Anthropic currently outweighs the one in digital assets. That judgment may prove correct or incorrect over time. What cannot be disputed is the clarity with which it was expressed and the size of the capital committed. In a market full of hedged language and small adjustments, such direct action stands out.
I keep returning to the portfolio chart Ma shared. Seeing one name at more than half of net worth forces a different kind of attention. It is easy to talk about conviction. It is harder to live with the volatility that follows. The coming months will test both the company and the investors who have concentrated around it. For those watching from the sidelines, the episode still provides useful information about how at least one experienced allocator is positioning for the next phase of AI development.
Capital continues to chase the strongest narratives. Right now, that narrative centers on companies building and deploying large models at scale. Anthropic’s successful raise and subsequent trading give it a prominent place in that story. Ma’s personal allocation decision simply makes the story more concrete for anyone following both crypto and public markets. The rotation is underway. How far it goes remains the open question worth tracking.