Anthropic is asking prospective investors to look unusually far into the future as it prepares for a potential blockbuster initial public offering, with bankers and investors using the company’s projected 2028 revenue to assess how much the artificial intelligence developer could be worth.

The Claude maker is projecting revenue of roughly $190 billion to $200 billion in 2028, according to two people familiar with its financials cited by Reuters. The forecast represents an extraordinary jump from the more than $47 billion annualized revenue run rate Anthropic reported in May and illustrates the scale of growth investors would be expected to price into the company ahead of a potential public listing.

Rather than relying primarily on current earnings, investors and bankers are using enterprise-value-to-revenue multiples applied to future projections, according to four people familiar with the process. That approach is common for rapidly expanding software companies that have yet to establish mature profit margins, but using forecasts two years into the future highlights the unusual difficulty of valuing an AI company whose financial profile is changing at exceptional speed.

Anthropic’s latest projections come as the company attempts to establish itself as one of the leading commercial AI platforms while spending heavily on computing infrastructure, model development, and personnel. The central question for investors is whether revenue can continue expanding faster than the enormous costs required to support increasingly capable AI systems.

The company’s recent financial performance has strengthened that case.

Anthropic’s revenue run rate was about $9 billion at the end of 2025 before climbing to more than $47 billion by May 2026. The company has projected second-quarter revenue of at least $10.9 billion and was on track for its first quarterly operating profit of about $559 million, according to people familiar with its financials.

Separate documents seen by Bloomberg showed preliminary second-quarter revenue of more than $11.5 billion, compared with $787 million a year earlier and $4.73 billion in the first quarter. The figures remain subject to revision.

The pace of expansion has put Anthropic in a direct contest with OpenAI for corporate AI customers, particularly in areas such as software development and coding. Anthropic’s annualized revenue has surpassed $47 billion, while OpenAI’s annualized revenue has topped $40 billion, although the companies may calculate their run rates differently.

Investors Are Being Asked to Price The Future

The proposed valuation methodology highlights a fundamental problem facing investors in AI companies: traditional measures such as current earnings provide limited insight into businesses still spending aggressively to build their competitive position.

Anthropic is investing heavily in GPUs and other computing capacity, model training, inference, and hiring. Those expenses suppress current profitability, but investors are effectively being asked to assume that the costs will decline as a percentage of revenue as the company reaches greater scale.

The underlying investment thesis is straightforward. If Anthropic can continue adding customers and increase usage of its models while improving computing efficiency, revenue could grow substantially faster than operating costs. Higher utilization, more efficient models and declining computing costs could then allow margins to expand.

The risk is that the opposite could happen. AI companies are locked in an infrastructure race that requires enormous capital commitments, while competition from OpenAI, Google, Meta and Chinese developers could force companies to spend more to maintain technological leadership.

That makes the $190 billion to $200 billion 2028 revenue forecast particularly important. Investors are not simply evaluating what Anthropic is earning today. They are assessing whether the company can become a business capable of generating hundreds of billions of dollars in annual sales within a relatively short period.

Palantir, Cloudflare and SpaceX Emerge As Valuation Benchmarks

Anthropic is also looking for public-company comparisons that can help investors determine an appropriate revenue multiple. Cloudflare, Palantir and SpaceX are among the companies being considered as reference points ahead of Anthropic’s analyst day, according to people familiar with the process.

Each provides a different valuation framework.

Palantir has become an important benchmark for investors valuing companies with rapid growth and significant exposure to AI. The company trades at roughly 53 times expected 2026 revenue, according to LSEG data.

Cloudflare, meanwhile, offers a comparison with a high-growth software and internet infrastructure company and trades at about 41.6 times expected 2026 revenue. SpaceX also trades at roughly 41.6 times expected 2026 revenue, although its business mix and capital requirements differ substantially from Anthropic’s.

Applying those kinds of multiples to Anthropic’s projected 2028 revenue could produce an enormous valuation, potentially pushing the company into the ranks of the world’s most valuable businesses.

But the comparison also demonstrates the risk. High revenue multiples require investors to maintain confidence in exceptionally strong future growth. Any slowdown in customer adoption, pricing pressure, or deterioration in AI margins could cause those multiples to contract sharply.

Anthropic’s IPO Could Reshape the AI Market

The potential listing would be significant beyond Anthropic itself. Analysts have touted it to become one of the largest tests yet of whether public-market investors are willing to assign extraordinary valuations to AI companies based largely on future scale.

The IPO would also arrive as investors have become more sensitive to the enormous capital requirements of the AI boom. Heavy spending on data centers, GPUs and electricity has contributed to concerns that the financial returns from AI could take longer to materialize than expected.

There have already been precedents for investors looking well beyond a company’s current earnings when valuing high-growth AI businesses. Backers of Cerebras Systems cited 2028 revenue projections ahead of its IPO, while SpaceX investors considered forecasts extending to 2029 before its public debut.

Anthropic’s case, however, is unusually large because the projected revenue base is so substantial. The company is also preparing for a potential IPO before OpenAI, while Chinese AI company DeepSeek is reportedly considering a listing as well. That could turn the next phase of the AI race into a competition not only for customers, computing power and talent, but also for public-market capital.

For investors, the central question will be whether Anthropic’s extraordinary growth can translate into durable economics.

“Could they (Anthropic) get a $2 trillion valuation, yeah they could and I just wonder if it would stay there over time,” said David Merkel, a principal at Aleph Investments.

The bigger issue, he said, is whether AI will generate enough additional productivity to justify such valuations.

That question is likely to sit at the center of Anthropic’s IPO. The company’s projected $190 billion to $200 billion in 2028 revenue may demonstrate the size of the opportunity, but investors will ultimately have to decide how much of that future growth is already embedded in the price.

Berkshire Hathaway made a roughly $17 billion investment in Alphabet during the second quarter, elevating Google’s parent company to its third-largest equity holding and giving investors a clearer view of how Warren Buffett’s successor, Greg Abel, is deploying the conglomerate’s enormous cash reserves.

Berkshire’s latest portfolio filing with the U.S. Securities and Exchange Commission showed the company owned nearly 106 million Class A and Class C Alphabet shares as of June 30. The position was worth about $36.6 billion based on current prices, placing Alphabet ahead of Coca-Cola at roughly $35.1 billion but behind Apple at $69.7 billion and American Express at $51.9 billion.

Berkshire added about 48.1 million Alphabet shares during the quarter, making the technology company by far its largest new investment in the period.

About 60% of those shares, or roughly 29 million, were acquired directly from Alphabet as part of a $10 billion private placement announced by the companies in early June. The remaining shares indicate Berkshire bought approximately $7 billion of Alphabet stock through the open market.

The scale of the purchase represents a significant shift in Berkshire’s equity portfolio, particularly because Alphabet had not previously ranked among the conglomerate’s largest disclosed holdings.

The investment also gives Berkshire greater exposure to the technology and artificial intelligence spending cycle. Alphabet operates Google, the world’s largest search engine, as well as its growing cloud business and Gemini artificial intelligence products. The company has been investing heavily in AI infrastructure and models as competition intensifies across the technology industry.

Alphabet was not Berkshire’s only notable addition during the quarter.

According to CNBC’s Warren Buffett’s Watch, Berkshire increased its holding in Delta Air Lines by 44%, adding roughly $1.6 billion to the position. The conglomerate now owns about 57.3 million Delta shares, worth approximately $5.1 billion.

The Delta investment is also notable because Buffett has historically had a difficult relationship with the airline industry. Berkshire initially invested in Delta and several other major airlines before selling its airline holdings at a loss during the first quarter of 2020 as the COVID-19 pandemic caused global air travel to collapse.

Buffett had previously expressed deep skepticism about airlines. In his 2007 shareholder letter, he joked that if a “farsighted capitalist” had been present at Kitty Hawk when the Wright brothers made their first flight, that investor would have benefited from preventing the flight.

Delta returned to Berkshire’s portfolio in the first quarter of this year, marking a renewed bet on the airline after Buffett’s earlier exit.

Berkshire also increased its Macy’s holding by 142%, although the relatively small size of the position meant the increase was worth only about $100 million.

The company added approximately $280 million to its investment in homebuilder Lennar during the same quarter in which Berkshire announced its $6.8 billion acquisition of homebuilder Taylor Morrison Home.

The latest filing also showed that Berkshire continued to reduce its exposure to financial companies, extending a selling trend evident in recent quarters.

Berkshire cut its stake in Ally Financial by 7% and reduced its Capital One position by 58%. Its Bank of America holding fell 5.9%. Because Berkshire’s Bank of America position remains one of its largest investments, the relatively modest percentage reduction translated into a decline of about $1.7 billion in the value of the stake, making it the company’s largest dollar reduction during the quarter.

Berkshire has now reduced its Bank of America position by 53% following eight consecutive quarters of selling. The portfolio changes provide an early indication of the investment priorities emerging under Abel, who succeeded Buffett as Berkshire’s chief executive.

That transition has attracted the attention of Michael Burry, the investor known for his successful bet against the U.S. housing market before the 2008 financial crisis. Burry criticized the way Berkshire has been deploying its cash reserves, arguing that Buffett’s successor may not possess the same willingness to wait for unusually attractive investment opportunities.

In a Substack post, Burry said his “biggest fear” had been that Buffett’s successor would not have the legendary investor’s “patience for the fat pitch.”

“Now, I believe this fear has come true,” Burry wrote.

Burry said he no longer considers Berkshire an attractive investment, although he acknowledged that the company still has an enormous cash position. Berkshire had roughly $360 billion in cash remaining, according to his comments.

His concern is less about the absolute amount of money Berkshire has spent than about the nature of the company’s initial moves under Abel.

“I believe this fear has come true,” Burry wrote, arguing that Abel’s “first steps look to be more framing moves than investment moves.”

Burry later clarified that he was not recommending that investors short Berkshire shares.

The debate goes to the heart of Buffett’s investment philosophy. For decades, Berkshire’s strategy has emphasized patience, allowing Buffett and his investment managers to hold cash until opportunities emerged at prices they considered sufficiently attractive.

Buffett frequently compared investing with baseball, arguing that investors have no obligation to swing at every opportunity.

“What’s nice about investing is you don’t have to swing at pitches,” Buffett told The New York Times in 2007. “You can wait for the pitch you want.”

Berkshire’s decision to invest heavily in Alphabet suggests that the company is willing to deploy significant capital into large technology businesses when management identifies what it considers an attractive opportunity. At the same time, the reduction in financial holdings indicates a meaningful rebalancing of the portfolio. The combination of the Alphabet purchase, the Delta increase, and continued sales of financial stocks marks a significant shift from the portfolio positioning Berkshire had maintained in recent years.

Investors have already reacted cautiously to the changes. Both classes of Berkshire shares fell more than 3% during the week following the company’s second-quarter spending disclosures, even as Berkshire resumed significant share buybacks for the first time in two years.

The market response suggests that some investors are questioning whether Abel’s early capital-allocation decisions will match the standards established during Buffett’s tenure.

The Alphabet investment, however, offers an important test of that transition. Berkshire is committing billions to one of the world’s largest technology companies at a time when AI is reshaping the technology sector and driving enormous capital spending.

JPMorgan Chase terminated its direct banking relationship with prediction-market operator Polymarket in October 2025 over regulatory concerns, according to a person familiar with the matter, highlighting the growing compliance risks banks face as event-based trading platforms expand into mainstream finance.

The decision, first reported by the Financial Times and confirmed by Reuters, did not amount to a complete break between the two companies. Polymarket said it continues to maintain relationships with multiple JPMorgan entities, including operational integrations and arrangements involving customer fund flows.

“We maintain a close, active relationship with JPMorgan across multiple entities, operational integrations, and material handling of customer fund flows,” a Polymarket spokesperson said.

“The strength of our relationship is highlighted by our CEO speaking at three of their flagship events in the past year alone. Any suggestion otherwise fundamentally mischaracterizes our relationship,” the spokesperson added.

JPMorgan’s decision concerned its banking relationship with Polymarket, rather than an across-the-board termination of all commercial dealings. Reports indicate that Polymarket subsequently found another banking partner, although the identity of that institution has not been disclosed. JPMorgan has also continued other interactions with the company.

The development comes as prediction markets have expanded rapidly since the 2024 U.S. presidential election, attracting users interested in trading contracts tied to elections, sports, economic indicators and other real-world events.

The growth has placed the industry at the intersection of financial markets, gambling, and cryptocurrency, creating a difficult regulatory environment for both the platforms and financial institutions that provide them with banking services.

Why Banks Are Becoming Cautious

Prediction markets allow users to buy and sell contracts whose value is linked to the outcome of an event. Supporters believe that the resulting prices can provide a real-time measure of collective expectations and market sentiment.

The regulatory question is whether some of these contracts amount to financial derivatives or unlawful gambling products.

Polymarket has already faced significant regulatory action. In 2022, the Commodity Futures Trading Commission ordered the company to pay a $1.4 million penalty and cease offering unregistered event-based binary options contracts to U.S. customers.

The company has since taken steps to establish a regulated U.S. presence, but prediction markets remain under scrutiny from regulators and lawmakers at both the federal and state levels.

That uncertainty creates a particular problem for banks.

Financial institutions must assess not only whether a customer has obtained a particular regulatory authorization but also the broader legal and compliance risks associated with processing its funds. A prediction-market platform can therefore become a difficult banking customer even while pursuing regulatory approval for its underlying business.

JPMorgan’s decision illustrates that distinction. Regulatory uncertainty can affect a company’s access to financial infrastructure independently of whether regulators ultimately allow it to operate.

Prediction Markets Face Growing Political and Legal Scrutiny

The banking decision comes as the prediction-market industry faces a wider regulatory challenge.

New York City Council Speaker Julie Menin this week accused major prediction-market companies of using predatory marketing practices that could exploit younger traders.

Separately, New York Attorney General Letitia James sued Polymarket rival Kalshi last month, alleging that its prediction-market operations violate state gambling laws.

The legal battles are significant because prediction-market operators are seeking to establish themselves as legitimate financial-market businesses rather than online gambling companies.

The distinction carries major consequences.

Analysts have noted that if the platforms are treated primarily as financial markets, they can potentially operate within the regulatory framework governing derivatives and exchanges. But if state authorities classify particular contracts as gambling, operators could face a different set of restrictions, licensing requirements and enforcement actions.

That uncertainty is particularly relevant to banks, which have extensive regulatory obligations of their own.

Polymarket’s continued expansion has made the relationship with established financial institutions increasingly important. The company has attracted significant investor interest as prediction markets have grown into a major new category of financial activity. At the same time, its business model remains exposed to regulatory disputes over which types of event contracts can legally be offered and to whom.

For JPMorgan, maintaining selected commercial ties while ending the direct banking relationship suggests a more nuanced approach than simply abandoning the company. The bank can continue engaging with Polymarket in areas where it is comfortable with the regulatory exposure while limiting direct involvement in banking services that may carry greater compliance risk.

However, Polymarket finding alternative banking arrangements demonstrates that access to financial infrastructure has become an important consideration as the company scales.

Prediction markets are increasingly being promoted as sources of information about market expectations, but the same characteristics that make them attractive to traders are creating concerns about gambling, market manipulation, insider information and consumer protection.

Those concerns are expected to become more significant as platforms expand beyond political events into sports, financial markets and other areas where large amounts of money can be wagered.

For now, the JPMorgan-Polymarket relationship illustrates the uneasy middle ground occupied by the industry. Prediction markets are attracting major investors, financial institutions and millions of users, yet banks remain sensitive to unresolved regulatory questions surrounding their business models.

The fact that JPMorgan ended one banking relationship with Polymarket while maintaining other commercial connections shows that Wall Street’s response is not necessarily a simple rejection of prediction markets. Instead, banks appear to be separating individual services and assessing the regulatory risk attached to each part of the relationship.