Anthropic’s Profits Could Change How Investors Look at AI Stocks

For the past two years, the story around AI companies has followed one pattern: huge growth, huge losses. The common worry has been simple. These companies are burning cash at a pace no business model can sustain, and at some point, the money runs out.

Anthropic just gave that worry a reason to pause.

On September 13, 2026, the Financial Times reported that Anthropic, the maker of the Claude AI models, told a small group of shareholders it expects positive adjusted operating income for a second straight quarter. The first profitable quarter was April to June 2026. The second is the current quarter, July to September.

Two profitable quarters in a row from a company that made just $787 million in revenue one year earlier, and is now on track for over $11 billion in a single quarter, is not a small story. It may be one of the most significant financial developments the AI industry has seen so far.

The Revenue Number, First

Anthropic’s revenue growth is hard to overstate.

In Q2 2025, a year earlier, Anthropic generated $787 million in quarterly revenue. By Q1 2026, that had climbed to $4.73 billion. By Q2 2026, preliminary figures reported by Bloomberg showed it had surged past $11.5 billion. That’s more than a 14x increase in a single year.

There’s another way to see how fast this has moved. Anthropic’s annualised run-rate revenue, a projection based on the most recent month’s numbers, stood at roughly $9 billion at the end of 2025. By July 2026, that run-rate had jumped to more than $65 billion, a sevenfold increase in about seven months.

For scale, $11.5 billion in a single quarter works out to a run rate of over $46 billion a year. Infosys, India’s second-largest IT services company, crossed the $20 billion mark in full-year revenue for fiscal 2026. Anthropic is now generating, in a single quarter, revenue that comes close to half of what one of India’s most established technology companies makes in an entire year.

The growth is overwhelmingly enterprise-driven. More than 300,000 businesses now use Claude tools, and enterprise and business clients account for roughly 80% of total sales. Claude Code, Anthropic’s AI coding assistant that launched in 2025, has grown into a major revenue driver in its own right, crossing $2.5 billion in annualized revenue by February 2026, with later reports pointing to a multiple of that figure by mid-2026. The business case for AI in the enterprise is clearly no longer theoretical.

The Gross Margin Story

Alongside the profitability claim, the FT reported that Anthropic’s gross margins sit above 80%, before accounting for two large cost items: the revenue it shares with distribution partners like Amazon, and the cost of training its AI models.

An 80% gross margin, even before those costs, puts Anthropic in the range of mature software companies. Salesforce and Adobe, two of the highest-margin businesses in tech, operate in this range too, after decades of building out their businesses.

Anthropic reaching these margins after just a few years says something about how AI model deployment actually works economically. Once a model is trained and deployed, the cost of serving each query is low relative to what enterprise customers pay for access. When the product works, and businesses are willing to pay, the unit economics of AI at scale look genuinely strong.

This is why the profitability news matters beyond one company’s results. It starts to answer the question that has hung over the entire AI industry: can these businesses actually make money at scale?

The Asterisk: What “Adjusted” Leaves Out

This is the part of the story that deserves just as much attention, and financial media has rightly focused on it.

The profit being claimed is adjusted operating income, a metric that excludes specific costs. In Anthropic’s case, those exclusions are significant. Stock-based compensation, which represents real economic value transferred to employees and a real cost to shareholders, is not included. Revenue shared with distribution partners like Amazon, which powers a large share of Claude’s distribution through AWS, is also excluded from the 80% gross margin figure. And the cost of training the underlying models, which can run into hundreds of millions or billions of dollars per major release, is left out too.

Short seller Jim Chanos, known for flagging questionable financial statements at major companies, responded to the FT report on X with two words: “Cost-Adjusted EBITDA.” The message was clear: this metric conveniently strips out the biggest and most structural costs of running the business.

Oracle’s Matt Asay was more direct, calling Anthropic’s profitability claim a “made-up accounting fiction” and arguing the company “weren’t and aren’t” actually profitable. The criticism isn’t that Anthropic is lying. It’s that this specific definition of profitability makes the business look considerably better than a full, GAAP-based accounting would show.

None of this means the underlying business is weak. It means investors should be asking the right question: not “is Anthropic profitable,” but “how profitable would it be with all real costs included, and how does that change as training costs shrink as a share of revenue over time?”

The IPO Context: Partly a Perception Exercise

The timing of the FT report, and Anthropic’s decision to share profitability signals with a small group of shareholders right now, is not a coincidence.

Anthropic is preparing for a potential IPO. Business Insider reported around the same time that Anthropic has chosen Nasdaq as its listing venue. Bloomberg separately reported the company could be valued at $2 trillion or more in that offering.

Anthropic’s own disclosed annualised run-rate had already passed $65 billion by July 2026. At a $2 trillion valuation, that works out to roughly 30 times revenue, a steep multiple, though one that looks somewhat less extreme once you use the company’s full run-rate rather than just a single quarter annualized. Either way, for that kind of valuation to hold up in public markets, Anthropic needs more than growth. It needs a credible path to real, sustained profitability.

Sharing the adjusted operating income figures with investors ahead of a potential IPO is partly about building that credibility. It signals that this is a business that can eventually stand on its own financially, not one that will always need fresh capital to survive.

There’s another layer worth noting. The day before the FT report, Anthropic CEO Dario Amodei published an essay calling on the AI industry to slow the pace at which it improves AI capabilities. Landing one day before a major profitability announcement, the timing drew mixed reactions. Some commentators called it self-serving positioning ahead of a landmark IPO. Others said there was no clear evidence the timing was deliberate. Either way, the coincidence was hard to ignore.

What This Means for Indian Investors?

This story connects to Indian investors in more ways than it might first appear.

The clearest link is through the Indian IT services sector. Infosys, TCS, Wipro, and HCL Tech have all been building AI practices, including partnerships with Anthropic’s Claude and competing models. How profitable AI is at enterprise scale directly affects the services opportunity these companies are chasing.

If Anthropic’s 80%-plus gross margins reflect genuine unit economics, then the revenue opportunity for IT services firms helping enterprises adopt these models looks real and durable, not a passing or transitional trend. The fear that AI would shrink IT services revenue more than it creates new work looks weaker if the underlying AI economics are this strong.

Amazon’s deep involvement with Anthropic, as both a major distribution partner and an investor, matters here too. A large share of Claude runs on AWS for enterprise customers. The more revenue Anthropic generates through AWS, the stronger the signal for cloud infrastructure demand, which in turn supports the Indian IT companies that have built major AWS practices.

The broader debate over an AI bubble versus a real AI business model also matters for how Indian investors think about AI-adjacent stocks. The typical skeptic’s argument has been that AI companies are burning cash unsustainably, and that the cycle will eventually reverse and drag down both AI companies and the Indian IT sector with it. Anthropic’s profitability claim, asterisks and all, weakens that argument.

If leading AI companies are genuinely moving toward profitability, the AI investment cycle is more likely to continue than collapse. That keeps enterprise AI adoption on track, and keeps the opportunity alive and growing for Indian IT services firms.

For Indian investors weighing technology sector allocation, the Anthropic story is a signal that the AI transformation of enterprise technology isn’t a hype cycle racing toward a crash. It’s a revenue-generating, margin-delivering business reality, and one still in its early years.

In Light of These Points

Anthropic claiming back-to-back profitable quarters doesn’t end the debate over AI profitability. The adjusted nature of the metric, real questions about what full-cost accounting would show, and the IPO motivations behind the timing all mean this story deserves careful reading, not uncritical celebration.

But the revenue trajectory, from $787 million a year earlier to over $11 billion in a single quarter, and gross margins above 80%, points to something real. The AI business model, at scale, with enterprise adoption behind it, works. It generates revenue. It has the makings of genuine economic returns.

For investors tracking technology, AI, and the Indian IT sector, this is the most important piece of financial data to come out of the AI industry this quarter.

At Bonanza Wealth, we track global technology and AI sector developments because they shape the outlook for Indian IT companies, cloud infrastructure plays, and technology-adjacent investments that many of our clients hold. If you want help thinking through how Anthropic’s financial trajectory, and the broader AI business model story, should factor into your portfolio, our team is here to walk through it with you.

Blog Disclaimer:

The stocks, companies, or financial instruments mentioned in this blog are for informational purposes only and should not be considered as investment recommendations. It is advised to consult with your financial advisor before making any investment decisions. Investment in securities markets are subject to market risks, read all the related documents carefully before investing. Investors are strongly encouraged to carefully read the risk disclosure documents prior to participating in market-related investments or trading activities. Due to the volatile nature of financial markets, no guarantees can be made regarding investment returns. Bonanza Portfolio  Ltd. does not offer any assured returns on market-linked securities. Please note that past performance of stocks or indices is not indicative of future results.

You can also join our community at LinkedIn, Instagram, and Twitter to stay updated.