Just three months after Elon Musk’s SpaceX (NASDAQ:SPCX) re-wrote the IPO playbook, UK retail investors have another potential multi-billion-dollar IPO on the cards… Anthropic. The company behind AI engine Claude could become one of the most important stock market events of 2026.
Anthropic has confidentially submitted a draft listing document S-1 to the SEC. It has not yet set the number of shares, offer price or firm listing date, although reports suggest a September or early-October 2026 debut could be targeted.
That puts Anthropic alongside SpaceX and OpenAI in an extraordinary wave of mega-IPOs. But SpaceX has already provided an important warning: a great technology story does not automatically make a great IPO investment.
SpaceX: Does a $2.1 trillion valuation make sense?
What is Anthropic’s business model?
Founded in 2021 by former OpenAI researchers and executives, and led by CEO and co-founder Dario Amodei, Anthropic develops Claude, an AI model competing with ChatGPT, Microsoft’s Copilot and Google’s Gemini.
Its business is increasingly enterprise-focused:
| Business | How Anthropic makes money |
| Claude.ai | Consumer subscriptions |
| Claude Enterprise | Per-seat subscriptions and usage |
| Claude API | Customers pay according to AI usage |
| Claude Code | AI-powered software development |
| AI agents | Automated business workflows |
Claude is also distributed through hyperscalers AWS, Google Cloud and Microsoft Azure, allowing enterprises to access it through existing cloud relationships.
The investment case is therefore less about selling a chatbot and more about becoming an AI platform embedded in corporate workflows.
Extraordinary revenue growth
Anthropic’s reported revenue trajectory is the strongest argument for the IPO. Annualised run-rate revenue reportedly exceeded $47bn in May 2026, compared with about $9bn at the end of 2025.
Annualised revenue run-rate*
*Run-rate revenue is not the same as reported annual revenue; it annualises the company’s current revenue pace.
Around 80%–85% of the business has reportedly been enterprise/developer-related rather than primarily consumer subscriptions.
That matters because Anthropic is increasingly selling AI as infrastructure for businesses.
Claude Code could be the key product
Investors should arguably focus as much on Claude Code as the Claude chatbot.
Claude Code is an AI coding agent capable of undertaking software development tasks. Software development is a high-value enterprise activity: if an AI tool saves a company substantial developer time, customers have a strong economic reason to keep using it.
The potential flywheel is:
Better models → more useful agents → more enterprise usage → more revenue → more compute → better models
If that works, Anthropic could become an important layer of the enterprise software stack.
Profitability is improving — but free cash flow is crucial
Anthropic’s economics have improved rapidly. Estimates suggest API (application programming interface) gross margins are already above 80%, while overall gross margins have been around the mid-60% range.
Reports have also suggested Anthropic could reach its first quarterly operating profit during 2026.
| Metric* | Current/indicative position |
| Revenue run-rate, May 2026 | >$47bn |
| Q1 2026 revenue | ~$4.8bn |
| Q2 revenue target | ~$10.9bn |
| Q2 operating profit target | ~$559m |
| Estimated Q3 EBIT | >$1bn |
| Overall gross margin | ~mid-60% |
| API gross margin | 80%+ estimated |
*to 31 Dec fiscal year end
But investors should not confuse gross margin, operating profit and free cash flow.
Anthropic is not a conventional software company. Every additional Claude query consumes computing resources, including GPUs, electricity, networking and cloud capacity.
The company has signed enormous infrastructure commitments, including additional AWS capacity and next-generation Google TPU capacity.
The bull case is that revenue grows faster than compute costs.
The bear case is that AI becomes increasingly commoditised and Anthropic will have to spend more simply to maintain its position.
Valuation is already enormous
Anthropic’s May funding round reportedly valued the company at $965bn.
Against a $47bn annualised revenue run-rate, that represents roughly 20.5x sales.
| Measure | Approx. figure |
| Latest private valuation | $965bn |
| Revenue run-rate | >$47bn |
| Valuation/revenue run-rate | ~20.5x |
| Possible IPO valuation | ~$1tn-$2tn |
According to the Financial Times, the company could seek a valuation $2tn+, SpaceX hubris perhaps. At these levels, investors are not paying for today’s Anthropic. They are paying for years of exceptional future growth.
The opportunity
Enterprise AI: Claude could become embedded in software development, finance, healthcare, professional services and other corporate workflows.
Exceptional growth: A $47bn revenue run-rate is substantial, even allowing for the distinction between run-rate and reported revenue.
AI agents: Claude Code and other agents could dramatically increase customer spending as AI moves from answering questions to completing tasks.
Strategic backing: Amazon, Google, Microsoft and major institutional investors including several Baillie Gifford managed funds – for example, Scottish Mortgage (LON:SMT), Baillie Gifford US Growth Trust (LON:USA), The Schiehallion Fund (LON:MNTN) – are deeply connected/invested in and to Anthropic’s ecosystem.
Multiple revenue streams: Anthropic can monetise consumers, developers, enterprises, APIs, coding and AI agents.
The risks
Valuation: At roughly 20-21x current sales, the shares leave relatively little room for disappointment.
Commoditisation: Google, OpenAI, Meta, xAI and open-source developers could drive model prices lower.
Compute economics: The more successful Claude becomes, the more computing capacity Anthropic may need.
Customer concentration: Reliance on major technology companies for infrastructure and distribution gives those partners significant bargaining power.
Regulation: Anthropic’s approach to AI safety and military applications could create political and government contract risks.
Dilution: Investors should scrutinise employee stock compensation, new share issuance and lock-up arrangements in the S-1 document.
What SpaceX teaches investors
SpaceX’s IPO offers perhaps the clearest warning.
The company raised a record $75bn in its June IPO at $135 a share. Predictably, the stock initially surged but subsequently fell below the IPO price.
Its first public earnings report showed revenue up 92% year-on-year to $7.8bn, yet investors focused on the scale of AI investment. Q2 capex was reportedly around $18.4bn, including approximately $15.8bn of AI-related spending, triggering a sharp share-price decline.
SpaceX Q2 2026 earnings: Strong results fail to justify lofty valuation as shares tumble
| SpaceX lesson | Anthropic equivalent |
| Huge revenue growth | Huge AI revenue growth |
| Strong technology | Leading AI models |
| AI opportunity | AI agents and enterprise AI |
| Huge capex | Huge compute commitments |
| Cash-flow concerns | Compute and infrastructure costs |
| Scarce IPO | Potentially huge IPO demand |
| Strong company | But valuation still matters |
The lesson is simple: don’t confuse the quality of the company with the attractiveness of the stock.
Bull vs bear case
🐂 Bull case
Anthropic becomes the enterprise AI operating layer.
Claude becomes deeply embedded in software development and corporate workflows. Revenue continues compounding at 50%+, inference costs fall faster than pricing, gross margins approach software-like levels and free cash flow turns positive.
At that point, a $1tn valuation could look reasonable.
🐻 Bear case
Frontier AI becomes a rapidly commoditising market.
Google, OpenAI, Meta, xAI and open models drive prices lower. Revenue growth slows, Anthropic continues signing huge compute commitments and margins fail to reach traditional software levels.
If investors then reduce the sales multiple from 20x to 10x, the shares could fall sharply even if revenue continues growing.
What UK retail investors should watch in the S-1
The IPO prospectus will matter more than the headlines. Focus on:
- Actual 2025 and 2026 revenue
- Revenue growth
- Gross margin
- Inference cost per dollar of revenue
- Operating cash flow
- Free cash flow
- Contractual compute commitments
- Customer concentration
- Stock-based compensation
- Share dilution and IPO valuation
The key issue isn’t ‘Is Claude better than ChatGPT?’
Instead,the crucial question is:
How much cash does Anthropic retain from every additional $1 of Claude revenue?
Investor verdict
Investors have another potential multi-billion-dollar IPO on the cards and Anthropic could be one of the most compelling technology businesses to reach public markets in years. Its extraordinary revenue growth, enterprise focus and improving margins make it very different from an unproven AI start-up.
But at a $1tn-$2tn valuation, investors would be paying today for a huge amount of tomorrow’s growth.
SpaceX has shown how quickly sentiment can change when investors focus on capital spending and returns rather than exiting technology and headline growth.
Of course, pre-IPO analysis may be moot for UK retail investors – we have no idea if there will be a retail offering, or if that will extend to overseas investors. Even so, the sensible approach may be to admire the company but scrutinise the IPO price and how that compares to a realistic returns profile.
Wait for the S-1, calculate valuation from actual financials and don’t assume the IPO price represents fair value. The most attractive entry point could ultimately come after the initial IPO excitement has subsided, rather than before or on day one.
You might also like:
