Anthropic was drumming up support for its IPO last week with investor roadshows, as it sought to reassure them of its profit making potential. Targeting a valuation $1.8Tn to $2.0Tn, it marks a watershed moment for Frontier AI capital economics, and the next era of tech investment.
It moves GenAI from private venture funding into the public market, with greater financial visibility and scrutiny. As the first Frontier AI lab to step into the public arena, the IPO serves as a stress test for the AI ecosystem. It forces a fundamental question: are valuations justified by the underlying revenue acceleration and productivity gains of foundation models, or are investors absorbing a bubble of unsustainable capital obligations?
Anthropic’s numbers are bonkers. Revenue surged by 45% to $16.8bn in the three months to the end of September, and in common with other AI firms, it is seeking to demonstrate it can turn this revenue to generate significant, sustainable profits despite spending ongoing billions on tech development and infrastructure. They announced a Q3 profit on an adjusted basis, an accounting method which ignores non-recurring costs, and investors left optimistic about the path to profitability.
It’s a bet, and a big one with potential ripple effects of some magnitude. By analysing the underlying drivers; of Anthropic’s listing we can unpack what this milestone reveals about AI sentiment and the immediate future of tech investment.
Anthropic’s trajectory to IPO is unprecedented. It illustrates how rapidly Frontier AI scaled beyond the capacity of traditional venture capital, and marks the end of the traditional VC-to-IPO pipeline. The volume of capital required to build Frontier models has altered the fundraising ecosystem. Along with OpenAI, they are bellwethers of the AI industry’s growth and ability to finance their infrastructure needed to develop the tech.
Dario Amodei admitted the IPO marks a shift in how future investors will evaluate AI. During the initial AI boom, investors accepted low or negative gross margins, treating GPU infrastructure overheads as customer acquisition costs. Anthropic disclosed divergent margins and some interesting financial drivers within its different business lines:
- Raw token APIs (low margin/high volume): Price competition and commoditisation press margins toward 20%–35%.
- Workforce automation & coding agents (high margin): End-to-end task execution platforms like Claude Code command SaaS like gross margins of 70%, because pricing is tied to labour value replacement rather than per-token consumption.
- Enterprise integration over consumer chatbots: While initial AI excitement was driven by consumer chatbots, institutional investor appetite focused on its enterprise foundation of embedded, security focused workflows.
- The hyperscaler circular economy: A key focus has been the degree of circular revenue between AI labs and big tech hyperscalers. Multi-billion dollar equity investments by cloud providers tied to compute purchase agreements prompted close examination.
- The capital commitment trap: Carrying over $500bn in non-cancellable compute obligations creates a high fixed cost floor. If enterprise spending slows or token pricing drops faster than compute efficiency improves, margins could compress.
- Open-Source and Frontier model convergence: The closing capabilities gap between proprietary models and open-weight alternatives continues to put pressure on API margins, accelerating the shift toward proprietary, agentic software platforms.
- Regulatory scrutiny and safety: As a public entity subject to national security oversight, export controls, and regulatory compliance, Anthropic’s operational flexibility will face ongoing public evaluation.
The market’s appetite for Anthropic and AI infrastructure has drawn capital away from non-AI tech startups. Mid-stage and late-stage SaaS companies without a compelling AI platform layer face compressed multiples, as institutional funds consolidate capital into foundational AI providers. Anthropic’s move to the public market tests whether public markets can absorb unprecedented capital requirements in exchange for a foothold in the future of computing.
Some of the disruption comes from rivals. In January a Chinese startup called DeepSeek came out of nowhere with a model that was almost as clever but required many fewer chips to train and use it. DeepSeek also made its code freely available to all, lowering barriers to entry into advanced model making. Competition is, of course, tech’s Schumpeterian lifeblood. Nothing stops Anthropic from making its models more efficient, including adopting some of DeepSeek’s ideas. The trouble is that the economics are changing in more fundamental ways too.
Will Anthropic ever make real money? Dario Amodei assured investors they would. Currently he is the darling of both the starch collared Davos set and Silicon Valley’s dishevelled techno-Utopians. Moneymen swept up in the AI mania need little persuading. What excites investor backing is the pace of AI innovation. It would be one thing for advances to be frequent, but the challenge for them and Anthropic is that they also frequently upend their economic model.
As for any startup, making money involves a series of steps: attract and retain brainboxes, have them create something clever, turn that something into a marketable product, sell more and more of that product until cash flow turns positive. Despite some high profile departures, Anthropic remains a talent magnet. The cleverness of its tech is indisputable. Dario Amodei’s claim that the latest models, with an enhanced ability to reason, display genius-level intelligence should be taken with a pinch of salt – but only a pinch.
This year it expects revenue to triple again, and inference costs to grow at the same rate. A shifting cost structure makes it hard to price products and plan budgets. A fixed subscription fee that once made sense looks unviable. You could retain operate a hybrid pricing strategy of subscriptions for core use and add a variable usage fee for inference heavy reasoning. But how long until the next model forces another rethink?
Any projections for revenue and costs beyond the next few months rest on brave modelling assumptions. Revenue and cash forecasts may as well be pulled out of a hat – not because it is too optimistic, but because it feigns certitude. This uncertainty may, perversely. make it easier to raise more capital. Yet it also highlights the uncertainty around what the company is truly worth and the scale of the challenge.
For tech founders, the takeaway is clear: the initial wave of speculative AI investment has given way to an era defined by unit economics, gross margins, and tangible financial returns. Anthropic’s IPO establishes a standard for evaluating AI. It also resets the agenda for the sentiment on capital deployment in the tech sector. Here are some thoughts on the impact on funding for founders looking to raise in Manchester.
1. Leverage strategic enterprise backing early. Anthropic scaled rapidly by tying up with cloud and platform providers who acted as both investors and distribution channels.
- Manchester thinking: Startups shouldn’t just look to local angel networks or VCs. Build strategic commercial partnerships with major enterprise anchors, and global tech vendors who can invest and provide your initial enterprise contracts.
2. Prepare for scrutiny on unit economics. Anthropic faces sharp investor scrutiny over high compute costs relative to revenue growth.
- Manchester thinking: Investors seek capital efficient growth. If your startup is building resource heavy tech, show a clear path to managing infrastructure overhead and achieving sustainable GM rather than burning cash blindly for top-line growth.
3. Build IP. High value tech companies succeed because of proprietary research and deep technical moats that competitors cannot replicate, creating genuine moats from their products.
- Manchester thinking: Manchester has a massive advantage with a world-class University. Bridge the gap between academic research and commercial funding early. Startups that leverage research attract premium valuations from investors.
4. Structure for insane ambition. Market leaders target expansion and scale rapidly, treating local markets as testing grounds. They immediately look to a long-term horizon.
- Manchester thinking: While Manchester provides a cost-effective base compared to London, your thinking and product roadmap should clearly articulate how you plan to scale past the UK into European and global markets.
5. Capitalise on the regulatory shift toward talent mobility. Regulatory environments are shifting to make it easier for tech talent to move freely and join high-growth startups, such as recent UK policy moves to curb restrictive non-compete clauses.
- Manchester thinking: Use your location as a selling point. Manchester has a massive local talent pool from regional universities and digital agencies. With reduced hiring friction and non-compete constraints, use equity incentives to poach top engineering talent away from legacy firms.
6. Diversify revenue. Anthropic’s S-1 document disclosure highlights a vulnerability: deep reliance on just one or two major cloud vendors for compute and sales channels.
- Manchester thinking: Ensure your startup maintains commercial independence. Build your tech stack with multi-cloud flexibility so that enterprise customers and future investors know your business isn’t overly fragile or at the mercy of a single partner’s pricing changes.
7. Position Manchester’s cost advantage as an extended runway. High burn rates shorten a startup’s lifespan, especially during market downturns or tightening IPO windows.
- Manchester thinking: Capital efficiency is a key driver. Operating out of Manchester typically means lower overheads (office space, salaries, and cost of living) compared to London or the US. Highlight to VCs that their investment buys a longer runway, allowing you to reach critical product milestones with less dilution.
8. Build compliance and trust into your product early. As tech companies scale toward public markets, regulatory compliance, data privacy, and safety become major risk factors flagged.
- Manchester thinking: bake robust security, ethics, and compliance frameworks into your MVP. Enterprise buyers (and institutional investors) will pay a premium for solutions that are audit-ready.
9. Master narrative-driven fundraising. Anthropic’s valuation is driven as much by a Dario Amodei’s compelling narrative and vision of the future as they are by potential financials.
- Manchester thinking: Manchester founders sometimes suffer from being overly modest and understated about their tech. When pitching national or international VCs, frame your local Manchester innovation within a grand, macroeconomic narrative. Show how your solution solves a massive global bottleneck – not quite with the bravado of Liam Gallagher though!
10. Leverage mentors with a ‘Made in Manchester’ exit. Tech ecosystems thrive because successful founders recycle their wealth, knowledge, and angel capital back into the next generation of startups.
- Manchester thinking: Manchester has an enviable heritage of entrepreneurial endeavour in its tech community. Successful local founder exits and scale-ups provide a rich source of mentorship. Seek out and learn from those who have already navigated seed-to-Series B journeys to avoid common scaling pitfalls.