Anthropic achieved a breakout moment this week, as investors bet the start-up has cornered the market selling AI products to businesses, worth hundreds of billions of dollars in revenue.
The AI lab has kept a lower profile than rivals OpenAI, Google and Meta, which have focused on consumer-facing products, instead pitching its models as tools for developers and companies.
That strategy came into focus over recent days with the release of new software that spooked public markets. Anthropic also captured attention with a Super Bowl advert that took aim at rivals, eliciting a tetchy response from OpenAI chief Sam Altman.
The five-year-old San Francisco group has stepped into the spotlight as it puts the finishing touches on a roughly $35bn funding round at a $350bn valuation and takes steps towards a blockbuster initial public offering this year.
Anthropic grew from $1bn in annualised revenue at the start of last year to more than $9bn by the end of 2025. The company has guided investors that annualised revenue will exceed $30bn by the end of this year and continue on a steep trajectory from there, according to people familiar with its finances.
Anthropic’s traction with businesses, product focus and stable leadership mean it is increasingly seen as a safer long-term bet than OpenAI, according to 12 investors who spoke to the FT.
“Anthropic is a well-run company with a simple capital structure that’s just working,” said Mike Paulus, a billionaire former Andreessen Horowitz partner whose family office backs the start-up. “Sentiment has moved to the idea that enterprise is really where you get paid for AI.”
The San Francisco-headquartered start-up was founded in 2021 by a group of ex-OpenAI researchers, including siblings Dario and Daniela Amodei, chief executive and president, respectively.
Since then, it has cultivated a sober, safety-orientated image, reinforced by lengthy blog posts by its CEO warning of the dangers of untrammelled AI.
Anthropic’s Claude Code tool for software engineering has become the industry leader since its launch a year ago. The system can read a company’s existing code, plan tasks and execute them. It marks an early demonstration of “agentic” capabilities investors hope will unlock massive new markets, as AI models gain the ability to carry out complex tasks independently.
It captivated workers and spawned the term “Claude benders” for marathon sessions building websites or apps with the tool.
“Anthropic wanted to build with Claude Code internally, [but] when they saw how good it was they productised it aggressively,” said Matt Murphy, a partner at Menlo Ventures, which first invested in Anthropic in 2023.
Anthropic faces fierce competition from Google and OpenAI, with OpenAI releasing updates to its own coding tool, Codex, this week.
An investor in both Anthropic and OpenAI insisted both could still thrive. “Anthropic’s magic around coding is totally different from OpenAI’s magic around ChatGPT replacing search.”
Another OpenAI backer highlighted Anthropic’s modest market share: “Coding does not equal enterprise, it equals developers.”
But the wager being made by investors in Anthropic’s ongoing funding round — set to include Nvidia, Microsoft and top-tier venture capital firms including Lightspeed Venture Partners, Sequoia Capital and Altimeter Capital — is that the company’s tools will go further, reinventing white-collar jobs.
“We took a view that AI is not ‘enterprise’ software in the traditional sense of going after IT budgets: it captures labour spend, at some point you’re taking over human workflows end to end,” said Sebastian Duesterhoeft, a partner at Lightspeed.
The firm wrote a $1bn cheque to Anthropic last year, its single largest investment.
In a bid to press home its advantage, Anthropic released a powerful new model on Thursday, called Claude Opus 4.6, and has pioneered techniques to train its models and manage their interactions with applications and databases.
Over recent days, it released a set of “plug in” tools catering to specific industries, including law, sales, finance, marketing and customer support. On Friday, Goldman Sachs announced it was working with Anthropic on an AI agent to automate roles at the bank.
These moves helped trigger a sell-off in public markets this week, with billions of dollars wiped from stocks related to data, enterprise software, advertising and publishing.
“Anthropic’s story has been very consistent: increasing intelligence will unlock greater market share. That feels very nebulous until you see what they actually mean,” said Lillian Li, an investment manager at Baillie Gifford, which first backed Anthropic in 2025.
Market reaction this week reflected “a realisation moment”, she added.
Li and other investors highlighted the start-up’s mission-oriented culture as crucial to its success, in attracting and retaining talent. “Culture is an afterthought at most places,” said one. “But for them it’s a religion, they’re evangelical about their beliefs over there.”
All seven Anthropic co-founders remain at the company. According to multiple investors, that stability provides an advantage over OpenAI, which has seen eight of its 11-strong founding team depart since it was launched in 2015. Altman himself was briefly ousted by the company’s board in 2023.
This week the company pledged not to introduce advertising into its products, distancing itself from rivals such as OpenAI, which has begun trial ads in ChatGPT in the push for new revenues.
Anthropic plans to broadcast the decision via a series of tongue-in-cheek television adverts due to run at this weekend’s Super Bowl. The ads are backed by a Dr Dre track with the lyrics: “what’s the difference between me and you? You talk a good one but you don’t do what you supposed to do”.
Daniela Amodei has claimed they are not directed at any other company.
Altman described the ads as “clearly dishonest” and “on brand for Anthropic doublespeak” on X, before telling the tech podcast TBPN that the war of words was “a sideshow”.
Altman said: “People are excited for a food fight between companies but the amazing capabilities of these models, the product, the groundswell of excitement around Codex, that feels a lot more important.”
Data visualisation by Clara Murray