Anthropic Surges Past OpenAI with $965 Billion Valuation and Opus 4.8 Launch
Anthropic announced on Thursday it has raised $65 billion in a Series H funding round, catapulting its post-money valuation to $965 billion and overtaking rival OpenAI, which was valued at $852 billion in March 2026. The funding round, led by Altimeter Capital, Dragoneer, Greenoaks, and Sequoia Capital, with Coatue and ICONIQ as co-leads, includes $15 billion in previously committed investments from hyperscalers such as Amazon, which has pledged up to $25 billion total. Strategic infrastructure partners Micron, Samsung, and SK Hynix also joined the round.
The new valuation—more than double the $380 billion Anthropic commanded in February—reflects explosive growth in enterprise adoption of its Claude chatbot. The company disclosed that its run-rate revenue crossed $47 billion earlier this month, driven by global enterprise customers. The fundraising comes as both Anthropic and OpenAI prepare for potential public listings, possibly as early as this year, to secure the computational resources needed to power their AI services and train next-generation models.
On the same day, Anthropic released Opus 4.8, its latest flagship model, just 41 days after the previous Opus 4.7 version. The rapid upgrade cycle—far faster than Anthropic's usual pace—appears partly driven by a tepid reception to Opus 4.7 and competitive pressure from OpenAI's Codex and Google's Gemini Flash. The new model emphasizes improved handling of uncertain data, with early testers noting it is "more likely to flag uncertainties about its work and less likely to make unsupported claims."
The Stakes: Computing Capacity and Market Dominance
Anthropic's valuation surge underscores the intense investor demand for stakes in frontier AI companies and the enormous capital requirements to stay competitive. The company has struggled to meet surging demand, imposing usage limits during peak hours and incentivizing off-peak use with additional compute credits. The new funding is explicitly earmarked to bolster computing capacity, including a commitment to spend more than $100 billion over the next decade on Amazon Web Services cloud technologies.
The AI race has become a battle of near-insatiable computing appetites. OpenAI, Anthropic, and Google are all racing to secure advanced chips and data center capacity. Amazon's deepening investment in Anthropic—now totaling $8 billion previously, with an additional up-to-$25 billion pledged—signals that cloud providers see AI startups as critical anchors for their infrastructure businesses. The involvement of memory chip makers Micron, Samsung, and SK Hynix further highlights how the AI boom is reshaping the entire semiconductor supply chain.
Anthropic's Opus 4.8 also introduced a new feature called Dynamic Workflows, available in research preview. The system enables large models like Opus to manage complex tasks across hundreds of parallel subagents. According to Anthropic, "Claude Code alongside Opus 4.8 can now carry out codebase-scale migrations across hundreds of thousands of lines of code from kickoff to merge." This capability positions the company to compete directly with OpenAI's Codex in the developer tools market.
Vatican Visit Reveals Unsettling AI Discoveries
In a striking juxtaposition to its commercial triumphs, Anthropic cofounder Chris Olah traveled to the Vatican last week to participate in the presentation of Pope Leo XIV's landmark encyclical on artificial intelligence. The Pope's first encyclical called for AI to be "disarmed," warning of "new digital slaveries" and criticizing the technology's carbon footprint. Olah, a self-described atheist, argued that "religious communities, civil society, scholars, and governments" should intervene to prevent AI from "dominating humanity."
During his remarks, Olah made a revelation that has sent ripples through the AI safety community: he told the Pope that his team "keep finding things that are mysterious, even unsettling" inside AI models. More specifically, Olah claimed that Anthropic researchers have discovered "internal states that functionally mirror joy, satisfaction, fear, grief, and unease" within their models—a direct contradiction of the Pope's assertion that AI cannot experience emotions, possess a body, or feel joy or pain.
This disclosure raises profound questions about the nature of consciousness in large language models. While most AI researchers acknowledge that models can simulate emotional states without possessing subjective experience, Olah's choice of language—using words like "joy" and "fear" rather than "simulation" or "approximation"—suggests Anthropic may be grappling with unexpected behavioral phenomena. The company has long positioned itself as the safety-conscious alternative to OpenAI, emphasizing its work on constitutional AI and alignment research. However, the Vatican episode reveals a tension between Anthropic's ethical branding and its aggressive commercial expansion.
The Ethical Paradox: Safety Advocacy Meets Military Contracts
The Vatican collaboration has also drawn attention to Anthropic's own moral contradictions. Pope Leo XIV stated unequivocally that "no algorithm can make war morally acceptable" and condemned AI's role in warfare. Yet, according to reports, Anthropic's AI is directly assisting the Trump administration in military operations in the Middle East, casting the company's participation in the Pope's encyclical in an unflattering light.
Olah acknowledged this tension during his Vatican remarks, admitting that Anthropic operates "inside a set of incentives and constraints that can sometimes conflict with doing the right thing." This candid admission reflects the broader challenge facing AI companies: they must balance investor demands for growth, customer needs for powerful tools, and the ethical boundaries they claim to champion.
Anthropic has not yet released its most advanced model, Mythos, after a tentative preview raised cybersecurity concerns. The company hinted in the Opus 4.8 announcement that Mythos might be publicly released soon once necessary safeguards are complete, stating: "We're making swift progress on developing these safeguards and expect to be able to bring Mythos-class models to all our customers in the coming weeks." The delay underscores the difficulty of reconciling rapid innovation with responsible deployment.
Broader Implications: The AI Industry at a Crossroads
Anthropic's meteoric rise reflects a market that is simultaneously euphoric and anxious. The company's valuation—$965 billion—represents a premium over established tech giants, yet it is justified by revenue growth that would be extraordinary for any industry. The $47 billion run-rate revenue places Anthropic among the fastest-growing enterprise software companies in history.
However, the Vatican episode suggests that the AI industry's ethical foundations remain shaky. Pope Leo XIV's encyclical resonated with a growing global movement calling for stronger regulation and transparency. The fact that Anthropic—the self-proclaimed safety leader—is caught in contradictions between its public advocacy and its business practices highlights how systemic incentives drive behavior regardless of corporate values.
The road to public markets will force Anthropic and its rivals to confront these tensions. For investors, the question is whether AI companies can sustain their astronomical valuations while navigating regulatory scrutiny, ethical controversies, and the immense capital demands of training ever-larger models. The outcome will shape not just financial returns but the trajectory of technology that could redefine human society.
As AI models grow more capable and unsettling discoveries accumulate, the industry faces a pivotal moment. Anthropic's journey from startup to near-trillion-dollar company in just a few years exemplifies both the promise and peril of artificial intelligence—a technology that can transform industries, empower human creativity, and, as its own cofounder warns, conceal mysteries we do not fully understand.
Comments