OpenAI, fresh off the largest private fundraising round in history, is facing mounting questions from some of its own backers over its $852 billion valuation and a whiplash-inducing pivot in strategy that prioritizes the higher-margin enterprise market at the expense of its consumer crown jewel - all because Anthropic is starting to drink their milkshake with enterprise contracts.
The company raised $122 billion last month from Silicon Valley and global capital - including SoftBank, Amazon, Nvidia, Andreessen Horowitz, Sequoia Capital and Thrive Capital. Yet even as Chief Financial Officer Sarah Friar hailed the oversubscribed deal as proof of “strong conviction” in the company’s direction, early investors are voicing skepticism. One told the Financial Times the pivot feels unfocused: “You have ChatGPT, a 1 billion-user business growing 50-100% a year - what are you doing talking about enterprise and code?”
Friar disagrees. "The suggestion that investors are not supportive of our strategy defies the facts," she said. "Our . . . raise, the largest in history, was oversubscribed, completed in record time and backed by a broad set of global investors, reflecting strong conviction in both our direction, current business momentum and long-term value."
The repositioning has indeed been swift and, to critics, symptomatic of the kind of strategic whiplash that often precedes trouble in hype-driven sectors. In December Chief Executive Sam Altman issued a "code red" urging staff to refocus on core business. High-profile consumer experiments have been quietly euthanized: the video-generation service Sora was shuttered, killing a planned $1 billion investment from Disney; an “adult” chatbot was mothballed; parts of the ambitious Stargate data-center project were ditched; and a $100 billion Nvidia deal was substantially scaled back. Even a recent “low hundreds of millions” acquisition of the tech talk show TBPN drew internal eye-rolling from investors who called it a distraction.
"I don’t get it frankly, it doesn’t make any sense to me," one investor told FT. "It’s a distraction and it irks me."
The new gospel is enterprise. OpenAI is reallocating computing resources toward its Codex coding tool, which insiders say could eventually eclipse ChatGPT in priority as the company chases nontechnical business users. Headcount is set to nearly double to 8,000 by year-end. Roughly half of revenue is expected to come from corporate customers, up from about 40% today. A new permanent office in London is in the works to anchor the largest research hub outside the U.S. The message from the C-suite: the market for corporate AI tools is “ours to win.”
However, fresh data from Morgan Stanley’s 1Q26 CIO Survey (fielded February 3–March 10 among 100 US and European CIOs - available to pro subs here) offers some early empirical support for the enterprise pivot - while also highlighting just how steep the climb is. Artificial Intelligence/Machine Learning has cemented its position as the clear #1 CIO priority (17.7% of respondents named it a top-three area, up from 16.3% in 4Q25), with 39% now calling it their single highest priority. Yet when CIOs were asked which vendors are poised to capture the largest incremental share of GenAI spending, Microsoft dominated both the one-year and three-year outlooks by a wide margin. OpenAI still ranked solidly inside the top tier - behind Microsoft, Amazon, Google, Salesforce and ServiceNow - and was also cited as a preferred vendor for building custom AI applications today and three years out.
The survey underscores the broader reality: overall 2026 IT budgets are growing only modestly at +3.7%, with Software the sole category expected to accelerate (+4.1%). Hyperscalers (led overwhelmingly by Microsoft via Azure OpenAI Service, Copilot, and its massive existing enterprise footprint) remain the dominant wallet-share winners in AI and cloud. AI labs and application vendors, including OpenAI, are making incremental gains on top of that foundation.
Rival Anthropic is making that claim harder to swallow. Founded by ex-OpenAI talent and led by Dario Amodei, the Claude maker has seen annualized revenue surge to $30 billion by the end of March from $9 billion at the close of 2025, fueled by demand for its coding and cybersecurity offerings. Secondary markets are now pricing Anthropic ahead of OpenAI for the first time. The startup has fielded multiple offers that could value it at $800 billion or higher - more than double its February tender valuation - though it has so far resisted. One investor who backs both companies noted that underwriting OpenAI’s latest round required assuming an IPO valuation north of $1.2 trillion.
Meanwhile, Anthropic has shrugged off a major national-security black eye that appears to have served as great marketing, after the Pentagon formally designated the company a “supply chain risk” to U.S. national security - the first time such a label, historically reserved for foreign adversaries like Huawei or Kaspersky, has ever been applied to a major American AI firm. The unprecedented move followed a bitter contract standoff in which Anthropic refused to strip safety guardrails from Claude that blocked its use for mass domestic surveillance or lethal autonomous weapons. Anthropic sued immediately, calling the designation retaliatory; courts have issued temporary blocks in some venues while litigation continues. Yet, this government smackdown has had no effect on private-market enthusiasm.
The two firms remain locked in a brutal arms race, each hemorrhaging billions annually on compute. OpenAI boasts a formidable infrastructure edge - 8 gigawatts secured now, targeting 30 gigawatts by 2030 -and claims it can simply serve a slightly inferior model if needed. Anthropic, by contrast, has cited outages and power constraints while promising restraint on further expansion. OpenAI’s new chief revenue officer, Denise Dresser, has accused Anthropic of overstating revenue by roughly $8 billion via cloud-partner gross-ups, though both sides insist they follow standard accounting.
Of course, there's an underlying catch: the lofty valuations rest on the assumption that enterprises will eventually pay up for these tools in volume. Yet a telling data point from the political arena suggests institutional buyers remain skittish. Republican campaigns are leaning into AI for messaging and voter targeting ahead of the 2026 midterms. The Democratic National Committee, however, has explicitly banned staff from using either ChatGPT or Claude, citing data-privacy and security risks.
OpenAI executives insist the repositioning towards enterprise is simply the necessary maturation of a company that has already reinvented itself multiple times. The massive war chest, they argue, provides “max flexibility” and “max optionality.” But with both startups still deeply unprofitable, compute burn rates that would make traditional tech CFOs blanch, and secondary-market momentum tilting toward the more focused rival, the narrative is shifting. What began as a consumer phenomenon is now a high-stakes bet that enterprise dollars will arrive fast enough—and in sufficient volume—to justify valuations that, to skeptics, increasingly look detached from today’s economics.