When your CFO publicly disputes your CEO's timeline for going public, that's not a scheduling disagreement. That's a signal.
OpenAI's Chief Financial Officer Sarah Friar has raised concerns about CEO Sam Altman's plan to take the company public in Q4 2026. Her worry: the gap between OpenAI's $852 billion valuation and its financial reality is too wide. The company is projected to lose $14 billion in 2026 despite $25 billion in annualized revenue. Meanwhile, Anthropic just passed OpenAI in revenue—hitting $30 billion ARR in April 2026, up from $9 billion four months earlier.
For enterprise buyers evaluating AI vendors, this isn't Silicon Valley drama. This is a vendor risk calculation.
The Numbers That Don't Add Up
OpenAI secured a $40 billion funding round from SoftBank earlier this year at an $852 billion valuation. That number puts it above Meta, Netflix, and Goldman Sachs. It's one of the highest valuations ever assigned to a private company.
Sarah Friar, OpenAI's CFO, reportedly told Altman the company isn't ready to go public. According to the Financial Times, she cited the disconnect between OpenAI's public narrative and its actual financial performance. When your own CFO doesn't want to face public market investors, that's worth understanding.
The core problem is unit economics. Running AI models at ChatGPT's scale costs extraordinary amounts. OpenAI is projected to lose $14 billion in 2026—not because it's a startup burning cash to grow, but because costs are growing faster than revenue. Microsoft has poured billions into cloud compute. Even with that subsidy, the math doesn't close.
The product roadmap hasn't helped. TechCrunch reported that OpenAI revised its plans at least twice in six months. GPT-5 launched with less differentiation than expected. Enterprise features promised to corporate clients shipped late. The company that once set every industry deadline is now scrambling to meet its own.
Anthropic's Quiet Takeover
While OpenAI was dealing with internal friction, Anthropic executed one of the fastest revenue ramps in enterprise software history.
Anthropic's growth trajectory:
- End of 2025: $9 billion ARR
- April 2026: $30 billion ARR
- Growth: 3.3x in four months
According to The AI Corner and multiple investor reports, Anthropic now leads OpenAI in annualized revenue. OpenAI hit $25 billion ARR by March 2026. Anthropic is at $30 billion. That's not a rounding error—that's a market shift.
Anthropic did this with roughly 5,000 employees, according to SaaStr. OpenAI has significantly more headcount. The efficiency gap is striking. On secondary markets, Anthropic now trades at a premium to OpenAI per dollar of revenue. Read that again: the company that was supposed to be the challenger is now valued more highly per revenue dollar than the category leader.
Photo by Tima Miroshnichenko on Pexels
What changed? Anthropic built its entire strategy around enterprise trustworthiness. Claude has become the default choice for companies that need reliable, safe, consistent output. While OpenAI chased consumer scale with ChatGPT, Anthropic focused on B2B contracts, government deals, and API stability.
Now OpenAI is trying to compete on Anthropic's home turf—enterprise sales—while simultaneously maintaining ChatGPT's consumer momentum. That's a difficult transition. If ChatGPT user growth flattens while OpenAI chases corporate contracts, the company ends up with neither a dominant consumer product nor enterprise leadership.
The IPO Pressure Cooker
Sam Altman wants to take OpenAI public in Q4 2026. Sarah Friar thinks that's too soon. Both are right about something, which makes this harder to resolve.
Altman's logic: The IPO window won't stay open forever. SoftBank, Microsoft, and other investors who put in billions at $157 billion (late 2023) and $852 billion (2025) need an exit path. Every month of delay increases pressure to justify those valuations. Going public gives OpenAI "currency"—stock it can use for acquisitions and talent retention.
Friar's concern: $14 billion in losses is a tough story to sell on a roadshow. Public market investors aren't venture capitalists. They want a path to profitability, not just a narrative about future dominance. The current financials don't provide that path. Reports indicate OpenAI won't reach profitability until around 2030.
Friar reportedly stopped reporting directly to Altman in August 2025 and was excluded from infrastructure and capital strategy discussions. That's not normal CFO-CEO dynamics. That's organizational fracture.
What This Means for Enterprise Buyers
If you're a CTO evaluating OpenAI vs. Anthropic for your next AI deployment, or a CFO approving a multi-year vendor contract, here's what to track.
For CTOs: Vendor Stability Risk
Talent retention. OpenAI has seen significant senior departures over the past year. Much of employee equity is tied to the $852 billion valuation. If that valuation corrects, retention becomes harder. Employees who stayed for the upside start calculating alternatives. In April 2026 alone, OpenAI lost its VP of Science (Kevin Weil), Sora research lead (Bill Peebles), and CTO of Enterprise Applications (Srinivas Narayanan) on the same day.
Product roadmap predictability. Enterprise contracts require stable APIs, predictable pricing, and consistent model behavior. If OpenAI is revising its roadmap every quarter to chase revenue, that introduces deployment risk. Late-shipping features and unexpected deprecations cost engineering time and executive trust.
Competitive pressure. Anthropic's revenue growth isn't just fast—it's accelerating. If that continues, Anthropic could become the default enterprise AI platform within 12-18 months. Switching costs are real but manageable at this stage. The question: do you commit to OpenAI now, or wait to see if Anthropic's momentum holds?
For CFOs: Financial Viability Questions
Vendor continuity. A company losing $14 billion annually isn't necessarily unstable—if it has $40 billion in recent funding and massive investor backing. But if the IPO gets delayed or the valuation corrects, capital availability changes. That affects pricing stability, support quality, and long-term roadmap confidence.
Contract terms. Multi-year commitments to OpenAI lock in pricing but also lock in vendor risk. If Anthropic offers better economics or OpenAI reprices aggressively to compete, you're stuck. Shorter contracts (12-18 months) preserve optionality but may cost more per token.
Acquisition currency. If OpenAI goes public, its stock becomes acquisition currency. That could mean faster consolidation in the AI tooling ecosystem—potentially beneficial if OpenAI buys your existing vendors, potentially disruptive if it doesn't.
For Both: The Anthropic Comparison
Anthropic's $30 billion ARR isn't just a number. It's validation that enterprise buyers are voting with contracts. The company's efficiency (5,000 employees vs. OpenAI's larger headcount) suggests sustainable unit economics. Its focus on safety and trustworthiness resonates with regulated industries.
The trade-off: OpenAI still has consumer scale and brand recognition. ChatGPT is a household name. Anthropic is not. If your use case requires cutting-edge consumer features (Sora video, multimodal ChatGPT), OpenAI may still be the better choice. If your use case is enterprise API reliability, Anthropic's momentum is hard to ignore.
Decision Framework: Three Scenarios
Scenario 1: OpenAI IPO Succeeds (Q4 2026)
- Valuation holds at $500-600B (down from $852B but still strong)
- Losses narrow to $8-10B by end of 2027
- Enterprise revenue surprises to the upside
- Implication: OpenAI remains a safe enterprise choice with public market transparency
Scenario 2: OpenAI IPO Delayed to 2027
- CFO concerns prove well-founded
- Anthropic widens revenue lead
- Some high-profile enterprise clients switch to Claude
- Implication: 12-18 month vendor risk window; prefer shorter contracts with both vendors
Scenario 3: Anthropic Pulls Ahead (Most Likely)
- Anthropic raises at $80B+ valuation in next 6 months
- OpenAI goes public at $400-500B (50% correction from $852B)
- Market normalizes around "two leading enterprise AI platforms"
- Implication: Dual-vendor strategy becomes standard (Claude for reliability, GPT for scale)
What To Watch Over the Next Six Months
Five signals to track:
-
Anthropic's next funding round. If it raises at $80 billion or more, that confirms the secondary market premium is real and growing.
-
OpenAI S-1 filing timing. If a registration statement doesn't appear by September 2026, the Q4 IPO window is almost certainly gone.
-
ChatGPT user growth. OpenAI stopped publishing monthly active user numbers consistently. If the numbers were accelerating, they'd be publishing them.
-
Senior executive and researcher departures. The April 2026 triple-departure (VP Science, Sora lead, CTO Enterprise Apps) is a pattern to monitor.
-
Enterprise deal momentum. Watch for press releases about Fortune 500 contracts. If Anthropic is winning more logos than OpenAI, that's the leading indicator.
The Bottom Line for Enterprise Leaders
This isn't about predicting which company "wins." Both will exist. Both will be viable vendors. The question is: how much vendor concentration risk are you willing to accept?
If you're a CFO: Shorter contracts preserve optionality. Multi-vendor strategies reduce switching risk. Budget for 20-30% annual AI spend growth regardless of vendor, and track unit economics (cost per output token) as closely as you track SaaS per-seat pricing.
If you're a CTO: Test both platforms in parallel for your next deployment. Anthropic's momentum makes it worth evaluating even if you're currently committed to OpenAI. Product roadmap stability matters more than cutting-edge features when you're deploying to production.
If you're both: The CFO-CEO friction at OpenAI is a yellow flag, not a red flag. It signals organizational stress, not imminent failure. But it does mean the IPO timeline is uncertain, which means pricing stability and roadmap predictability are uncertain. Plan accordingly.
The gap between an $852 billion valuation and $14 billion in annual losses has to close somehow. Either revenue accelerates dramatically, or costs come down, or the valuation corrects. Anthropic's $30 billion ARR suggests the market has options. Your job is to make sure your AI strategy does too.
Continue Reading
Sources
- OpenAI's $852B Valuation: Three Cracks That Could Break It - Humai Blog
- Anthropic Passed OpenAI in Revenue: $30B ARR April 2026 - The AI Corner
- OpenAI CFO Raises Concerns Over Sam Altman's 2026 IPO Plans - Economic Times
- Anthropic Only Has ~5,000 Employees. Almost No One Has Ever Been This Efficient - SaaStr
- Latest AI News (April 10, 2026) - OpenTools AI
Have questions about AI vendor strategy? Connect with me on LinkedIn, Twitter/X, or via the contact form.
Want to calculate your own AI ROI? Try our AI ROI Calculator — takes 60 seconds and shows projected savings, payback period, and 3-year ROI.
Photo by