For the first time since ChatGPT launched, Anthropic has overtaken OpenAI in US business adoption. Ramp’s April 2026 AI Index, published on May 13, puts Anthropic at 34.4% of tracked enterprise spending versus OpenAI at 32.3%. Anthropic added 3.8 percentage points in a single month while OpenAI fell 2.9. A year ago, OpenAI sat near 32% and Anthropic was under 8%. The lead is one month old, but the trajectory is brutal.
If you’re choosing between Claude und ChatGPT, or trying to make sense of which company you actually want to bet on, the honest answer is messier than the headline. OpenAI is still the biggest AI brand on the planet, with roughly 900 million weekly ChatGPT users and an $852 billion valuation. Anthropic is smaller in users but ahead on revenue growth, enterprise share, and possibly the next valuation round. This guide breaks the comparison into the parts that actually matter in 2026: money, market share, models, safety, and which one you should personally pay for.
The Key Takeaways
- Enterprise adoption: Anthropic 34.4% vs OpenAI 32.3% (Ramp, April 2026, US businesses).
- Valuation: OpenAI $852B (March 2026); Anthropic $380B after Series G (Feb 12, 2026), with talks for a new round at $900B+ pre-money reported in May 2026.
- Q1 2026 revenue: OpenAI $5.7B; Anthropic about $4.7B (The Information). OpenAI ahead on trailing revenue, Anthropic ahead on forward ARR per Sacra’s $43B April estimate.
- Consumer reach: ChatGPT around 900M weekly users; Claude is a much smaller, mostly paid, mostly developer and enterprise base.
- Flagship models: Claude Opus 4.7 (Anthropic) vs GPT-5.5 / GPT-5.5 Instant (OpenAI), both generally available across major clouds.
Anthropic vs OpenAI at a Glance
If you only have 30 seconds, this table covers the core differences in May 2026. Every number is dated so you can spot when it goes stale.
| Category | Anthropic | OpenAI | Ahead | Notes |
|---|---|---|---|---|
| Founded | 2021 | 2015 | OpenAI (older) | Anthropic founders are ex-OpenAI |
| Founders | Dario Amodei, Daniela Amodei + ex-OpenAI researchers | Sam Altman, Elon Musk, Greg Brockman, Ilya Sutskever et al. | Tie | Different missions, same talent pool |
| Flagship model (May 2026) | Claude Opus 4.7 | GPT-5.5 / GPT-5.5 Instant | Tie | Both general-purpose frontier |
| Latest valuation | $380B (Feb 2026); $900B+ in talks (May 2026) | $852B (March 2026) | Anthropic, if next round closes | Pre-money vs post-money; see note below |
| Q1 2026 revenue | ~$4.7B (The Information) | $5.7B (The Information) | OpenAI on trailing | OpenAI was ~$1B ahead in Q1 |
| Forward ARR (April 2026) | ~$43B (Sacra estimate) | ~$25B | Anthropic on momentum | Sacra is an analyst estimate, not official |
| US enterprise share | 34.4% (April 2026) | 32.3% (April 2026) | Anthropic | First time ever |
| Consumer reach | Smaller, paid-heavy, dev-heavy | ~900M weekly ChatGPT users | OpenAI | OpenAI dominates the consumer market |
| Best for | Coding, long documents, regulated industries | Multimodal, image/video, casual chat, agents | Depends on use case | See verdict section |
The table reads like a split decision because it is one. Anthropic leads on money and enterprise; OpenAI leads on people and brand. That’s the whole 2026 story in one line.
What Are Anthropic and OpenAI?
Anthropic and OpenAI are the two largest US generative AI companies. Anthropic, founded in 2021 by former OpenAI researchers Dario and Daniela Amodei, focuses on AI safety and enterprise customers, and makes the Claude family of models. OpenAI, founded in 2015 by Sam Altman and others, focuses on consumer AI through ChatGPT and a broader stack of multimodal products like DALL-E, Soraund Codex.
The interesting wrinkle is the lineage. Most of Anthropic’s founding team came from OpenAI in 2021 over disagreements about how aggressively to commercialize and how seriously to take alignment research. Dario was VP of Research at OpenAI before leaving. That history shapes everything from product design to investor pitches today.
8 Differences Between Anthropic and OpenAI in 2026
1. Mission and origin
Anthropic sells itself as a safety-first lab. Its method, called Constitutional AI, trains models against a written set of principles rather than only against human raters. The pitch to enterprise customers is that Claude is the model you can ship to a bank, a hospital, or a government without getting fired.
OpenAI sells itself as the lab building artificial general intelligence for the broadest possible benefit. Safety work exists, but the public face is a consumer product company that ships fast, broadens distribution constantly, and integrates with Microsoft, Apple, and now its own consulting arm DeployCo.
This isn’t marketing fluff. It shows up in the customer mix.
2. Revenue
Read the revenue question two ways and you get two different winners. By trailing Q1 2026 revenue, OpenAI was ahead: $5.7 billion vs Anthropic’s roughly $4.7 billion, per The Information. That’s a single-quarter snapshot, and OpenAI still owns the bigger book of business this minute.
By forward annualized run rate, Anthropic is ahead. Sacra estimates Anthropic at about $43 billion ARR as of April 2026, up from roughly $9 billion at the end of 2025. OpenAI’s ARR sits around $25 billion on the same basis. A core driver of Anthropic’s ramp is Claude Code, the developer agent, which hit $2.5 billion in annualized revenue in February 2026 and has more than doubled since the start of the year per Sacra.
Two caveats are worth bolding. First, ARR is an extrapolation of the most recent monthly revenue and can move sharply if growth accelerates or stalls. Second, both numbers come with footnotes about how AI vendors count revenue (gross vs net of credits, committed vs realized, run rate vs trailing). The honest read is: OpenAI is the bigger business today, Anthropic is the faster-growing one.
3. Valuation
OpenAI is valued at $852 billion after a $122 billion round that closed on March 31, 2026, led by Amazon ($50B), Nvidia ($30B), and SoftBank ($30B). That had been the largest private valuation in tech history. Anthropic closed a $30 billion Series G auf February 12, 2026, at a $380 billion post-money valuation, led by GIC und Coatue with co-leads including D.E. Shaw Ventures, Dragoneer, Founders Fund, ICONIQund MGX.
The bigger story is the next round. By May 22, 2026, Bloomberg reported that Anthropic’s next financing was expected to close as soon as the week of May 26 at a pre-money valuation above $900 billion. The round is co-led by Sequoia Capital, Dragoneer, Altimeter Capitalund Greenoaks Capital Partners, with each putting in roughly $2 billion. Founders Fund und General Catalyst are also participating. If the round closes at that level, Anthropic would overtake OpenAI as the world’s most valuable AI startup, about a year after being a fraction of its size. Until paperwork is signed, treat the number as reported rather than final.
This is the one that broke containment in May 2026. Ramp’s AI Index measures vendor share among US businesses paying for AI tools through Ramp’s corporate spend platform. The April 2026 data, published May 13, puts Anthropic at 34.4% (up 3.8 percentage points in a month), OpenAI at 32.3% (down 2.9 percentage points), and Google at just under 10%, far behind both.
The annual picture is more dramatic. Over the past year, Anthropic quadrupled business adoption while OpenAI grew enterprise adoption by 0.3%. As Axios reported, Ramp’s own economist Ara Kharazian called it “a stunning reversal,” but also cautioned that one month is not a moat. He’s right. Still, the direction is undeniable.
Anthropic’s own numbers back it up: more than 300,000 business customers as of October 2025, 100,000+ of them running Claude on Amazon Bedrock by April 2026, and over 1,000 customers spending more than $1 million per year, a number that roughly doubled in two months.
5. Consumer reach
OpenAI wins this one by a huge margin and probably will for years. ChatGPT serves around 900 million weekly active users, has the most downloaded AI app on iOS and Android, and is the default assistant in most people’s heads when they say “AI.”
Anthropic has Claude.ai and the Claude apps, but the user base is smaller and skews developer, technical, and paid. Anthropic doesn’t publish a weekly active user number, and most third-party traffic data puts Claude well below ChatGPT and below Google’s Gemini app too.
If you measure by “who has the most humans typing into a chat box,” OpenAI is the clear leader and not close. That’s why headlines like “Anthropic passes OpenAI” need an asterisk: it’s enterprise adoption, not raw users.
6. Products and platform
OpenAI ships more product surface than Anthropic. The OpenAI lineup includes ChatGPT across free, Plus ($20), Pro ($200), Business, and Enterprise tiers, with GPT-5.5 und GPT-5.5 Instant as the default models and GPT-5.5 Thinking for harder problems. On the creative side, OpenAI ships DALL-E for images, Sora for video (now positioned more as a research preview after Sora 2 was discontinued), and Codex for code. Round it out with the OpenAI API, Custom GPTs, an operator-style agent, and the newly spun-out DeployCo consulting arm built around the Tomoro acquisition.
Anthropic’s surface is narrower but increasingly enterprise-grade: Claude in three sizes (Opus 4.7, Sonnet, Haiku), the Claude API, Claude Code, Claude Cowork for collaborative work, distribution through Amazon Bedrock, Google Cloud Vertex AIund Microsoft Foundry, and a growing managed-agents program for long-running workflows.
A useful way to think about it: OpenAI is trying to be every layer of the AI consumer stack; Anthropic is trying to be the model layer that every other platform pays to embed.
7. Pricing for individual users
For a personal user on a Mac or iPhone, the two flagship subscriptions cost the same. ChatGPT Plus is $20 a monthund Claude Pro is $20 a month. The next tier up is where it diverges: ChatGPT Pro is $200 a month for high-end and unlimited-leaning use, while Claude Max sits between Pro and ChatGPT Pro depending on the plan.
OpenAI’s API pricing is cheaper for ultra-light workloads thanks to the Nano tier, which lands as low as $0.10 per million input tokens. Anthropic doesn’t ship a comparable budget tier; the cheapest Haiku model is materially more expensive per token. If you’re running high-volume cheap inference, OpenAI wins on price. If you care about quality per dollar at the top end, the answer is closer.
8. Public safety stance
Anthropic’s brand is built around safety. Constitutional AI, a published responsible scaling policy, public model cards, and the Claude Mythos cybersecurity program announced in 2026 are the visible artifacts. Enterprise buyers in regulated industries cite this as a reason for choosing Claude.
OpenAI has invested heavily in safety too, with red-teaming programs, preparedness frameworks, and an external safety board. But OpenAI’s reputation has been bruised by repeated executive departures over safety disagreements and by the broader perception that the company optimizes for product velocity. Whether that’s fair depends on who you ask. What’s measurable is that risk-sensitive buyers vote with their wallets and they’re voting Anthropic more often than they used to.
Anthropic vs OpenAI Revenue and Valuation in Plain Numbers
Both companies are growing in absolute terms. The gap is direction.
OpenAI is a giant business with a near-monopoly on consumer AI mindshare. $5.7 billion in a single quarter is not a small number; it makes OpenAI one of the fastest-growing software companies in history. Enterprise revenue now makes up over 40% of OpenAI’s mix, up from around 30% a year earlier, and the company is on track to hit consumer-enterprise parity by year-end. The DeployCo subsidiary, with $4 billion of initial investment and 150 engineers from the Tomoro acquisition, is a clear bet on owning the consulting layer.
Anthropic is the faster line on the chart. Going from $1B to $43B ARR in 15 months is a 30x increase, almost all of it driven by enterprise contracts and Claude Code demand. The reported $900B+ pre-money valuation in May 2026 would mark the moment Anthropic overtakes OpenAI on paper as the most valuable private AI company. If the round closes at that level, expect the framing of this whole comparison to flip in real time.
Valuations are not the same thing as profits. Neither company is profitable in any conventional sense. Both are spending billions per quarter on compute, talent, and data center commitments. The race for revenue is also a race for the cash flow needed to keep paying for those commitments.
Enterprise Adoption: Why Companies Are Switching to Claude
The Ramp number is real, but it doesn’t tell you why. Talking to engineering leaders, three reasons come up over and over:
The first is Claude Code. Anthropic’s developer agent has become the default tool inside many engineering orgs, especially for large refactors, autonomous task execution, and long-context codebase work. When the dev team picks Claude, the rest of the org usually follows.
The second is context window and long-document handling. Claude’s million-token context, available across Opus and Sonnet, makes it the natural choice for legal review, financial analysis, and any workflow where you can’t reasonably chunk a document.
The third is brand safety. Compliance, legal, and risk teams have an easier time approving a vendor that puts safety front and center in its marketing and has fewer public incidents. Whether that’s a real durable advantage or a temporary perception edge, it’s moving budget today.
OpenAI is fighting back hard. DeployCo, expanded Microsoft Foundry distribution, enterprise-only model variants, and aggressive pricing on GPT-5.5 Instant are all part of the response. The Q3 2026 Ramp data will be the next real signal of whether the flip holds or reverses.
Consumer Reach: ChatGPT vs Claude in Daily Use
If you’re not buying for a business, the consumer comparison is the one that matters. For everyday questions, brainstorming, writing assistance, and image generation, ChatGPT remains the broadest tool. The product has more surface area, more integrations, image and video generation built in, voice mode, custom GPTs, and a mobile app polish that Claude still hasn’t matched.
For long documents, dense reasoning, careful writing, and any task where you want a model that follows nuanced instructions over thousands of words, Claude is the better daily driver. The interface is calmer, the responses are typically more careful, and the model is harder to bait into making things up.
If you want the deeper feature-by-feature breakdown, we cover it in our Claude vs ChatGPT comparison. This article is about the companies; that one is about the products.
Safety: Constitutional AI vs RLHF
Both companies use reinforcement learning from human feedback (RLHF) as a baseline. The difference is what they add on top.
Anthropic’s Constitutional AI trains models to critique and revise their own outputs against a written set of principles, then uses those self-critiques to fine-tune behavior. The intent is to scale alignment beyond what human raters can review one prompt at a time, and to make the model’s values explicit and inspectable rather than hidden in millions of preference pairs.
OpenAI emphasizes a layered approach: pre-training filters, RLHF, instruction tuning, deliberative alignment, and external red-teaming, supported by the Preparedness Framework that scores models on dangerous capability categories before release. It’s a less branded approach, but it covers similar ground in different language.
For a consumer choosing between Claude Pro and ChatGPT Plus, the practical safety difference is small. For a Fortune 500 buyer signing a multi-year contract, the perception of safety leadership is one of the reasons the Ramp chart looks the way it does.
Who Should Use Anthropic and Who Should Use OpenAI
Pay for Claude if you do a lot of coding and want Claude Code, you write long documents, you work with sensitive material that needs careful handling, or you’ve found that ChatGPT hallucinates too aggressively on your kind of work. Claude rewards people who write detailed prompts and care about the quality of long outputs.
Pay for ChatGPT if you want the broadest single AI tool, you care about image and video generation in the same product, you use voice mode regularly, you build custom GPTs for niche workflows, or you’re a casual user who values speed and breadth over depth.
Pay for both if you can afford $40 a month and you want the best of both worlds, which is what a lot of power users actually do. The cleaner alternative, especially on Mac and iPhone, is to use a single multi-model app that gives you all of them under one subscription.
Anthropic vs OpenAI on Mac and iPhone
Both companies ship native Mac and iOS apps and both are good. ChatGPT’s Mac app is more polished, has Apple Intelligence integration through the system extension, and ships voice mode and image generation in one window. Claude’s Mac app is leaner and faster, focused on text and document work, and pairs well with terminal workflows where Claude Code is involved.
If you don’t want to choose, the third option is a multi-model app like Fello AI, which lets you call Claude, ChatGPT, Gemini, Grok, and DeepSeek from a single subscription. The point of multi-model isn’t ideology; it’s that the right model depends on the prompt, and switching apps every time is friction you don’t need. One price, many models, including newer ones like DeepSeek that don’t have polished Mac apps of their own.
A Timeline of the 2025–2026 Race
The shift didn’t happen overnight. Here’s how it looked through the year.
| Date | Company | Milestone |
|---|---|---|
| April 2025 | OpenAI / Anthropic | OpenAI commands roughly 32% of US enterprise AI adoption per Ramp. Anthropic is at about 8%. |
| Mid-2025 | Anthropic | Enterprise pipeline begins compounding hard, driven by the Claude family and the early Claude Code beta. |
| End of 2025 | Anthropic / OpenAI | Anthropic ARR hits approximately $9B per Sacra. OpenAI continues to scale ChatGPT consumer revenue and pushes deeper into enterprise. |
| February 12, 2026 | Anthropic | Closes $30B Series G at a $380B post-money valuation. CFO Krishna Rao says “Claude is increasingly becoming critical to how businesses work.” |
| February 2026 | OpenAI | ChatGPT reaches 900 million weekly active users und 50 million paying subscribers. |
| March 31, 2026 | OpenAI | Closes a $122B round at an $852B valuation, led by Amazon, Nvidia, and SoftBank. |
| April 23, 2026 | OpenAI | Launches GPT-5.5, which posts strong benchmark numbers and becomes the default ChatGPT model. |
| April 2026 (data) | Anthropic / OpenAI | Ramp vendor share: Anthropic 34.4%, OpenAI 32.3%. Anthropic ARR estimated at $43B by Sacra. OpenAI Q1 2026 revenue reported at $5.7B, about $1B ahead of Anthropic’s Q1 per The Information. |
| May 13, 2026 | Anthropic | Ramp publishes the April data. Axios and others run the “Anthropic passes OpenAI” headline. |
| May 22, 2026 | Anthropic / OpenAI | Bloomberg reports Anthropic’s next financing is expected to close as soon as the week of May 26 at a pre-money valuation above $900 billion. Anthropic’s flagship model is Claude Opus 4.7, generally available across Claude products and major clouds. OpenAI launches DeployCo consulting subsidiary with $4B initial investment after acquiring Tomoro. |
The next data point that actually matters is the May 2026 Ramp index, which will tell us whether OpenAI’s product velocity and DeployCo push reverse the enterprise flip or whether Anthropic widens the lead.
Conclusion: Who’s Actually Winning?
There isn’t one winner. Read the result by lens.
If “winning” means revenue growth, Anthropic is winning, and the gap is widening. If it means trailing quarterly revenue, OpenAI is still ahead by about $1B as of Q1 2026. In case it means enterprise market share in US businesses, Anthropic is winning, by a one-month lead that may or may not hold. If it means valuation, OpenAI is ahead today and Anthropic is days away from passing it if the reported May round closes at $900B+. And if it means consumer reach, OpenAI is winning by an order of magnitude and shows no sign of slowing.
For most readers, the practical takeaway is simpler. Pay for Claude if you code, write long, or work with sensitive material. Pay for ChatGPT if you want the broadest single AI tool with multimodal built in. Get both if you can; use a multi-model app like Fello AI if you’d rather have one bill and access to every major model from your Mac or iPhone. Pick the company you want to bet on for valuation reasons only if you’re an investor, and even then, the next twelve months could change the answer twice.
If you want to go deeper on the product comparison, our Claude vs ChatGPT guide has the feature-by-feature breakdown. For the wider field including Google and the Chinese challengers, see our Best AI Models 2026 hub.
FAQ
Is Anthropic owned by OpenAI?
No. Anthropic is an independent company. Its founders, Dario and Daniela Amodei, previously worked at OpenAI and left in 2021 to start Anthropic. The two companies are direct competitors.
Does Amazon own Anthropic?
No, but Amazon is a major investor and Anthropic’s primary cloud partner. Amazon has committed billions to Anthropic across multiple rounds and offers Claude through Amazon Bedrock. Anthropic remains an independent company with its own board and leadership.
Who founded Anthropic and why?
Anthropic was founded in 2021 by Dario Amodei, Daniela Amodei, and other former OpenAI researchers, primarily over disagreements about the pace of commercialization and how seriously to prioritize alignment and safety research.
Is Anthropic actually bigger than OpenAI now?
It depends on the measure. By US enterprise adoption as of April 2026, yes, by a narrow margin of 34.4% to 32.3% per Ramp. By forward annualized run rate, Anthropic is ahead at an estimated $43B vs OpenAI’s roughly $25B per Sacra. In trailing Q1 2026 revenue, OpenAI was about $1B ahead. According to consumer users and current valuation, OpenAI is still bigger, though a reported $900B+ Anthropic round in late May 2026 could flip the valuation lead.
Should I subscribe to Claude or ChatGPT in 2026?
Both cost $20 a month for the standard tier. Pick Claude for coding, long documents, and careful writing. Pick ChatGPT for multimodal use, image generation, voice mode, and breadth. If you do both kinds of work, the cleanest answer on Mac and iPhone is a multi-model app like Fello AI that gives you Claude, ChatGPT, Gemini, Grok, and DeepSeek under one subscription.
