5WPR Study: AI's 15 Biggest Players Rank Worlds Apart on Transparency

PR firm 5WPR released the first structured benchmark of how 15 leading AI companies actually communicate with the outside world. The headline finding: only about a third of them regularly disclose concrete metrics — weekly active users, revenue, or enterprise customer counts — even as they reshape entire industries.

The study, published May 28, evaluates how OpenAI, Anthropic, Google DeepMind, Microsoft AI, Meta AI, and ten other major AI players talk to enterprise buyers, developers, regulators, and consumers. The gap between the most transparent and the most opaque is wider than most observers realized.

Who Got Ranked, and How

The 15 companies covered are OpenAI, Anthropic, xAI, Google DeepMind/Gemini, Microsoft AI/Copilot, Meta AI, Perplexity, Mistral, Cohere, Character AI, Stability AI, Runway, Midjourney, Inflection AI, and Hugging Face.

Each company was graded across three dimensions: implementation messaging (how they describe what their AI does), usage and adoption disclosure (whether they publish real numbers), and founder voice posture (how their CEO shows up publicly).

Why this matters: AI is now a sector where multi-billion-dollar funding decisions are being made based on press releases and selectively leaked benchmark scores. A structured comparison of who actually backs up their claims with data is overdue.

The Most Transparent

5WPR ranked five companies as most transparent on usage and revenue: OpenAI, Anthropic, Microsoft AI/Copilot, Google DeepMind/Gemini, and Perplexity. These are the labs that publish (or regularly leak) hard numbers on revenue, model usage, and enterprise deployment.

A separate ranking for strongest enterprise customer roster put Anthropic, OpenAI, Microsoft AI/Copilot, Google DeepMind/Gemini, and Cohere at the top. Anthropic in particular has leaned heavily into named enterprise disclosures — KPMG, PwC, Deloitte, and a long list of financial services firms.

The pattern: companies with mature enterprise sales motions disclose more because their buyers demand it. Procurement teams at Fortune 500 firms want to see customer lists, security certifications, and revenue trajectory. So the labs serving those buyers publish that information.

The Most Opaque

On the other end, the study identified five companies as "most opaque despite scale": Midjourney, xAI/Grok, Meta AI, Stability AI, and Character AI. Each of these has substantial usage — Midjourney has tens of millions of users, Character AI gets billions of monthly messages — and discloses almost nothing structured about it.

A separate "quietest company relative to scale" ranking covered Midjourney, Character AI, Stability AI, Inflection AI, and Mistral. Mistral standing out here is notable: it's a major European AI player positioning itself as a sovereign alternative, but it discloses very little publicly compared to its US peers.

The opacity isn't accidental. Several of these companies are either privately held with no public-market reporting obligations, or in the middle of legal and regulatory entanglements that make selective silence the safer move.

Loudest Founder Voices

The study also ranked CEOs by public posture. The five loudest founder voices: Elon Musk (xAI), Sam Altman (OpenAI), Mark Zuckerberg (Meta AI), Dario Amodei (Anthropic), and Aravind Srinivas (Perplexity).

The interesting pattern is the mix between high-volume tweeters (Musk, Altman) and CEOs who post less frequently but get amplified hard when they do (Amodei's longer essays, Srinivas's product-focused commentary). Zuckerberg's inclusion reflects his shift toward direct public communication on AI strategy over the past 18 months.

The implicit point: in AI, founder voice has become a measurable competitive asset. Companies whose CEOs don't show up publicly lose narrative ground to those who do.

The Anthropic vs. OpenAI Asymmetry

The study's most striking comparison is between the two revenue leaders. Anthropic's annualized revenue hit roughly $30 billion in April 2026, exceeding OpenAI's $25 billion as of February. Yet OpenAI's consumer user base is about 20 times larger, and its press coverage volume is "several multiples greater."

In other words: OpenAI dominates the cultural narrative; Anthropic quietly out-earns it on a per-user basis. The two companies communicate that reality very differently. OpenAI leans on consumer announcements, product launches, and Altman's personal megaphone. Anthropic leans on enterprise customer lists, safety publications, and Amodei's longer-form essays.

Both approaches work — they just produce very different public perceptions of who's "winning."

Why This Matters for Buyers and Regulators

For enterprise buyers, the 5WPR findings are a procurement tool. If a vendor refuses to disclose user counts, revenue trajectory, or customer references, that's a data point worth weighing — especially for multi-year platform decisions.

For regulators, the study quantifies what many already suspected: voluntary disclosure in AI is wildly inconsistent, and the loudest companies aren't necessarily the most transparent. That argues for standardized disclosure requirements, similar to how public companies file 10-Ks.

For competing labs, the findings highlight a clear strategic lever. Companies that have been quiet about adoption (Mistral especially) may be leaving narrative value on the table by not publishing structured updates.

What Industry Watchers Are Saying

Coverage has framed the study as confirmation of the "disclosure asymmetry" that has defined the AI sector since 2023. The AI Journal called it "the first structured benchmark" of AI communications, and Morningstar picked it up.

In AI policy circles, the report is being read as ammunition for calls to mandate AI-specific disclosure rules — something the EU AI Act partially does but US regulators have so far avoided.

What's Next

5WPR says this is the first edition of an ongoing benchmark. Expect updated rankings later in 2026 as new entrants reach scale.

If transparency becomes a buyer-driven differentiator, currently opaque companies will face pressure to publish more — or risk losing enterprise deals to labs that do.

The Bottom Line

AI's biggest companies are not on the same page about how to talk about themselves. A third disclose like public companies; a third say almost nothing; the rest sit between. For anyone evaluating AI vendors, knowing where each lab sits on that spectrum is now table stakes.