Even OPENAI Can Be Too Descriptive: What Founders Should Learn from the EU Trade Mark Refusal
A famous name is not always a strong trade mark.
This is one of the most important lessons from the General Court’s judgment concerning the EU trade mark application for the word sign OPENAI.
OpenAI, Inc. applied to register OPENAI as an EU trade mark for a broad range of goods and services, including software, SaaS, PaaS, cloud computing, artificial intelligence-related technologies and certain identification and verification services. The application was refused in part by EUIPO on the basis that the sign was descriptive and devoid of distinctive character for the goods and services concerned. The Board of Appeal upheld that refusal, and the General Court dismissed OpenAI’s action.
The case is a useful reminder that trade mark law does not ask only whether a name is famous, commercially successful or associated in practice with one company.
It asks a more basic question first:
Can this sign function as a trade mark for the goods and services for which protection is sought?
Why was OPENAI refused?
The refusal was based on descriptiveness under Article 7(1)(c) of the EU Trade Mark Regulation.
Under that provision, signs that may serve, in trade, to designate characteristics of the goods or services cannot be registered as EU trade marks. The purpose is public interest: descriptive signs should remain available for all market participants to use.
The Court confirmed that the assessment must be made by reference to:
the goods and services covered by the application; and
the understanding of the relevant public.
In this case, the relevant public was the English-speaking public in the European Union, including both the general public and professionals.
The Court accepted EUIPO’s analysis that the relevant public would recognise the two elements of the sign:
“open”, which may mean available, unrestricted, freely accessible or, in a technological context, connected with openness, access or transparency; and
“AI”, a widely understood abbreviation of “artificial intelligence”.
When used for software, cloud services, SaaS, PaaS, technological services and related verification services, the sign OPENAI could therefore be understood as referring to accessible, unrestricted, transparent or explainable artificial intelligence.
That was enough.
The Court held that OPENAI was descriptive of the nature, function or purpose of the goods and services concerned because those goods and services could be based on, driven by or provided with the aid of accessible artificial intelligence.
Only one descriptive meaning is enough
One of the most important points in the judgment is that it does not matter if a sign has several possible meanings.
A sign must be refused if at least one of its possible meanings designates a characteristic of the goods or services concerned.
This is very important in practice.
OpenAI argued that the word “open” has multiple meanings and that the combination OPENAI should be treated as a composite, fanciful sign without one clear lexical meaning. The Court rejected that argument.
Even if “open” may have other meanings, that did not matter. What mattered was that, in relation to the goods and services covered by the application, one possible meaning of OPENAI was descriptive.
In other words, if one plausible meaning of the word element tells consumers about the nature, function, purpose or characteristics of the goods or services, the sign may fall under the absolute ground for refusal.
A brand name does not need to appear in a dictionary to be descriptive.
It does not need to be commonly used in everyday language.
It does not even need to be the only possible interpretation.
If the relevant public can immediately understand one descriptive meaning in the context of the goods and services, that may be enough.
Joining the words together did not help
Another practical point: combining words into one expression is not automatically enough to create distinctiveness.
The Court held that the terms “open” and “AI” were easily and immediately recognisable in the sign OPENAI. The expression followed normal English syntax: an adjective followed by a noun.
The absence of a space or hyphen did not make the sign fanciful.
For founders, this is particularly relevant.
Many tech and AI brands are built by combining descriptive or semi-descriptive words into a single expression. That may work very well from a marketing perspective. It may be clear, intuitive and easy to understand.
But the same clarity can become a legal weakness.
If the name is only the sum of its descriptive parts, it may be difficult to monopolise as a word mark.
Recognition is not the same as inherent registrability
OpenAI also relied on the recognition and reputation of the sign.
That argument did not succeed at this stage.
The Court explained that, when assessing descriptiveness under Article 7(1)(c), the relevant issue is the intrinsic characteristics of the sign in relation to the goods and services. Actual use and market recognition are not decisive in that assessment.
They may become relevant under Article 7(3), which concerns distinctive character acquired through use.
This distinction matters.
A sign may be refused because it is inherently descriptive, even if the market already associates it with a specific company. The applicant may still try to prove acquired distinctiveness, but that is a different, evidence-heavy route.
For a company operating at the scale of OpenAI, proving acquired distinctiveness may be possible. But for many startups, this route is expensive, uncertain and difficult.
The better strategy is to avoid the problem before launch.
Why this matters for founders
The OPENAI case is striking because it concerns one of the most recognisable companies in the world.
If even a global technology company can face this problem, founders should not assume that a name is legally safe simply because it is memorable, popular or commercially effective.
A descriptive name can be attractive at the beginning because it explains the product immediately.
It tells users what the business does.
It improves discoverability.
It may help with marketing, SEO and investor communication.
But from a trade mark perspective, that same name may be weak, narrow or difficult to protect.
The real problem often appears later, when the brand has already gained traction.
By then, the company may have invested in:
a website and domain,
product design,
app interfaces,
pitch decks,
advertising,
social media accounts,
packaging,
customer recognition,
PR,
investor materials,
international expansion.
At that stage, a trade mark refusal is no longer just a filing issue.
It becomes a brand architecture problem.
The company may continue using the name commercially, but it may not obtain the broad EU trade mark protection it expected for the word sign across all Member States. Enforcement may become more difficult. Licensing may become less clean. Investors may ask uncomfortable questions. Competitors may have more room to use similar descriptive language.
And if the company wants to rely on acquired distinctiveness, it must be ready to prove it.
This is where IP Scan becomes valuable
This is exactly the kind of risk an IP Scan is designed to detect early.
An IP Scan is not just an administrative review of what a company has already filed.
It is a structured audit of the intellectual property behind the business: names, logos, software, content, know-how, contracts, designs, trade marks, copyright and other intangible assets.
For founders, one of the most practical benefits is the ability to assess whether the brand name is legally fit for purpose before the business becomes too attached to it.
A proper IP Scan can help answer questions such as:
Is the brand name descriptive, weak or difficult to register?
Does the word element have enough distinctive character?
Would a figurative version be easier to protect than the word alone?
Are there earlier rights that create conflict risk?
Are the selected classes aligned with the actual business model?
Is the company protecting the right assets first?
What should be changed before the cost of change becomes too high?
The OPENAI judgment shows that trade mark risk is not theoretical.
It can affect even highly visible, well-funded and globally recognised companies.
For smaller businesses, the consequences may be even more serious because they usually do not have the same budget, evidence base or legal resources to fight a refusal or prove acquired distinctiveness across the European Union.
The practical lesson
A good brand name should do two things.
It should communicate.
But it should also distinguish.
If the name only describes the technology, function, audience or intended purpose of the product, it may be commercially useful but legally fragile.
That does not mean every descriptive or suggestive name must be abandoned. Sometimes the right strategy may be to combine a weaker word element with a distinctive logo, develop a stronger house mark, adjust the specification of goods and services, or build evidence of use over time.
But that decision should be made consciously.
Not after the brand is already everywhere.
Before launch, founders usually ask:
“Does this name sound good?”
The OPENAI case shows that they should also ask:
“Can this name actually be protected as a trade mark for what we sell?”
That is a legal question, not just a branding question.
And it is much cheaper to answer it before the market answers it for you.
