There are now more types of credentials in the United States than there are words in most dictionaries. Credential Engine's latest national count identified 1,850,034 distinct credentials offered by more than 134,000 providers, up from about 1.08 million just three years earlier. Digital badges alone passed one million, overtaking certificates (roughly 486,000) and degrees (roughly 264,000) as the largest category. Behind the count sits the money: the same report tallies $2.34 trillion invested annually in U.S. education and workforce development.
Read those numbers as a market analyst would. A sector with 134,000 sellers, near-doubling product variety in three years, and category growth concentrated in the cheapest-to-mint product line (the digital badge) is not a sector optimizing for scarcity of meaning. It is a sector optimizing for volume. And volume is exactly what a proof-of-competence market should not optimize for, because proof derives its value from discrimination: from the things it refuses to certify.
Follow the revenue, find the incentive
The certification economy's defining feature is that the issuer is paid at the moment of issuance, not at the moment the certified skill succeeds or fails in the world. A training provider collects tuition when the course completes. A badge platform collects fees as badges flow. A workplace-culture certifier, to take one documented example, charges companies a program fee for the survey and analysis behind a badge that lasts twelve months and must then be repurchased. In every case the revenue event is the credential, so the rational issuer maximizes credentials.
Nothing in that logic requires bad faith, which is what makes it durable. Each issuer can honestly believe in its rubric. But an industry paid per proof will, in aggregate, lower the effective difficulty of proof, multiply its categories, shorten its validity windows to create renewal revenue, and market the anxiety that makes proof feel necessary. Every one of those behaviors is visible in the count: 1.85 million varieties, subscription-style renewals, and a badge category growing precisely because it is frictionless to award.
The buyers' side completes the machine. Workers facing opaque hiring filters buy legibility. Employers facing evaluation costs buy pre-screened applicants. HR platforms parse credential keywords, so credentials proliferate to match the parsers. Each participant acts sensibly; the system output is inflation, in the strict sense of more tokens chasing the same underlying stock of competence.
The cost is paid in attention and trust
Who loses? First, the worker with modest resources, for whom the proliferating menu functions as a maze with tolls: 1.85 million options generate real spending on credentials whose labor-market value nobody can state, because with 134,000 issuers nobody, including employers, can hold the map. Credential Engine's own stated mission is credential transparency, and the size of the count is the size of the transparency problem.
Second, the genuinely rigorous credential loses. Rigor is expensive to build and invisible at a glance; in a crowded market the rigorous certificate and the weekend badge occupy the same line on a CV and the same keyword slot in a filter. Bad proof crowds good proof off the shelf, a Gresham's law of evidence.
Third, employers lose the thing they were trying to buy: reduced uncertainty. A signal that anyone can acquire signals little, so screening moves elsewhere, often to the very proxies (pedigree, network) that certifications were supposed to democratize away.
What proof without the complex looks like
The alternative is not the abolition of certification; regulated safety-critical skills genuinely need it. The alternative is remembering what proof is for and who should generate it. Evidence of work (portfolios, samples, trials, references from people who watched the work happen) is produced by the worker and the workplace, priced near zero, and impossible to mint in volume. Its verification cost falls on the evaluator, which is precisely why an industry grew up to relieve evaluators of it, and precisely why the relief keeps disappointing.
The test for any credential purchase, for a worker or an employer, is the incentive question: who is paid, and when, and does their payment depend on this proof meaning something in five years? For 1.85 million products, the honest answer is that the issuer's interest ends at issuance. What happens after issuance is the part that was supposed to matter.
Sources
- Credential Engine (2025). Counting U.S. Postsecondary and Secondary Credentials (Dec. 2025 report). 1,850,034 unique credentials from 134,000+ providers, up from 1.08M in 2022; badges (1.02M), certificates (~486K), degrees (~264K); $2.34T annual investment in education and workforce development. https://credentialengine.org/2025/12/09/new-report-finds-1-85-million-credentials-and-opportunities/
- Great Place To Work (accessed 2026). Certification FAQ. Example of the paid-badge model: program fee covers survey and analysis; certification is valid for 12 months, then renewable. https://www.greatplacetowork.com/solutions/certification-faq