Paper Unicorns: A $1 Billion Valuation Is Not $1 Billion in Revenue

By Olivia Hart

The price of hitting a fixed level of AI performance fell 13-fold every year since 2023, according to an Epoch AI study cited by Ilman Shazhaev. For startups that just resell model access, he argues, falling prices expose the company to severe price competition.

Shazhaev sets a higher benchmark as 250 companies joined The Crunchbase Unicorn Board in the first half of 2026. A $1 billion valuation still makes a startup a unicorn on paper, he says; his higher test is $1 billion in revenue sustained by customers who keep paying.

His question for investors is blunt: “What will customers keep paying for?” Shazhaev spent five years at Harbin Institute of Technology and Shanghai Jiao Tong University, where “fail fast, improve faster” shaped his approach to building technology.

What does revenue prove?

Fast revenue deserves attention, Shazhaev says, because founders who convert pilots into paying deployments solve a real problem. Strong sales can also make the next fundraising round easier. Neither outcome proves durability when an annualized run rate depends on discounts, experimental budgets or generous compute subsidies.

Shazhaev cites OpenRouter and its $113 million Series B: developers can already swap between models through one shared interface. That interchangeability, in his view, makes price the obvious competitive pressure for companies with little beyond model access. He considers software embedded in customer approvals, records and daily decisions harder to replace than the model doing the underlying work.

Shazhaev recommends three checks before assigning lasting value:

• Renewals at full price
• Contract expansion
• Margins after inference costs

His takeaway is plain: paper revenue can win attention, but full-price renewals test whether the business can retain its customers.

Why should funding cross borders?

Once founders know which customers keep buying, Shazhaev advises raising across regions because a local backer can double as a route into that market. Borders barely slow an AI product down, he says, but enterprise sales still depend on local budgets, procurement procedures and relationships. He values a foreign investor when those connections convert a demonstration into a paid, renewable contract.

The $30 million Series A for London- and Gulf-based 1001 AI offers the example he favors. U.S. venture firm Lux Capital led the financing, alongside Saudi PIF-backed Sanabil Investments, with engineers in London and customers in the Gulf. Shazhaev cautions that the deal cannot establish future retention by itself; it shows only how capital, talent and customers can sit on different maps.

He would ask every investor what they bring besides money. One warm introduction can beat a higher valuation, he says, if it turns into a contract the company can actually deliver and renew.

For Shazhaev, a broader cap table helps only when investor access matches customer demand and the founder can still expand.

What obligations come with the money?

Shazhaev rejects a simple choice between fast venture capital and patient sovereign money. In his view, the practical difference lies in obligations: venture funds need returns within their fund lives, while some sovereign-backed funds carry domestic-development mandates. He notes that neither category is uniform; many sovereign funds pursue conventional returns, and traditional VCs can tolerate long enterprise sales cycles.

U.S. and European venture investors have asked Shazhaev how quickly his company can turn product adoption into repeatable sales before competitors catch up. Commercial momentum and a sizable exit guide those funds’ decisions.

He says that pressure can sharpen execution, but it can also push founders toward fast expansion without counting delivery costs. The result may be a pipeline of discounted experiments instead of durable enterprise relationships.

Sovereign-backed investors tend to ask where user data is processed, which technical capabilities must remain in-country and whether local customers can rely on the service years later. Some funds carry explicit domestic-development mandates and treat those concerns as investment criteria.

Shazhaev says the benefit can be deeper local infrastructure and enterprise adoption; the cost can be longer diligence, localization spending and contract terms that impede expansion elsewhere. Government-affiliated capital also offers no automatic procurement win.

Shazhaev says founders should read governance rights, exclusivity requests and promised commercial introductions before accepting a check. The relevant test is whether the investor’s regional reach justifies those constraints, not whether the capital carries a familiar label.

For Shazhaev, a larger investor network helps only when its obligations leave the founder room to grow.

What should investors fund?

Shazhaev’s preferred sequence starts with a working product, then follows the customer through renewal. Venture money can accelerate development and sales once the business model works. Regional strategic investors become more useful when infrastructure and procurement rules make local expertise essential.

He would show investors a small cohort renewing without discounts rather than a glossy forecast dependent on the next funding round. His watch list:

• Net revenue retention
• Production deployments
• Margins after inference and delivery costs

Paper unicorns will continue to appear and disappear each quarter, Shazhaev predicts, as market sentiment changes. Teams that reach $1 billion in lasting annual revenue will likely combine venture momentum with regional obligations, leaving global cap tables that mirror the enterprise markets they serve.

Shazhaev is the founder and CEO of Dizzaract, an AI and gaming infrastructure company headquartered in Abu Dhabi. His biography describes Dizzaract as the Middle East’s largest game development studio and says his work spans distributed AI infrastructure, human behavioral data and interactive environments where AI systems are trained and tested.

He studied in China before building an AI company in Abu Dhabi. He also serves as a United Nations expert under the UNODC program, advising on AI for social problems in developing economies. Shazhaev has written 46 scientific articles, including research on AI reading a player’s psychological state from ordinary gaming hardware, and holds 10 registered invention patents.

The 250 companies added to The Crunchbase Unicorn Board in the first half of 2026 must clear Shazhaev’s test: renewal at full price, production deployments and margins after delivery costs.