The hum of servers filled the air as the lead engineer, Sarah Chen, stared at the thermal readings. The new M300 chip, slated for a late 2026 launch, was running hotter than expected. “Damn,” she muttered, “or maybe that’s how the supply shock reads from here.”
It’s a scene playing out across AI startups. The pressure to achieve unicorn status is intense, and some founders are using a novel valuation mechanism to do it. The practice, which involves selling the same equity at two different prices, is drawing scrutiny from investors and analysts alike. The core problem: It inflates valuations, making a company appear more valuable than it might be in a traditional funding round. This allows them to raise capital at a higher valuation, which is particularly attractive in a market where valuations are already high, and investor confidence can be fragile.
The mechanism often works like this: A company will sell equity to different investors at different prices, effectively creating two separate valuations for the same shares. Early investors, who take on more risk, might get a lower price per share. Later investors, riding the wave of the company’s perceived success, pay a premium. The difference in price can be substantial, with some companies seeing as much as a 20% to 30% difference between the two prices, according to a recent analysis by Deutsche Bank.
The practice isn’t illegal, but it’s raising eyebrows. “It’s a bit like selling the same house twice,” said Mark Johnson, a senior analyst at JPMorgan, during a recent briefing on the AI market, “except the second buyer doesn’t know the first one got a discount.” Johnson pointed out that the practice is most common in companies with strong growth projections, especially those in the AI space, where the promise of future returns is high, and the underlying technology is complex enough to obscure the valuation discrepancies.
The implications are significant. First, it can create a false sense of market value, which makes it easier for companies to raise capital. Second, it can distort the playing field, making it harder for companies with more realistic valuations to compete. Third, it can create an environment where investors are incentivized to chase the highest valuations, rather than focus on the underlying fundamentals of a business.
Consider the Baidu/Kunlunxin example. Baidu’s Kunlunxin chips, designed for AI applications, have faced the same supply-chain constraints and export controls as other advanced chips. The initial M100 chip, launched in 2024, saw limited production due to restrictions. The company projected a significant increase in capacity for the M300, scheduled for a 2026 release, aiming for a 40% increase in sales. However, the reliance on SMIC, a Chinese chip manufacturer, puts them at odds with U.S. export controls. These macro issues are the same ones Chen and her team are wrestling with, right now.
The practice also raises questions about transparency and accountability. Investors have a right to know the true value of their investments, and the current mechanism makes it difficult to ascertain that value. This lack of transparency can erode trust in the market, making it harder for companies to raise capital in the future. The pressure to achieve unicorn status, coupled with the complexity of AI technology, is creating a perfect storm for this kind of valuation manipulation.
What happens next? Expect more scrutiny from regulators and investors. The SEC and other regulatory bodies may step in to create more stringent guidelines. Investors will likely demand more transparency from the companies they invest in, and the market may begin to differentiate between companies with realistic valuations and those that are inflating their worth. The real value is always in the thermal readings.
About the Author
