Artificial intelligence has become the financial Rorschach test of our time. Investors see inevitability while founders see destiny and bankers see fees. To make matters worse markets see numbers that seem to defy gravity or market fundamentals. Everyone is looking at valuations through their own telescope of rationalization.
The question is whether the lofty AI valuations are rational or being rationalized and what is the rationale?
At the heart of the current AI valuation frenzy lies a critical question: are these soaring numbers rational, derived from disciplined financial analysis, or are they rationalized, shaped more by emotional distortion and self-fulfilling narratives than by economic fundamentals?
The “Telescope Model of Rationalization” (described by Hedley and Potarazu) helps illuminate this dilemma by showing how perceptions can be altered through internal “lenses” — core beliefs, emotions, and thinking errors that bend logic much like curved glass bends light. Just as these psychological lenses can distort an executive’s view of business reality and justify inflated claims about performance, they can also warp investor and founder beliefs about the future of AI, causing optimism and fear of missing out to overshadow disciplined valuation analysis. In this landscape, lofty AI valuations may reflect not just projections of future profits, but also a collective distortion in how opportunity, pressure, and self-justifying narratives interact to shape market pricing
AI companies are commanding extraordinary valuations under structures that are often opaque, rapidly changing, and difficult to benchmark against traditional financial metrics. Beneath the excitement lies a more complicated question: are these valuations grounded in math—or myth?
One of the defining features of the current AI boom is opacity. Unlike mature public companies, many high-profile AI firms disclose limited financial detail. Revenue figures are selectively reported, margins are unclear, customer concentration is rarely transparent, and capital expenditures—particularly for compute infrastructure—are often embedded in complex arrangements. This is a traditional challenge for entrepreneurs to not get caught in the web of future potential to not focus on reality and fundamentals of the present.
READ: Sreedhar Potarazu | India’s AI content crackdown: Why it matters to Big Tech and the US (February 10, 2026)
Investors may rely heavily on forward projections, total addressable market estimates, and strategic positioning rather than profitability. In this environment, valuation becomes less about discounted cash flow and more about velocity and the potential just as we saw in the internet boom. The trap that many of us as entrepreneurs found ourselves in can be a dangerous one.
Another emerging dynamic is the rapid escalation of valuations across near-term funding rounds. Startups raise capital in a round led by a prominent investor at one valuation—only to return weeks later with a significantly higher price tag, often anchored by a new strategic participant.
These step-ups can occur before material changes in fundamentals. The signaling effect of a lead investor can create a feedback loop: prestige drives demand, demand drives scarcity, scarcity drives price. The result is a valuation curve that appears exponential, even if underlying financial performance is linear. This was the longstanding mantra in the valley that “perception drives momentum and momentum drives reality.”
Traditionally, accounting and valuation theory rest on the concept of a transaction between a willing buyer and a willing seller, neither under compulsion and both with reasonable knowledge of relevant facts. But what happens when the willing buyer believes they are securing privileged access to a transformative technology? Or when access to the deal itself becomes the scarce commodity?
If a buyer knowingly accepts a premium to gain strategic positioning, is that still fair value—or is it strategic value? The distinction matters. Accounting frameworks assume rational market participants operating with symmetry of information. But often traditional metrics are not in synchrony with the methods being used for AI startups with little revenue and no profit. Today’s AI marketplace often operates with asymmetry, urgency, and competitive fear of missing out.
Complicating the picture further is the widespread use of stock-based compensation. Many AI startups, flush with investor capital but prioritizing growth over profitability, compensate employees with equity grants that may ultimately convert into substantial payouts.
While equity incentives are a legitimate tool to attract talent, they also dilute ownership and can obscure the true economic cost of labor. When employees are effectively paid in equity that is valued at rapidly increasing private-market marks, the notional wealth creation feeds the valuation narrative itself. Moreover some employees are getting the opportunity to sell these shares in public offerings sooner.
A stark example of how valuation narratives can diverge from traditional financial anchoring is evident in move by Anthropic announced today kicks off share sale for staffers of up to $6 billion.
The AI company, recently valued at roughly $380 billion in a $30 billion funding round, has arranged a secondary market transaction where outside investors will buy up to $5-6 billion of employee shares at an implied $350 billion valuation — even though no new operating performance has been independently verified to justify that price.
What this reveals is not merely employee liquidity, but how private markets can re-affirm sky-high valuations through subsidiary transactions that lack the rigorous price discovery of public markets.
Such secondary share sales simultaneously reinforce inflated marks and create the illusion of widespread “confidence,” yet they may not reflect broader market prices or disciplined valuation analysis. In other words, the very mechanisms that are touted as proof of economic value — high valuations anchored in recent deals — can themselves be products of narrative momentum rather than underlying profitability or cash-flow substantiation.
READ: Sreedhar Potarazu | AI sovereignty race: US and China lead, India watches (
Large technology firms, like Meta, add another layer to the distortion. Stock-based compensation is expensed under accounting rules, but it does not represent an immediate cash outflow. On cash flow statements, these equity payouts are often added back in operating cash flow calculations because they are non-cash expenses.
The result is that reported cash flow can appear robust even when substantial economic value is being transferred to employees in the form of equity. When investors focus on adjusted cash metrics without fully appreciating dilution and long-term cost, valuations can drift further from underlying economic reality.
My own experience has taught me that valuation is rarely a single number and can have many consequences. It is an opinion—sometimes many competing opinions—about future profitability. In traditional business analysis, sustained profitability and free cash flow ultimately anchor value.
Yet in today’s market, numerous companies with limited or no profitability command valuations that rival—or exceed—long-established enterprises with consistent earnings. History reminds us that narrative-driven markets can persist longer than skeptics expect, but fundamentals eventually reassert themselves.
Read more columns by Sreedhar Potarazu
The math of valuation has not changed: long-term enterprise value is a function of expected future cash flows, discounted for risk. What has changed is the degree to which projections, optionality, and strategic positioning are substituting for realized earnings.
AI may well reshape industries and justify extraordinary valuations over time. But if the numbers are built more on competitive signaling and capital velocity than on demonstrable economic return, the myth may eventually collide with the math.
In the end, markets do not punish innovation—they punish overconfidence. AI’s promise is real. Whether its valuations are, remains an open question.