AI’s Valuations Surge, Leaving Pre-ChatGPT Startups Behind

Just five years ago, the venture capital landscape was awash with capital, fueling American startups with soaring valuations that often preceded profitability. From subscription lingerie services to intricate scheduling software, companies were showered with funding, many achieving “unicorn” status — a valuation of $1 billion or more — long before demonstrating a sustainable business model.

This was an era characterized by readily available cheap money and a pandemic-induced surge in demand, creating a frothy market for nascent businesses. Even as the Federal Reserve began its campaign of interest rate hikes in 2022, many founders remained optimistic, believing their companies could grow into their inflated valuations, according to insights shared with CNBC by investors.

Then, the landscape was dramatically reshaped by the advent of ChatGPT.

“The ChatGPT moment was a paradigm shift, signifying that the next generation of entrepreneurs would leverage natural language as their primary coding interface,” explained Samir Kaul, a partner at Khosla Ventures, an early investor in OpenAI. “We’re now witnessing scenarios where a team of 50 engineers can achieve what previously required 500. This fundamentally altered our approach to valuing these companies.”

While publicly traded software giants like Salesforce, ServiceNow, and Workday experienced significant stock price declines this year, partly due to the competitive threat posed by artificial intelligence, a more subdued but equally impactful reckoning has been unfolding in the private markets.

The recent AI boom, which saw over $250 billion channeled into key players like OpenAI and Anthropic ahead of their anticipated blockbuster IPOs, has left numerous startups, established before ChatGPT’s arrival in late 2022, in a precarious position. These companies, often burdened by pre-AI technology and inflated valuations, find themselves cut off from venture funding and not yet profitable enough to attract public market investors.

According to PitchBook data, there are 857 U.S. startups valued at $1 billion or more. However, PitchBook, a private markets data firm, reports that nearly half of these companies have not secured new funding in the past three years, rendering their valuations outdated.

PitchBook’s own valuation estimates indicate that startups that last raised capital in 2021 have seen their valuations decline by an average of 68%, while those from 2022 have experienced a 52% drop.

Consequently, PitchBook data shared exclusively with CNBC reveals that over 220 companies that once commanded billion-dollar valuations are now considered “fallen unicorns.” These estimates are derived from metrics such as headcount growth and comparisons with public market counterparts.

“Many of these companies are fundamentally pre-AI, not just in their cost structures but also in their product offerings,” commented Immad Akhund, CEO of Mercury, a banking services provider for early-stage venture-backed firms, which recently raised $200 million. “They are in a challenging situation. With all the attention focused on AI, companies that are not AI-first require exceptionally strong financial performance to attract investment.”

The roster of fallen unicorns includes prominent direct-to-consumer (D2C) brands such as Glossier, The Farmer’s Dog, Rothy’s, Brooklinen, and Savage X Fenty. These companies were emblematic of a trend that predicated their growth on the assumption that digital retail could achieve software-like profit margins.

The list also features companies frequently advertised on podcasts, including the nutritional supplement brand AG1 and the robo-advisor pioneer Betterment, alongside the online ticket marketplace SeatGeek.

These businesses rose to prominence in an environment that prioritized rapid growth and tolerated high valuations, underpinned by two core assumptions: sustained low interest rates and the consistent demand for engineering talent as an acquisition driver for larger tech firms.

However, the emergence of generative AI has fundamentally reshaped the venture capital landscape. Capital is now disproportionately flowing into AI-native companies, while making it exceedingly difficult for older startups to justify their prior valuations.

The hardest-hit sector appears to be enterprise software. Scheduling software startup Calendly, for instance, leads the pack among fallen unicorns in this category. PitchBook’s data identifies 75 Software-as-a-Service (SaaS) companies on its list, a figure double that of fintech companies, the next largest group. This concentration reflects both the substantial valuations commanded by software startups during the 2021 boom and the disruptive impact of generative AI on the sector’s foundational assumptions.

David Zhu, a former head of engineering at DoorDash, observed a seismic shift in the software industry following the “ChatGPT moment.” His analysis spanned startups, privately financed mid-sized firms, and major public SaaS companies, all of which he believed were facing significant disruption.

“My thesis was that all workflow-driven enterprise SaaS companies would either be disrupted or become obsolete within the next decade,” Zhu told CNBC.

The SaaS model, which often involves embedding services into employee workflows and charging on a per-user basis, is particularly vulnerable to the rise of autonomous AI agents. After leaving DoorDash, where he managed a team of over 200 engineers, Zhu founded Reevo, an AI platform designed to automate corporate sales and marketing functions.

Zhu argues that companies established before generative AI are encumbered by inefficient staffing models and legacy software, making it difficult for them to adapt. “Unless they undertake a radical, 180-degree pivot to rebuild their core offerings from the ground up, they are destined for gradual failure,” Zhu stated. “This reality means investors are more inclined to back new entrepreneurs at more reasonable valuations rather than reinvesting in older, struggling startups.”

The majority of the 20 fallen unicorns highlighted by CNBC either did not respond to multiple requests for comment or declined to provide a statement.

A spokesperson for Skydio, a drone manufacturer whose valuation PitchBook estimates have fallen from $2.5 billion to $509 million, issued a statement asserting, “This third-party speculation is inaccurate and not reflective of Skydio’s operations or the substantial growth we are experiencing in revenue and customer acquisition.”

While an AG1 spokesperson did not offer a specific comment for this article, Reuters reported following CNBC’s inquiry that the supplement maker was exploring strategic options, including a potential sale of the entire company or a significant stake, at a valuation of $2 billion, a figure that reportedly includes the company’s existing debt.

Investors and founders widely agree that companies which have not secured funding since 2021 or 2022 are unlikely to attract venture capital in the future. Without access to venture funding or a clear path to an IPO, many fallen unicorns face acquisition at a fraction of their previous valuations.

“When we observe companies ceasing fundraising activities, it’s a significant red flag,” noted PitchBook analyst Andrew Akers. He added that this typically indicates either stagnant or declining growth.

While some startups might be abstaining from fundraising due to robust profitability, Akers emphasized that this is an anomaly. “Beneath the surface, I anticipate a significant number of further revaluations,” he predicted.

This year has seen some encouraging signs of recalibration among certain startups.

In February, Stash, a financial technology platform focused on investing and savings, was acquired by Singapore-based Grab for an enterprise value of $425 million. This figure is notably below the approximately $660 million investors had previously injected into the company.

That same month, another fintech company, Step, was acquired by YouTube personality MrBeast. While the acquisition price remains undisclosed, investors speculate that it fell considerably short of the roughly $500 million the startup had raised prior to the deal.

“Many of these businesses simply do not hold the same value anymore, which is why we are seeing them being acquired at substantial discounts,” observed Ryan Falvey of Restive Ventures, an investment firm specializing in fintech. “Valuations have compressed significantly from their peak in 2021, when they traded at 50 times future revenues. This means a company with identical revenue is now worth approximately 85% less than it was just five years ago,” Falvey told CNBC.

Before this market reset, it was common for startups to be acquired by larger technology companies seeking to absorb their engineering talent. According to Khosla Ventures’ Kaul, such acquisitions typically commanded about $2 million per engineer, translating to a valuation of at least $200 million to $300 million for a firm with 100 engineers. This valuation floor, he explained, has largely evaporated with the advent of AI coding tools, which empower smaller teams to develop sophisticated products, thereby reducing the demand for large engineering departments and limiting exit opportunities.

Falvey notes that startups founded after the introduction of GPT-3.5 are now outperforming their older counterparts. He described investments made by his firm in the post-ChatGPT era as “undoubtedly the best” they have ever made.

“By 2023, we observed that companies we invested in post-ChatGPT were already generating more revenue than most of the companies we had invested in prior to ChatGPT,” Falvey shared.

Generative AI has the potential to significantly reduce the capital required to build successful software companies, challenging a fundamental tenet of the venture capital boom of the past decade. This market consolidation is likely just beginning, as the transformative impact of AI continues to ripple through the entire business funding ecosystem, from venture capital to private credit and even large public corporations.

Kaul posits that older software companies continue to rely on business models centered around per-employee user fees, an approach he believes will be undermined by AI as companies increasingly automate white-collar tasks. To remain competitive, software providers will need to transition to outcome-based pricing models and adopt AI-native infrastructure.

“My consistent question to any company presenting their case is: ‘Why can’t OpenAI, Anthropic, or Google achieve this?'” Kaul stated. “For the majority of them, the answer is sobering: ‘They can.'”