The Dangers of the Generative AI Boom: Increasing Calls for Caution

Soon, Silicon Valley may reflect on recent developments as the moment when the enthusiasm for generative AI might have become excessive.

This summer, there was skepticism among investors about whether top AI stocks could maintain their lofty valuations due to minimal returns on significant AI investments. As Autumn draws near, key AI industries—such as chip manufacturing, large language models (LLMs), and AI devices—have regained support, though there are growing reasons for prudence.

Cerebras: A chip competitor with substantial risk

Cerebras, a chip startup, is aiming to compete against Nvidia by creating processors intended to enhance sophisticated LLMs. Nvidia, a significant entity in the AI surge, witnessed its market capitalization soar from $364 billion at the start of 2023 to over $3 trillion.

Cerebras, however, heavily depends on a single client: the Abu Dhabi-based AI company G42. In 2023, G42 was responsible for 83% of Cerebras' revenue, and in the first half of 2024, this increased to 87%. Even with G42's backing from industry giants like Microsoft and Silver Lake, this reliance presents a risk. While Cerebras has secured a deal with Saudi Aramco, the dependency on one customer could be worrisome as it seeks a $7-8 billion valuation for its IPO.

OpenAI’s significant funding—but with conditions

OpenAI grabbed headlines by raising $6.6 billion with a valuation of $157 billion, marking the largest investment round in Silicon Valley history. However, the company has advised investors against backing rivals like Anthropic and Elon Musk's xAI—an atypical move in venture capital, where diversification is usual. Critics like Gary Marcus labeled this strategy as "running scared."

Among OpenAI's supporters are "bubble chasers" such as SoftBank and Tiger Global, known for investing in companies at peak valuations, often leading to significant losses. With key figures like CTO Mira Murati stepping down and projected losses amounting to $5 billion this year despite increasing revenues, OpenAI is facing significant hurdles.

Meta’s substantial investment in AI wearables

Meta joined the AI competition by introducing Orion, its augmented reality glasses. These wearables aim to bring AI into everyday living, receiving an endorsement from Nvidia’s CEO Jensen Huang. Nonetheless, a production cost of $10,000 per unit poses a significant barrier.

Meta must reduce production costs and overcome consumer reluctance, since earlier AI-powered wearables—like Snapchat's glasses, Google Glass, and the Humane AI pin—struggled to attract consumers.

The challenges ahead

What lies ahead for AI? OpenAI must demonstrate it can substantiate a $157 billion valuation while facing financial losses. Cerebras needs to convince investors that relying on a solitary client won't be detrimental. Meanwhile, Meta must persuade consumers to embrace an entirely new mode of interacting with AI.

If these entities triumph, this period could herald a pivotal point in the AI evolution. However, as seen in tech history, navigating high-stakes markets is seldom straightforward.