AI chipmaker Nvidia has finally broken its silence following the firm setting a new record for the biggest single-day loss of any company in history.

Chinese startup DeepSeek made a major splash over the last few days, claiming that its AI chatbot can rival the best OpenAI has to offer but at a tiny fraction of the cost. It also happens to be open source and free for all to use.

The announcement sent ripples across the industry, sparking a global selloff. The tech sector overall is on track for a "$1 trillion wipeout."

Nvidia, which develops and sells the chips AI companies use to train their increasingly compute-intensive models, has chosen the high road — even after the startup eviscerated its stock.

"DeepSeek is an excellent AI advancement and a perfect example of Test Time Scaling," an Nvidia spokesperson told CNBC today.

Put simply, test time scaling refers to a new trend in the AI world allowing models to test and verify different responses before arriving at a final answer, a primitive form of "reasoning" that can lead to more reliable answers. Both OpenAI's o1 model and DeepSeek's latest R1 model make use of forms of test-time scaling.

"DeepSeek’s work illustrates how new models can be created using that technique, leveraging widely-available models and compute that is fully export control compliant," the spokesperson added.

It's a strikingly measured response, considering the sheer amount of money at stake.

Further complicating the situation is that DeepSeek is reportedly making use of a capped version of Nvidia's ever-popular H100 AI chip, called the H800, which was designed for the Chinese market. That's despite the US government banning the export of both chips in late 2023.

Today's stock bloodbath highlighted widespread skepticism over soaring costs in the AI industry. Just last week, US president Donald Trump touted an AI infrastructure project dubbed Stargate, which aimed to raise $500 billion in just four years, with Meta CEO Mark Zuckerberg committing to spending $60 billion of his own on AI this year.

But DeepSeek's claims have investors spooked, leading to some tough questions: have they been massively overspending? And when will generative AI actually break even, if ever?

Some analysts are hopeful that DeepSeek's advancements could clear a slightly less murky path to profitability.

"If model training costs prove to be significantly lower, we would expect a near-term cost benefit for advertising, travel, and other consumer app companies that use cloud AI services, while long-term hyperscaler AI-related revenues and costs would likely be lower," Bank of America Securities analyst Justin Post wrote in a note, as quoted by CNBC.

Nvidia's statement doesn't appear to have immediately calmed investors. The company's shares have yet to rebound, cementing a catastrophic 16 percent loss (although to put the ascendant AI sector in perspective, that's just four months of gains.)

DeepSeek's explosion onto the scene has left a massive black eye, and OpenAI is likely still reeling from the whiplash. CEO Sam Altman has been noticeably quiet since DeepSeek rocketed past ChatGPT to score the top iPhone App Store spot.

In short, the dust hasn't even begun to settle as investors are forced to do some serious soul-searching.

More on DeepSeek: Trump Embraced AI and It Exploded Spectacularly in His Face


Share This Article