Skip to main content
TipRanks
Go Ultimate75% OFF

Market Currents: NVISIONING the Bottom?

Market Currents: NVISIONING the Bottom?
Image source: Shutterstock

The terrible, horrible, no good, very bad week for stocks: the AI trade loses its Godfather, chips lose their loopholes, and Anthropic’s debut loses its pop.

Stocks suffered a brutal selloff, notching one of their worst weeks outside of recessions. The Dow Jones Industrial Average (DJIA) declined for the fifth straight day on Friday, wrapping up the weekly loss of 8.02%. The S&P 500 (SPX) logged its deepest seven-day slide since February 2020, including the weekly drop of 8.61%. The Nasdaq-100 (NDX) tumbled 10.43% – one of the largest weekly declines in the past 40 years – with much of that loss attributed to Friday’s broad selloff in technology stocks. The large-cap tech index is on the verge of a bear market, having lost 19.8% since its October 6 peak. The extent of the selloff has erased nearly all of the year-to-date gains of the DJIA and the S&P 500, while pushing the Nasdaq-100 into the red.

Roughly 40% of the market cap loss in the S&P 500 since its recent high is attributable to NVIDIA (NVDA). The AI chip giant erased $905 billion in market cap over the past week alone, and about $2.8 trillion since October 6. Despite a near-perfect beat-and-raise, the chip leader – reporting for the first time in its 33-year history without Jensen Huang – failed to soothe investor fears regarding both its own future and the broad AI ecosystem. Operationally and financially, NVIDIA’s report was more than stellar, with the guidance upgrade reflecting soaring demand alongside improving supply availability.

However, what past results and near-term guidance revealed was already known, and that’s part of the problem. The quarter reaffirmed the extraordinary strength of the technological and strategic roadmap built and implemented by Jensen Huang with such brilliance that it now allows the most important tech company in the world not only to chug along, but to accelerate into the next fiscal year. Beyond that horizon, however, no one knows how much of NVIDIA’s technological leadership depended on Huang’s personal ability to anticipate and shape technological change – or how much of that capability the company retains without him. That’s why the NVIDIA board’s announcement that it doesn’t plan to bring Jensen back squashed the stock’s attempted rebound, which began on rumors of the rock-star CEO’s possible return.

Interim CEO Colette Kress reaffirmed NVIDIA’s commitment to supporting AI infrastructure financing but promised greater selectivity and discipline in future investments and guarantees. Markets inferred that there would be much less financing and fewer backstops for the weaker AI names and AI startups, as NVIDIA continues confronting its own mess, stretching well beyond leadership uncertainty – and even beyond SEC, FBI, and congressional investigations.

One notable fallout from the Jensen Huang controversy was the sudden urgency within the U.S. government to crack down on semiconductor export loopholes, which allowed Chinese AI developers to access cutting-edge U.S. AI chips remotely through data-center leasing. The U.S. Commerce Department now requires re-export licenses for markets including Singapore, Malaysia, Thailand, Indonesia, and Brazil. As of now, the extent of that gray-bordering-on-black market demand remains unknown, but it’s likely far from negligible. A slate of media reports has tied China’s frontier progress directly to two channels: remote access to NVIDIA’s most powerful AI accelerators via Malaysian and Singaporean data centers, and hardware and training data smuggled to China through third countries, with the help of actors like the now-infamous Megaspeed.

The fear is that tighter export controls will lower chip demand and depress prices – which over time would lead to lower investment and even an earnings slowdown. Adding to the anxiety, President Trump promised to impose massive taxes on U.S. companies building data centers abroad to prevent crucial tech from falling into Chinese hands. The iShares Semiconductor ETF’s (SOXX) slide accelerated, wiping out over 17% during the week and extending its drawdown from its June peak to roughly 43%. Even if Friday turns out to have been a trough, this will still go down as one of the toughest chip selloffs of all time.

Anthropic’s (ANT) historic IPO managed to do two opposite things at once: confirm that the AI narrative is alive and well and leave much of the AI trade bleeding. The Nasdaq debut briefly lifted the Claude maker’s market cap above $2 trillion, sucking in institutional and retail flows that rotated out of other tech names to fund ANT allocations. But while the AI lab raised an extraordinary amount of money, the party didn’t last long. Bearish tech sentiment prevailed, and the stock had shed more than 20% from its opening pop by Friday’s close. Below the surface of Anthropic’s mind-blowing frontier-model capabilities and surging revenues lie the same hard questions that have been weighing on the trade for the past few months: when will surging demand for AI compensate for the eye-popping costs of data centers, chips, networking equipment, and electricity needed to power the technology?

Companies across the AI industry are spending hundreds of billions to build and run AI infrastructure before they see a return on that investment on their bottom lines. And that’s happening at a time when the macro setup has driven financing costs higher, while lenders are becoming increasingly reluctant to bankroll tech capex without checking thoroughly under the hood. The risk isn’t confined to a few neoclouds merging or going out of business – it’s a more systemic one: less access to financing means fewer orders for the chipmakers and equipment suppliers counting on the AI buildout. A slower and/or more expensive data-center expansion means frontier-model development and training may face even higher costs. Between the anti-data-center backlash and the hawkish Fed minutes, that risk is becoming more material by the hour.

Which brings us back to the question: how much of NVIDIA’s role as the “AI central bank” will disappear? And it doesn’t center on financing alone, although roughly $165 billion in backstops, commitments, and guarantees – alongside orders, direct investments, and more – were among the key supports of the AI buildout. Beyond money, Jensen Huang’s currency in the AI trade was confidence. Just like a central banker, NVIDIA’s cofounder could stabilize a momentary wobble in the trade with words alone. While NVIDIA may continue providing its guarantees and financing to a large extent, the “Godfather of AI” is gone for good.

Currently, the stars have aligned quite negatively for the AI trade and for the broader tech stocks, which have more or less coalesced around the narrative in the past few years. Given the unprecedented weighting of technology in broad-market indexes and the growing alignment of previously non-tech industries – like utilities and builders – with technology, a flagging tech trade should weigh heavily on the stock market in general, with few, if any, insulated niches. A bear market is one small jolt away.

Still, there’s hope. Stocks are extremely oversold, and the so-called “Fear & Greed Index” is pointing to extreme fear. All this while the U.S. economy is growing at a solid clip, unemployment is near historic lows, and the S&P 500 companies are reporting their strongest EPS growth in any quarter outside post-crisis rebounds, having beaten estimates for 13 consecutive quarters. Meanwhile, the average forward P/E of the tech sector has collapsed to its lowest since March 2020. The AI revolution is now undoubtedly facing one of its biggest tests to date, if not the biggest one. However, the market appears to be pricing in a far more pessimistic outcome than what the underlying fundamentals currently suggest. Unless something unexpectedly breaks, this is a perfect setup for a powerful bounce.

Disclaimer & DisclosureReport an Issue

Add as preferred source
Related Articles