The artificial intelligence (AI) boom has turned Nvidia (NVDA) into one of the biggest winners on Wall Street, with its powerful graphics processing units (GPUs) becoming the backbone of the global AI infrastructure buildout. But as demand for AI computing continues to surge, competition for a share of this lucrative market is intensifying. Alphabet’s (GOOG) (GOOGL) Google is among the companies stepping up efforts to challenge Nvidia’s dominance, and the latest expansion of its custom AI chip lineup suggests the battle for AI computing supremacy is entering a new phase.
Google’s announcement came at a particularly sensitive moment for Nvidia. Just as the search giant unveiled its latest AI hardware, Nvidia was grappling with a controversy surrounding then-CEO Jensen Huang’s extramarital affairs. The revelations triggered a sharp selloff in NVDA shares, and the turmoil culminated in Huang’s resignation on Thursday.
But the implications extend well beyond Nvidia. Google’s latest move and Huang’s sudden resignation come as investors reassess the AI trade following a period of heightened volatility. Could intensifying competition and the loss of Nvidia’s longtime leader undermine the chipmaker’s investment case, or might these developments ultimately create new opportunities across the broader AI ecosystem? Let’s take a closer look.
Google’s New TPU Systems Signal a Bigger Push Into the AI Chip Market
On Monday, Alphabet’s Google unveiled two new versions of its tensor processing units (TPUs), custom-designed chips used to develop and run AI software. The announcement came as market leader Nvidia was grappling with a leadership crisis that ultimately led to the immediate resignation of CEO Jensen Huang.
Google said it expanded its eighth-generation TPU lineup with two new systems, the TPU 8tx and TPU 8i Max:
- The TPU 8tx was developed for government agencies and regulated industries that need to process sensitive information securely. It features hardware-level data protection and AMD's (AMD) EPYC processors, making it suitable for applications with strict security requirements. The system also allows developers to fine-tune models and train AI agents without having to transfer data between separate computing clusters.
- The TPU 8i Max combines two TPU 8i chips into a single package to deliver greater computing power and memory capacity. It is designed for workloads such as advanced coding assistants, complex reasoning tasks, and AI agents that process vast amounts of information. The chip also aims to address the so-called “memory wall,” a bottleneck that can limit the performance of large AI models.
Importantly, these are new system configurations based on existing TPU silicon rather than entirely new chip architectures. The TPU 8tx uses the same silicon as the training-focused TPU 8t, combines it with custom fifth-generation AMD EPYC CPU hosts, and incorporates the Boardfly network originally developed for the TPU 8i. Meanwhile, the TPU 8i Max combines two standard TPU 8i chips into a single package, as noted earlier. Google chip chief Amin Vahdat told Bloomberg in an interview that the company reconfigured the chips to handle workloads customers had been requesting. That means Google can respond more quickly to evolving customer needs by adjusting the systems built around its current line of processors rather than developing new chips from scratch.
The company introduced its original TPU 8 chips, developed in partnership with Broadcom (AVGO), earlier this year. The latest additions bring the number of TPU systems unveiled this year to four, highlighting a faster pace of product development at Google’s chip division. Previously, the division introduced a new AI processor every two years.
The expanded lineup comes just as Google ramps up efforts to market its AI hardware. The company recently announced that its TPUs would become available on Oracle’s cloud infrastructure platform beginning in 2027.
All of this tells us that Google wants a bigger slice of the pie in the lucrative AI chip market. And for good reason. According to Bloomberg Intelligence, Google’s TPU sales, including related software and hardware, are projected to exceed $135 billion by 2028, up from an estimated $12 billion this year. And that forecast doesn’t even include revenue from customers accessing the chips through Google Cloud.
Nvidia's Woes Complicate the Picture
As Google intensifies its challenge to Nvidia’s dominance in the AI chip market, the chip giant has endured a rough few days, and the worst may still be ahead. NVDA shares plunged a combined 9.7% on Monday and Tuesday following revelations that its CEO, Jensen Huang, had an affair with a businesswoman suspected of spying for China. The stock staged a modest rebound on Wednesday, but the controversy surrounding Huang deepened shortly after the closing bell. The New York Times reported that Huang had also disclosed a second relationship with an Nvidia executive. That appears to have been the final straw, as the chipmaker’s board announced Huang’s immediate resignation as president and CEO on Thursday, sending the stock down more than 4% at the time of writing.
On Sunday, Nvidia’s board launched an investigation into CEO Jensen Huang following his admission of an affair with a woman described in multiple reports as a suspected Chinese spy. The Wall Street Journal earlier reported that U.S. officials determined last year that Chinese national Alice Huang had been passing information to the Chinese government while working with Nvidia on data center development in Asia. Authorities in Singapore and Taiwan suspected that she had been recruited by Chinese intelligence services after discovering what they believed was a close relationship between her and Nvidia’s CEO. Nvidia’s board hired Joon H. Kim, a partner at Wachtell, Lipton, Rosen & Katz, to investigate whether Jensen Huang shared sensitive national-security information or confidential company information with Alice Huang. Separately, U.S. President Donald Trump told reporters on Monday that the Federal Bureau of Investigation was also looking into the relationship.
On Wednesday, The New York Times reported that Huang had an extramarital affair with an Nvidia employee in addition to one with the suspected Chinese spy. The employee was identified in the report as Kimberly Powell, a senior executive who heads the company’s healthcare business. Shortly afterward, The Wall Street Journal reported that Nvidia’s board was preparing to oust Huang as CEO. And what seemed unthinkable just a week ago became a reality on Thursday: Jensen Huang, the face of the AI boom, was out at Nvidia.
Google’s New TPUs Could Threaten Nvidia’s AI Chip Pricing Power
But let’s set aside the drama surrounding Huang and look at what Google’s latest TPU move means for the chip giant, all else being equal.
Nvidia’s GPUs have become the default choice for many AI developers thanks to their processing power, well-established CUDA software ecosystem, and widespread adoption. However, not every AI workload requires the same type of processor. Training a massive foundation model, running an AI chatbot, processing sensitive medical data, and powering an autonomous coding agent can each place very different demands on hardware.
And that opens the door for specialized chips. For example, Google’s TPU 8i Max is designed for workloads requiring massive context windows, while the TPU 8tx targets applications that demand strong security safeguards. If these systems deliver better performance per dollar or per watt for their intended workloads, customers could have a compelling reason to choose them over Nvidia’s GPUs.
With that, the key takeaway for NVDA investors is that intensifying competition could gradually erode the company’s pricing power in the AI computing market. Even as Nvidia continues to sell more chips, its gross margins could come under pressure if customers gain access to credible, lower-cost alternatives. Still, the chipmaker’s most pressing challenge now is finding a successor capable of executing as effectively as Huang did. Despite hitting NVDA stock with a downgrade, StoneX analysts expect no impact on Nvidia’s revenue or EPS estimates through 2028 from Huang’s resignation but warn that longer-term visibility has diminished. Nvidia Chief Financial Officer Colette Kress will serve as interim CEO while the board searches for a permanent successor.
What Google’s TPU Push and Huang’s Exit Mean for the AI Trade
Beyond the implications for Nvidia, Google’s latest TPU expansion highlights an important shift in the broader AI investment landscape. As AI workloads become more diverse, demand for specialized processors and supporting hardware could pick up. Google’s decision to incorporate AMD EPYC CPUs into its TPU 8tx system is a good example, showing how the rise of custom AI accelerators could create new opportunities for other chipmakers.
But an even bigger question is what Huang’s resignation means for the broader AI trade. For starters, AI demand shows no signs of slowing and will likely strengthen further regardless of who takes the helm at Nvidia. CoreWeave’s (CRWV) earnings report on Tuesday, along with a Bloomberg report that OpenAI’s quarterly revenue topped $10 billion for the first time, provided the latest evidence that AI demand remains strong. Nvidia’s rivals, meanwhile, could seize the opportunity to gain market share while the chip giant grapples with leadership turmoil. That helps explain why shares of Advanced Micro Devices and Broadcom are trading higher at the time of writing.
Finally, smaller AI cloud companies that have relied on Nvidia’s financial backing may face the greatest risk, as it remains to be seen whether a new CEO will continue Huang’s investment strategy, which has become known on Wall Street as “circular financing.”
On the date of publication, Oleksandr Pylypenko had a position in: NVDA , GOOGL . All information and data in this article is solely for informational purposes. For more information please view the Barchart Disclosure Policy here.