OpenAI is showing growing dissatisfaction with some of Nvidia’s latest artificial intelligence chips, marking an important shift in the balance of power shaping the global AI boom, and introducing fresh uncertainty into a partnership that has, until now, symbolized the sector’s rapid rise.

The development is also emerging as a meaningful stress test for the chipmaker’s dominance, and a sign that the next phase of the AI boom may be shaped less by who trains the biggest models and more by who can run them most efficiently.

Eight sources familiar with the matter say the ChatGPT-maker has, since last year, been exploring alternatives to Nvidia hardware for parts of its computing needs, with a particular focus on inference, according to Reuters. Inference is the stage where trained AI models generate responses to user prompts, power applications, and handle real-time workloads. It is also where costs scale rapidly as usage grows, making chip efficiency and pricing critical.

Nvidia remains the clear leader in chips used to train large AI models, an area where its GPUs and software ecosystem have become deeply entrenched across the industry. But inference has increasingly become a separate and more commercially sensitive battlefield. As AI tools move from experimental deployments into everyday consumer and enterprise use, inference workloads now account for a rising share of total computing demand.

Sources say OpenAI believes some of Nvidia’s newer chips are more heavily optimized for training rather than the kind of high-volume, always-on inference workloads that now dominate its operations. That perception has pushed OpenAI to assess other options, including specialized inference chips and alternative suppliers that promise better performance-per-watt or lower operating costs.

The shift underscores a broader challenge facing leading AI developers. Running models at scale is expensive, energy-intensive, and difficult to optimize. For OpenAI, whose products serve hundreds of millions of users, even marginal gains in efficiency can translate into significant savings. Inference chips, unlike training hardware, must balance speed, cost, and power consumption, particularly as regulators, customers, and investors scrutinize energy use more closely.

This search for alternatives comes at a sensitive moment for Nvidia. The company has built its position not just on powerful hardware, but on an integrated stack that includes software, developer tools, and tight relationships with top AI labs. Any sign that its most high-profile customer is hedging its bets, even partially, introduces uncertainty about how durable that dominance will be as AI matures.

It also complicates a parallel set of discussions between the two companies. In September, Nvidia said it intended to invest as much as $100 billion in OpenAI, a deal that would give the chipmaker an equity stake while providing OpenAI with capital to secure scarce advanced chips. The talks highlighted how interdependent the two companies have become, with OpenAI relying on Nvidia’s supply and Nvidia benefiting from OpenAI’s scale and influence in shaping AI workloads.

OpenAI’s exploration of alternatives does not amount to a break with Nvidia. Sources stress that Nvidia remains central to OpenAI’s training infrastructure and will continue to supply a significant portion of its computing needs. Switching inference infrastructure at scale is technically complex and would likely happen gradually, if at all. Still, the move sends a signal across the AI sector that even Nvidia’s largest customers are unwilling to rely on a single supplier indefinitely.

The implications extend beyond the two companies. Inference is widely expected to become the dominant source of AI chip demand over the coming decade, as models are deployed across search, productivity software, customer service, and consumer devices. That creates an opening for rivals and for new chip designs tailored specifically to inference workloads, rather than the brute-force training that has defined the AI arms race so far.

The moment is believed to underline the need for Nvidia to defend its leadership while adapting to a market that is shifting from building models to running them cheaply and reliably at scale. It also reflects OpenAI’s strategic effort to control costs, reduce dependency, and maintain flexibility as AI usage continues to grow.

Palantir Technologies’ shares surged 11% in premarket trading on Tuesday after the data analytics firm delivered a stronger-than-expected fourth quarter, underscoring how rising government and enterprise spending on artificial intelligence is translating into real revenue momentum — even as political and ethical scrutiny around its work intensifies.

The company reported fourth-quarter revenue of $1.41 billion, comfortably ahead of Wall Street expectations of $1.33 billion, according to LSEG data. The results capped a volatile stretch for the stock, which had slumped sharply late last year amid broader concerns that enthusiasm around AI-linked software companies was running ahead of fundamentals.

Despite November marking Palantir’s worst month in two years, the stock still ended 2025 up 135%. That rally has cooled somewhat in early 2026, with shares down about 17% year to date at Monday’s close before the earnings-driven rebound.

Chief executive Alex Karp struck a confident tone following the release, calling the performance “the best results that I’m aware of in tech in the last decade” in an interview with CNBC. His remarks reflected a belief that Palantir has moved beyond speculative hype into a phase where AI adoption is driving durable, large-scale contracts, particularly in the public sector.

Founded as a data intelligence company focused on complex analytics for governments, Palantir has increasingly positioned itself as a core infrastructure provider for AI-driven decision-making. Its software is used by U.S. agencies, including the Department of Defense, the Internal Revenue Service, and the Department of Homeland Security, as well as by corporate clients seeking to integrate AI into operations.

That government focus is now a central pillar of its growth story. Palantir said U.S. government revenue rose 66% year on year, highlighting accelerating adoption of its platforms across defense, security, and public administration. The company has secured a string of large, multi-year deals that have bolstered confidence in the visibility of future earnings.

In July, Palantir signed a software contract worth up to $10 billion with the U.S. Army, one of the largest agreements in its history. That was followed in December by a $448 million contract with the U.S. Navy aimed at accelerating shipbuilding production, a deal that underscored the Pentagon’s push to use AI and data analytics to modernize procurement and logistics.

Investors have long wrestled with Palantir’s valuation, which has often been described as stretched relative to traditional software peers. Yet some analysts now argue that the company’s pricing looks more defensible in the context of the wider AI ecosystem.

“Although Palantir’s valuation is still frothy, it appears more reasonable relative to recent venture rounds for companies tied to the AI ecosystem,” said Louie DiPalma, an analyst at William Blair, in a note published ahead of the earnings release.

He added that Palantir’s operating margin could expand from around 50% to as much as 65% over the next five years, driven largely by growth in high-margin government and defense contracts.

That margin story is central to Palantir’s longer-term appeal. Unlike many AI-focused firms that continue to burn cash as they chase scale, Palantir has emphasized profitability and disciplined cost control, arguing that its platforms are already deeply embedded in customer workflows and therefore expensive to replace.

Still, the company’s close ties to law enforcement and immigration agencies remain a flashpoint. In recent weeks, Palantir’s work with U.S. Immigration and Customs Enforcement has drawn renewed attention following protests in Minneapolis, where federal agents shot two demonstrators during clashes tied to immigration enforcement operations.

While Palantir was not directly involved in the incidents, the controversy has revived broader debates about the role of private technology firms in surveillance, policing, and immigration control.

This has exposed Palantir to stark juxtaposition. On one hand, it is benefiting from governments’ willingness to spend heavily on AI tools amid geopolitical tensions and national security concerns. On the other hand, that same dependence on state power exposes the company to reputational and political risks that do not show up neatly in earnings models.

However, Palantir appears well-positioned to capture a significant share of accelerating AI spending in 2026, particularly from defense and public sector clients seeking to turn vast data troves into operational advantage. The challenge for investors is weighing that opportunity against the ethical, regulatory, and political questions that continue to shadow the company’s ascent.

Elon Musk on Monday confirmed that SpaceX has acquired his artificial intelligence startup xAI in a deal that reshapes the global technology landscape and sets a new benchmark for mergers and acquisitions.

The transaction formally unites Musk’s ambitions in artificial intelligence and space infrastructure, combining the world’s most valuable private rocket company with the developer of the Grok chatbot.

The decision marks the clearest step yet toward building a tightly integrated personal conglomerate that spans space, AI, communications, mobility, and neurotechnology, binding his most ambitious bets into a single corporate orbit.

Musk said on Monday that SpaceX had acquired xAI in a deal that values SpaceX at about $1 trillion and xAI at roughly $250 billion, according to people familiar with the transaction. Investors in xAI will receive 0.1433 shares of SpaceX for every xAI share, while some executives are being offered a cash alternative priced at $75.46 per share. The combined entity is expected to price shares at around $527, one of the people said.

The transaction, first reported by Reuters last week, overtakes the $203 billion Vodafone–Mannesmann takeover in 2000 to become the largest merger ever recorded, based on data compiled by LSEG. It also comes as SpaceX is preparing for what could be a landmark initial public offering later this year, with people familiar with the matter saying the listing could value the company at more than $1.5 trillion.

Musk framed the deal in sweeping terms, describing it as the next phase in a shared mission to extend intelligence beyond Earth. Behind the rhetoric, the industrial logic is more concrete. xAI, the developer of the Grok chatbot, is one of the most capital-intensive businesses in the AI race, with costs driven by chips, data centers, and electricity. SpaceX, through its Starlink satellite network, already operates one of the world’s largest private communications infrastructures and generates steady cash flow.

By bringing xAI under the SpaceX umbrella, Musk is effectively internalizing the infrastructure stack needed to compete with rivals such as Alphabet’s Google, Meta, Amazon-backed Anthropic, and OpenAI. Starlink offers global data distribution and a growing enterprise and government customer base, while SpaceX’s launch and satellite operations provide leverage over where and how future compute and data assets are deployed.

Ali Javaheri, a senior emerging spaces analyst at PitchBook, said Starlink was already a cash-flow engine and that the addition of AI created a new revenue layer. He noted that Starlink could also become a distribution surface for AI services and data, particularly if policy shifts allow certain categories of customer data to be used for model training. The longer-term prospect of orbital or space-based data centers, while still speculative, adds to a narrative that positions SpaceX as an integrated infrastructure platform rather than a pure rocket company.

The deal also underscores how Musk has been consolidating his businesses into a more unified structure. Last year, he merged social media platform X into xAI through a share swap, giving the AI startup direct access to real-time data and a built-in distribution channel. Earlier in his career, he used Tesla stock to acquire solar installer SolarCity, folding energy generation into his electric-vehicle ecosystem. Alongside Tesla, Neuralink, and the Boring Company, the SpaceX–xAI tie-up tightens what investors increasingly view as a single, interconnected industrial empire.

However, that strategy is believed to have come with risks. Investors have already expressed unease about aggressive spending and valuation assumptions across Musk’s companies. SpaceX was last valued at about $800 billion in a recent insider share sale, while xAI was valued at roughly $230 billion as recently as November, according to the Wall Street Journal. Combining them at higher figures concentrates both upside and downside into one balance sheet, just as public markets remain sensitive to capital intensity in AI and infrastructure.

Regulatory and governance questions are also likely to follow. SpaceX holds billions of dollars in contracts with NASA, the U.S. Department of Defense, and intelligence agencies, all of which have some authority to review transactions for national security implications. The deal may also draw attention to conflicts of interest, given Musk’s overlapping leadership roles and the potential movement of engineers, data, and proprietary technology between his companies.

Still, the timing suggests strategic calculation. Hyperscalers and AI developers are locked in a global build-out of compute and data-center capacity, with AI-related data-center deals hitting a record $61 billion in 2025. By merging xAI into SpaceX ahead of a potential IPO, Musk is presenting investors with a broader story: not just rockets and satellites, but a vertically integrated platform that links launch, connectivity, data, compute, and AI services.

In effect, the transaction formalizes what has been taking shape for years. Musk is no longer running a collection of loosely related ventures. He is assembling a personal conglomerate, designed so that each business feeds the others, and positioning it for the public markets as a single, expansive bet on the infrastructure of the future.