DeepSeek has rolled out a preview of its V4 artificial intelligence model built to run on Huawei chips, a move that highlights China’s growing ability to develop advanced AI systems outside the U.S. semiconductor ecosystem, and one that is likely to keep competitors, particularly in the United States, on edge.

The new model marks a shift from DeepSeek’s earlier dependence on processors from Nvidia, though the company did not specify the exact chips used for training. Instead, it emphasized close integration with Huawei’s Ascend AI systems, which are central to Beijing’s push to reduce reliance on foreign chip technology.

Huawei said the collaboration ensures broad compatibility. “The entire Ascend supernode product line now supports the DeepSeek V4 series models,” it said, signaling that the model is designed to run across a wide range of domestic high-performance computing infrastructure.

The V4 is being released in preview form, allowing DeepSeek to gather user feedback before final deployment. It also includes a lower-cost “flash” version, reinforcing the company’s approach of undercutting rivals on pricing while maintaining competitive performance, a combination that helped drive its rapid rise last year.

According to DeepSeek, the pro version of V4 outperforms other open-source models on world-knowledge benchmarks, trailing only Gemini-Pro-3.1, a closed-source system from Google. That positioning, if sustained, strengthens DeepSeek’s role as a leading player in the open-source segment, where accessibility and cost efficiency are becoming decisive factors.

The release comes at a delicate moment in U.S.-China relations. A day earlier, Washington accused China of acquiring intellectual property from American AI labs “on an industrial scale,” intensifying scrutiny of companies like DeepSeek ahead of a planned summit between the two countries.

DeepSeek has been a focal point in that debate. U.S. officials have alleged it circumvented export controls to obtain advanced Nvidia chips, while OpenAI and Anthropic have said it may have improperly “distilled” their proprietary models. DeepSeek has acknowledged using Nvidia hardware but has not confirmed whether those chips were restricted, and it has said its earlier V3 model relied on naturally collected web data rather than synthetic outputs from competing systems.

China has rejected the accusations. The Chinese Embassy in Washington called them “baseless,” adding that the country places importance on protecting intellectual property.

Beyond the political friction, the V4 launch illustrates a deeper shift in how AI systems are being built. With U.S. export controls limiting access to cutting-edge chips since 2022, Chinese firms have been forced to redesign their technology stacks, pairing domestic hardware with increasingly optimized software. The collaboration between DeepSeek and Huawei reflects that adjustment, narrowing performance gaps through tighter integration rather than relying solely on raw computing power.

This approach is beginning to show results. While Huawei’s chips are still often seen as less advanced than Nvidia’s top-tier offerings, improvements in software efficiency and model design are helping offset hardware limitations. That dynamic is central to China’s effort to build a self-sustaining AI ecosystem.

DeepSeek’s rise has already unsettled the competitive industry. Its earlier models demonstrated that high-performing AI systems could be developed and deployed at significantly lower cost, challenging assumptions about the scale of investment required to compete. Each successive release has forced rivals to reassess pricing, efficiency, and deployment strategies.

The V4 preview is likely to have a similar effect. Like previous launches, it raises the bar for performance within the open-source segment while reinforcing cost pressure across the industry. For U.S.-based developers, it adds urgency to an already intense race, particularly as competition extends beyond model capability to include infrastructure, cost, and global accessibility.

The immediate market reaction in China reflected that pressure. Shares of Zhipu AI fell 9%, while MiniMax dropped 7%, suggesting investors see DeepSeek’s latest model as a competitive threat to domestic peers as well.

Interest in DeepSeek itself is growing. The company, backed by High-Flyer Capital Management, is reportedly seeking funding at a valuation above $20 billion, with Alibaba and Tencent said to be exploring potential stakes.

The broader implication is that the AI race is becoming more distributed. Rather than a single dominant ecosystem, parallel development tracks are emerging, shaped by geopolitical constraints and differing approaches to cost, openness, and infrastructure. DeepSeek’s latest release does not settle the contest, but it reinforces a pattern. Each time the company introduces a new model, it shifts expectations around what is possible with fewer resources and alternative hardware.

That pattern is likely to keep competitors, especially in the United States, recalibrating their own strategies, as the gap between the two ecosystems narrows in both capability and ambition.

Tesla’s long-promised robotaxi revolution is encountering the messy realities of real-world deployment, with CEO Elon Musk delivering an unusually measured update on the company’s autonomous ambitions during Wednesday’s first-quarter earnings call.

For months, Musk had sketched out a breakneck expansion: robotaxis serving half the U.S. population by the end of 2025, followed by “hyper-exponential” growth. On Wednesday, he struck a markedly different chord. He now expects driverless operations in “a dozen or so states” by year-end and stressed a deliberate, safety-first posture designed to avoid any headline-grabbing accidents or fatalities that could halt progress cold.

Details on the newly announced Dallas and Houston rollout remained sparse, leaving investors to read between the lines of a tone that felt more boardroom prudent than launch-pad exuberant.

The shift marks another recalibration in a decade-long pattern of ambitious targets that have repeatedly slipped. Musk himself poked fun at his reputation in January 2025, calling himself “the boy who cried wolf” on self-driving technology, before insisting this time the wolf was real.

Last July, fresh off a limited Austin pilot, he forecast far broader coverage by now. On Wednesday, the message was clear: scale will come, but only after rigorous validation.

Tesla is pinning its autonomous future on the Cybercab, the sleek two-seater unveiled last year with no steering wheel or pedals. Production has started at Giga Texas, Musk confirmed, yet he warned the initial ramp will be “very slow,” with “exponential” growth only toward the end of this year and into 2027.

Long term, he sees the Cybercab dominating Tesla output. Paid robotaxi miles nearly doubled in the first quarter, but the fleet remains modest and heavily supervised in limited geographies.

The limiting factor, Musk said bluntly, is not manufacturing muscle but software safety.

“The limiting factor for expansion is really rigorous validation, making sure things are completely safe,” he said.

A forthcoming software update is expected to clear the path for wider deployment, but until then, the company is holding back. That caution pushed back earlier projections: robotaxis were once slated to become “material” to the bottom line by mid-2026; now Musk sees them as “not super material this year” but potentially significant next year.

Wall Street took note of the tempered language. William Blair called the call “low energy” and observed that Musk sounded “reserved and cautious” on his signature topic. Morgan Stanley, among the more bullish voices on autonomy, conceded the rollout is “progressing slower than investor expectations,” trimming near-term stock upside. Barclays highlighted that Tesla still operates only a “nominal number” of fully driverless vehicles.

Morningstar’s Seth Goldstein gave the caution a vote of confidence, arguing the stakes are simply too high for shortcuts: one serious incident could invite regulatory crackdowns and erode years of goodwill.

However, not everyone is ringing alarm bells. CFRA’s Garrett Nelson pointed out that veteran Tesla watchers have grown accustomed to “Elon time.” As long as the company proves the business model works in a few well-chosen markets, he said, investors will likely keep extending credit.

The broader context is unforgiving, as much of Tesla’s roughly $1.5 trillion market capitalization still rests on the belief that a vast robotaxi fleet and millions of autonomous software subscriptions will become the next growth engine as traditional EV sales mature.

Competitors such as Alphabet’s Waymo have spent years wrestling with the same operational thicket—mapping edge cases, securing permits city by city, and proving unsupervised driving can be both safe and profitable. Tesla is betting its camera-and-neural-net approach can leapfrog that experience curve, but the ground game is proving stubborn.

Shares fell more than 3 percent in afternoon trading Thursday, a modest rebuke that suggests the market is still willing to wait—provided the next software update and early Cybercab volume deliver tangible proof of progress. This time, Musk has traded the usual fireworks for a quieter message. He indicates that the technology is coming, but only when it is truly ready. In the high-stakes world of autonomous vehicles, that may be the most reassuring thing investors have heard in a while.

Apple is entering a consequential leadership transition that could redefine its trajectory, as longtime hardware chief John Ternus prepares to take over as chief executive in September from Tim Cook, ending a tenure that reshaped the company into a services and operations powerhouse.

The shift is already prompting debate among investors about what Apple’s next chapter should look like — and how much it should diverge from the Cook era. Longtime shareholder Fran Tarkenton framed the moment as one that demands independence rather than continuity, pointing to advice once given by Steve Jobs to Cook ahead of his own succession in 2011.

“He said to Tim Cook, When you’ve got to make a decision, don’t say to yourself, what would Steve Jobs do? You do what you think is the best thing for the company,” Tarkenton said.

That guidance, which Cook later recalled during a memorial tribute to Jobs, now serves as a benchmark for Ternus as he steps into a role defined by both high expectations and shifting industry dynamics.

Tarkenton was explicit about the need for differentiation. “He cannot be Tim Cook,” he said.

The comment reflects a broader investor sentiment: while Cook’s leadership delivered scale, profitability, and resilience, the challenges facing Apple today are materially different. Growth in core hardware segments has moderated, competition has intensified, and the center of gravity in the tech industry is shifting toward artificial intelligence.

Ternus’s appointment signals a potential recalibration. With more than two decades at Apple and a background rooted in hardware engineering, he represents a return to the company’s product-driven DNA — a contrast to Cook’s operational focus, which emphasized supply chain efficiency, global expansion, and the rapid scaling of services.

Tarkenton endorsed the choice, describing Ternus as “the right guy at the right time,” but stressed that execution will depend on how effectively he leverages that experience.

“He cannot be Tim Cook,” he repeated, adding that Ternus must rely on his 25 years inside Apple to navigate the transition.

Cook’s tenure leaves a complex legacy. Under his leadership, Apple expanded its services business into a multi-billion-dollar engine spanning subscriptions, payments, and digital content, in some periods outpacing hardware in revenue growth. That diversification strengthened margins and reduced reliance on flagship products like the iPhone.

However, it shifted investor expectations. Apple is now being judged not only on financial performance but on its ability to deliver the next major technological leap, particularly in AI.

The timing of the transition amplifies that pressure. Rivals have moved aggressively to integrate generative AI into both consumer and enterprise ecosystems, raising questions about Apple’s pace and positioning. For Ternus, the challenge will be to integrate AI into Apple’s tightly controlled hardware-software ecosystem in a way that aligns with its emphasis on privacy, performance, and user experience.

His hardware background could prove central to that effort. Apple’s historical advantage has come from vertical integration, designing chips, devices, and software as a unified system. Extending that model into AI could allow the company to differentiate without competing directly on scale with cloud-based AI providers.

Still, the tension lies largely in Cook’s model. Cook’s model prioritized predictability and operational discipline. Ternus inherits that foundation but faces a market that increasingly rewards visible innovation and category creation.

Tarkenton, who began investing in Apple around 2015 and has maintained a long-term position, said his confidence in the company remains unchanged.

“I’ve never sold any of the stock, and I reinvest the dividends every quarter,” he said.

His conviction is partly rooted in Apple’s internal culture and leadership pipeline. “The teams that win have the best coaches, and the best coaches make sure they get the best players,” Tarkenton said. “Apple does that naturally.”

The governance structure also provides continuity. Cook is expected to remain as executive chairman, offering oversight while allowing Ternus to assume operational control. The arrangement is designed to preserve institutional knowledge while enabling a shift in leadership style.

Internally, Ternus’s rise has been anticipated. Tarkenton said his contacts within Apple had been discussing him as a potential successor for years, suggesting the transition is the result of deliberate succession planning rather than a reactive decision.

As recently as last month, Cook downplayed speculation about stepping aside, saying he could not imagine his life without Apple. His continued presence as chairman may help reassure investors during the transition, even as strategic direction evolves.

The broader question is how far that evolution will go. A return to product-centric leadership could signal renewed focus on breakthrough hardware, deeper AI integration, and potentially new categories. At the same time, the services ecosystem built under Cook remains a critical pillar of profitability.

The challenge for Apple is not choosing between the two models but reconciling them, combining operational strength with renewed innovation.