SoftBank Group’s aggressive wager on artificial intelligence has turned into a windfall. The Japanese tech conglomerate posted a 2.5 trillion yen ($16.6 billion) net profit for the second quarter ended September, more than doubling its results from a year earlier.

The turnaround was driven primarily by valuation gains in OpenAI, the U.S. company behind ChatGPT, marking SoftBank’s strongest performance since 2022, according to Reuters.

The results underline Masayoshi Son’s high-stakes pivot toward AI as the core of SoftBank’s future. Once reeling from the collapse of several Vision Fund-backed startups such as WeWork and Oyo, Son has re-emerged as one of the world’s most bullish investors in AI — a stance he describes as “an all-in conviction on the future of intelligence.”

SoftBank’s Vision Fund, which manages two investment vehicles launched in 2017 and 2019, posted a 3.5 trillion yen gain in the quarter — a reversal from the fund’s previous streak of losses. Nearly two-thirds of that came from its OpenAI holdings, which gained 2.16 trillion yen in value following the U.S. firm’s successive fundraising rounds and soaring secondary market demand.

The scale of SoftBank’s exposure to OpenAI has made it one of the company’s largest outside backers. In March 2025, SoftBank led a funding round of up to $40 billion at a valuation of $300 billion, and by October, it joined a consortium of investors acquiring $6.6 billion worth of employee shares at an implied valuation of $500 billion. By December, its total investment is expected to reach $34.7 billion — a sum larger than the entire GDP of some developing nations.

The AI windfall also pushed SoftBank’s share price to a record high, prompting the company to announce a four-to-one stock split to make its shares more accessible to retail investors. The stock has nearly quadrupled in value in the past six months amid a broader rally in tech-related equities.

Financing the AI gamble

Son’s AI ambitions demand vast capital. To bankroll them, SoftBank has moved swiftly to liquidate legacy assets and raise fresh cash. The firm sold its entire 32.1 million-share stake in Nvidia in October for $5.83 billion, and also offloaded part of its T-Mobile holdings for $9.17 billion.

In parallel, SoftBank has issued bonds in three currencies — yen, dollars, and euros — worth more than 1 trillion yen ($6.4 billion) since April. It also secured two major bridging loans totaling $15 billion: an $8.5 billion loan to finance its OpenAI investment and another $6.5 billion for the planned acquisition of U.S. chip design firm Ampere Computing, which remains pending.

SoftBank’s Chief Financial Officer, Yoshimitsu Goto, defended the company’s aggressive financing strategy at a Tokyo briefing.

“There are various opinions, but SoftBank’s position is that the risk of not investing is far greater than the risk of investing,” he said.

Goto hinted that the company sees OpenAI’s ecosystem — from model training to infrastructure — as a central pillar of global digital transformation.

AI boom — or speculative bubble?

While SoftBank’s results reinforce its reputation as a bellwether for tech cycles, analysts warn that the AI sector may be overheating. Many of the leading companies in the space, including OpenAI, are reporting mounting losses despite sharp valuation increases.

“Son is a savvy investor, so selling the entire [Nvidia] stake must mean he is no longer optimistic about the share price,” said Wong Kok Hoi, CEO of Singapore-based APS Asset Management. “Big tech companies may continue to invest heavily in GPU chips but not at this year’s level for many years.”

The concern is that the massive sums flowing into AI infrastructure — particularly data centers, semiconductor production, and model training — may not yet translate into sustainable profits. OpenAI, for instance, has seen its operating costs surge due to the expense of training and deploying its GPT models, which require vast computing power.

Still, SoftBank maintains that the potential upside of AGI — artificial general intelligence — outweighs the short-term financial risks. Son has framed the company’s strategy as a long-term play that echoes his earlier success with Alibaba, which turned a $20 million stake into more than $50 billion at its peak.

Son’s investing style has always been marked by extremes: bold vision backed by heavy leverage. His early triumph with Alibaba gave him global prominence, but later missteps — particularly with WeWork, which filed for bankruptcy in 2023 — left the Vision Fund battered and investors wary.

SoftBank’s new trajectory represents both a recovery and a reckoning. The group is now positioning itself as a cornerstone investor in the global AI ecosystem, participating in infrastructure ventures, chip design, and data center construction, alongside stakes in leading AI software developers.

The company’s strategy also aligns with Japan’s broader push to reclaim technological leadership in semiconductors and next-generation computing, sectors long dominated by the U.S. and China.

Amid this technological competition, the US motivates EU and Asian companies to relocate their members and ventures to the country, increasing the global visa services activity around the world to ensure global dominance in the AI sector.

However, SoftBank’s latest surge signals a sustainable AI renaissance or a temporary bubble remains uncertain. The current rally has drawn comparisons to the late 1990s dot-com boom, when valuations soared on enthusiasm for a technology whose business models were still unproven.

For now, Son’s gamble has paid off handsomely. But with OpenAI still burning cash, the broader market tightening, and competition intensifying across AI research and infrastructure, SoftBank’s balance between vision and vulnerability may soon be tested again.

SoftBank has sold its entire stake in U.S. chipmaker Nvidia for $5.83 billion, marking one of its boldest portfolio realignments as the Japanese conglomerate doubles down on artificial intelligence, especially through its massive investment in ChatGPT maker OpenAI.

The company said in its second-quarter earnings report that it sold 32.1 million Nvidia shares in October, alongside a $9.17 billion sale of part of its T-Mobile holdings. SoftBank Chief Financial Officer Yoshimitsu Goto described the moves as part of a broader “asset monetization” strategy designed to keep liquidity strong while aggressively funding new AI ventures.

“We want to provide a lot of investment opportunities for investors, while we can still maintain financial strength,” Goto told investors during a presentation. “Through those options and tools, we make sure that we are ready for funding in a very safe manner,” he added, in comments translated by the company.

A Shift From Profit-Taking to Strategic Repositioning

The Nvidia sale raised eyebrows, given the U.S. chipmaker’s central role in powering the AI boom. But some analysts quoted by CNBC say the move is less about losing faith in Nvidia and more about freeing up cash for what SoftBank calls its “all-in” bet on OpenAI and related projects.

“This should not be seen, in our view, as a cautious or negative stance on Nvidia,” said Rolf Bulk, equity research analyst at New Street Research. “Rather, it must be viewed in the context of SoftBank needing at least $30.5 billion of capital for investments in the Oct–Dec quarter, including $22.5 billion for OpenAI and $6.5 billion for Ampere.”

Bulk noted that the total represents “more in a single quarter than it has invested in aggregate over the two prior years combined,” underscoring the scale of SoftBank’s current push into artificial intelligence.

Morningstar’s Dan Baker echoed that sentiment, saying the Nvidia sale “wasn’t any view on Nvidia” but rather a strategic reallocation.

“At the end of the day, they are using the money to invest in other AI-related companies,” he said.

Nvidia and SoftBank: A History of Strategic Entanglement

SoftBank’s relationship with Nvidia has long been cyclical. Its Vision Fund first built a $4 billion stake in the chipmaker in 2017, before selling all its shares in January 2019. The latest divestment, therefore, marks the second time SoftBank has cashed out of Nvidia after a period of strong valuation growth.

Despite the exit, the two companies remain intertwined through overlapping ventures. SoftBank is a partner in several AI and data infrastructure projects that depend on Nvidia’s GPU technology, including the ambitious $500 billion “Stargate” project for U.S. data centers.

Nvidia shares fell 0.95% in premarket trading on Tuesday, but analysts say the sale is unlikely to dent confidence in the company’s long-term trajectory, given record demand for AI chips and data center components.

SoftBank’s Vision Fund posted a $19 billion gain in the quarter — its best performance in years — largely due to soaring valuations of AI firms in its portfolio. Goto said the success was partly driven by the company’s early investment in OpenAI last year, which has since reached a $500 billion valuation by fair value estimates.

“The reason we were able to have this result is because of September last year — that was the first time we invested in OpenAI,” Goto said. He described OpenAI’s valuation as one of the world’s largest, ranking it alongside major tech giants by market worth.

The Vision Fund’s renewed momentum marks a turnaround for SoftBank, whose earlier bets on startups like WeWork and Katerra led to significant losses. Now, the company is positioning itself as a central player in the global AI race, investing heavily across the value chain — from semiconductor firms and data center operators to robotics and large language models.

Still, not everyone is convinced the AI rally will sustain. SoftBank’s stock has slumped over the past week, reflecting investor unease about what some analysts call an “AI bubble.” The fear is that valuations of AI-focused firms are running far ahead of actual revenue growth, echoing patterns from past tech booms.

Goto addressed those concerns, saying SoftBank remains focused on “providing as many investment opportunities as possible” to shareholders while maintaining balance sheet flexibility. The company’s planned four-for-one stock split, he said, is part of that strategy to “broaden access” and attract more investors.

“We know our share price has been moving dynamically,” he said, acknowledging the market volatility. “But this is part of a larger effort to strengthen our financial base and support long-term growth.”

However, with SoftBank now firmly pivoting toward AI, the company appears ready to play a defining role in shaping the industry’s future infrastructure. Analysts expect further announcements of partnerships and acquisitions in the coming months, particularly as it channels billions into AI chip design, robotics, and large-scale data centers.

While its exit from Nvidia may have startled markets, the move points to a company repositioning itself for a new era — one where SoftBank is betting that artificial intelligence, not semiconductors alone, will be the true engine of the next technological revolution.

Good People, let us dissect the imminent future. Uber, the global ride-hailing leviathan, is seeing the writing on the wall. They understand the profound disruptive enzyme that self-driving technology represents, a force ready to utterly disintermediate their core service model. Why? Because the very function they aggregate, the driver, is destined for obsolescence.

I project that the arrival of full autonomy is not a distant aspiration but a fast-approaching business equilibrium. By 2028, I expect to see advanced, fully driverless car capabilities scaling rapidly across the Chinese tech ecosystem, and by 2030, the United States market will embrace this new paradigm at full scale. Simply put, autonomous capability will no longer be an add-on; it will be a non-negotiable, embedded core component of every vehicle manufactured. When this happens, the aggregated network effect, Uber’s historical competitive moat built on matching human drivers to human riders, will collapse at scale.

To survive this seismic shift, Uber is not standing still; it is charting a bold, new trajectory. CEO Dara Khosrowshahi has declared a strategic transformation: Uber is evolving far beyond mere rides and deliveries into a global “platform for work.” This is their defense mechanism. They are now focused on connecting their millions of users with digital tasks, including jobs training and the burgeoning demands of the Artificial Intelligence ecosystem, a strategic pivot designed to diversify income and remain viable in the age of automation.

Uber Technologies is broadening its scope far beyond rides and deliveries, with CEO Dara Khosrowshahi declaring that the company is transforming into a global “platform for work.”

The company’s newest ambition is to connect its millions of app users with digital tasks — including jobs training, and artificial intelligence — as part of a strategic shift toward diversifying income opportunities in an era of automation.

This is the crux, Good People: Numerous established business models are marked for extinction as AI permeates the heart of global economies. From the learning-content models of platforms like Chegg to the aggregation models of Uber, the status quo is under existential threat.

Therefore, the critical question for every enterprise leader is this: What is your company’s AI Resilience Score? Are you proactively evaluating the viability of your business model against the relentless march of intelligent automation?

The most profound utility of AI is not its deployment as a mere operational tool for efficiency gains. No! AI’s true, transformative power lies in its capacity to re-engineer and fundamentally transform the business model itself. If AI does not force you to reimagine your value proposition and your economic structure, you are not deploying it right.

Uber is strategically planning for its post-driver future. You must do the same for your mission. Here at Tekedia Institute, we understand this principle intimately. We have noted that the sheer Supply of Knowledge is rapidly being disintermediated by Large Language Models and AI platforms. Anyone can now curate any basic lecture note. Consequently, our value has shifted. We are doing more live, interactive sessions because the value today resides not in the content itself, but in the contextualization, the filtering, and the interpretation of that knowledge, removing the friction that too much abundance creates.

I ask you again, leader: What is your business’ AI Resilience Score as disruption picks up across markets and geographies? Failure to plan is a plan to fail.

Nvidia Corp. Chief Executive Officer Jensen Huang has urged Taiwan Semiconductor Manufacturing Co. (TSMC) to increase chip supply, as global demand for artificial intelligence hardware shows no signs of slowing.

The move, first reported by Bloomberg, comes at a time when the semiconductor supply chain remains tight, and competitors are struggling to secure enough production capacity from the world’s largest contract chipmaker.

Speaking to reporters on Saturday in Hsinchu, Taiwan, where he attended TSMC’s annual sports day, Huang described Nvidia’s business as “very strong,” adding that it continues to expand rapidly.

“The business is very strong, and it’s growing month by month, stronger and stronger,” he said, underscoring that Nvidia’s AI chip orders are surging as companies across the globe race to build the computational infrastructure needed for next-generation AI models.

According to Huang, Nvidia’s three major AI memory suppliers — SK Hynix, Samsung Electronics, and Micron Technology — have all scaled up their production “tremendously” to meet rising orders. These suppliers provide the high-bandwidth memory (HBM) chips critical for Nvidia’s flagship H100 and H200 processors, which power the world’s largest AI systems, including those used by OpenAI, Microsoft, and Amazon Web Services.

Huang’s appeal for more capacity came during a private dinner with TSMC CEO C.C. Wei on Friday, followed by public comments the next day. Wei confirmed that Huang requested more wafers and added that TSMC expects to “continue breaking sales records every year” as demand for high-performance chips intensifies.

The timing of Huang’s visit notably came just as Wall Street turned cautious on the AI rally. Nvidia’s shares slid earlier in the week after reports surfaced that OpenAI’s infrastructure spending and funding commitments might be under review, and hedge fund Scion Asset Management, led by investor Michael Burry, disclosed bearish bets against Nvidia. The broader technology sector, including Apple and Microsoft, also saw declines amid investor fears of overheating valuations.

Still, Nvidia remains the world’s most valuable company, ahead of both Apple and Microsoft, with a market capitalization of about $5 trillion. Its chips dominate roughly 80% of the global market for AI accelerators, used in training and deploying generative AI systems such as ChatGPT, Gemini, and Claude.

Huang credited TSMC for its central role in Nvidia’s rise, calling the company indispensable to the global AI ecosystem. “No TSMC, no Nvidia,” he said, highlighting the deep dependence of U.S. chip design firms on Taiwanese manufacturing. Nvidia relies heavily on TSMC’s 4-nanometer and 3-nanometer processes, which are among the most advanced in the world, to produce its cutting-edge GPUs.

His optimism about AI’s trajectory mirrors that of Qualcomm CEO Cristiano Amon, who told Bloomberg TV earlier in the week that “the world is underestimating how big AI will become.” Amon said demand for AI processing chips is “the strongest technology wave since the smartphone era,” with potential to reshape computing, telecommunications, and industrial automation.

TSMC, meanwhile, faces an ongoing challenge in balancing surging demand from multiple tech giants, including Apple, AMD, Broadcom, and Intel, all of which compete for its limited advanced fabrication capacity. Wei said in October that the company’s capacity remains very tight, and that it is “working hard to narrow the gap between demand and supply.”

The global AI race has created one of the most intense supply crunches in semiconductor history. Data centers worldwide are competing for Nvidia’s GPUs, and companies are now prepaying billions of dollars to secure future deliveries. In the United States and Europe, governments are also racing to build domestic chip plants to reduce dependence on Asia, though Taiwan continues to hold the world’s most critical semiconductor production capabilities.

Analysts say Huang’s visit to Taiwan underscores both Nvidia’s influence and its vulnerability. Without access to sufficient wafer supply from TSMC, Nvidia risks losing ground to rivals like AMD, Intel, and Qualcomm, which are all developing AI chips to capture a slice of the expanding market.

For now, Nvidia’s demand remains unmatched. The company’s planned Blackwell architecture, expected to begin shipping soon, has already drawn huge interest from hyperscalers and sovereign AI projects. TSMC is set to be the exclusive manufacturer of the new chips.

As the AI boom accelerates, both companies are likely to see their fates remain intertwined. Maintaining access to TSMC’s cutting-edge capacity is crucial to sustaining Nvidia’s dominance. On the other hand, Nvidia’s insatiable demand is not just a business win — it’s a signal that the world’s next computing revolution still runs through Taiwan’s foundries.