Physical Intelligence, the San Francisco-based robotics startup founded two years ago, is reportedly in talks to secure around $1 billion in fresh funding at a valuation surpassing $11 billion, as per Bloomberg. This agreement would essentially double the firm’s valuation of $5.6 billion within a mere four months.

Founders Fund is expected to be involved, with Lightspeed Venture Partners also in discussions to invest alongside previous supporters Thrive Capital and Lux Capital, according to Bloomberg. The negotiations are still in the preliminary phases, and specifics may alter, the outlet noted.

In January, TechCrunch visited the headquarters of Physical Intelligence, where co-founder Sergey Levine articulated the company’s vision succinctly: “Imagine it as ChatGPT, just for robots.” At that point, the firm had raised slightly over $1 billion and employed approximately 80 individuals focused on developing general-purpose AI models capable of enabling robots to execute an array of tasks, including folding laundry and peeling vegetables.

Co-founder Lachy Groom mentioned to TechCrunch that the company does not have a set timeline for commercialization, a stance that its investors appear to accept. “There’s no cap on how much funding we can realistically deploy,” Groom stated. “There’s always additional computing power you can apply to the challenge.”

SoftBank has secured a new loan amounting to $40 billion to finance its $30 billion pledge to invest in OpenAI, part of the AI developer’s unprecedented $110 billion fundraising last month, the Japanese firm announced on Friday.

Notably, the loan is unsecured and has a term of 12 months, implying it must be paid back or refinanced by the next year. This may suggest that lenders expect OpenAI’s much-anticipated public offering will materialize later this year, as some sources, including CNBC, have indicated. JPMorgan Chase, Goldman Sachs, and four Japanese banks are facilitating the loan.

If OpenAI’s IPO turns out to be one of the largest in history this year, it would likely provide SoftBank with the necessary liquidity to address the debt within such a brief timeframe. With its latest $30 billion investment in OpenAI, SoftBank’s total investment in the creator of ChatGPT exceeds $60 billion.

SK hynix, a leading South Korean memory chip manufacturer already traded on the KOSPI, is preparing for a potential listing in the U.S. that could reportedly bring in approximately $10 billion to $14 billion.

This week, the company revealed that it has confidentially submitted a Form F-1 for the listing, with a focus on the latter half of 2026.

However, the key issue isn’t merely the amount it can generate: it’s whether a U.S. listing could enhance its trading value as one of the most vital players in the AI chip supply chain.

Despite its significant contribution to high-bandwidth memory (HBM), a critical element supporting AI technologies from firms like Nvidia, the stock has previously been trading at a lower valuation compared to global competitors, based on insights from a Seoul-based semiconductor analyst. While its market capitalization stands at around $440 billion, the valuation multiples are still below those of U.S.-listed semiconductor companies, raising concerns over whether geographic location, rather than actual business fundamentals, is influencing this disparity.

This action is widely regarded as an endeavor to elevate its valuation to align more closely with global counterparts such as Micron.

“SK hynix’s U.S. listing could aid in rectifying a persistent valuation gap with global rivals. Even with production capacity comparable to – or in some sectors superior to – U.S.-based chipmakers, the Korean firm has historically been undervalued, partly due to its primary listing being in Korea,” the analyst commented to TechCrunch.

The analyst also pointed out structural elements influencing the deal. “SK Square, the largest shareholder of SK hynix, holding 20.07% as of December 2025, is mandated to retain at least a 20% stake according to Korea’s holding company laws.”

According to current share prices, issuing around 2% in new shares could generate $10 billion to $14 billion while allowing SK Square to uphold its ownership requirement, as stated by the analyst. (Under Korea’s Fair Trade Act, holding companies must maintain minimum ownership levels in subsidiaries, at least 20% for listed companies, to keep control.)

There exists a precedent. Taiwan Semiconductor Manufacturing Company (TSMC), for instance, has experienced instances where its U.S.-listed shares traded at a premium over its domestic shares, especially during times of elevated AI demand, indicating that cross-listing can affect investor pricing of the same underlying business.

This initiative is already causing ripples throughout the broader Korean chip landscape. After SK hynix’s filing, some investors are advocating for Samsung Electronics to contemplate a similar U.S. listing. Artisan Partners, a significant shareholder, mentioned on Friday that a U.S. listing (technically termed an American depositary receipt, or ADR) could also enhance Samsung’s valuation and provide U.S. retail investors the opportunity to purchase its stock, according to a Bloomberg report.

A capital push to meet AI-driven demand

SK hynix’s planned ADR listing is also largely interpreted as a strategy to secure funding ahead of boosted capital expenditures to satisfy the growing demand for memory from AI semiconductors.

During its annual general meeting on March 25, CEO Noh-Jung Kwak stated that financial capacity will be crucial for maintaining growth in the AI age, adding that the company aims for around $75 billion (over 100 trillion KRW) in net cash to support long-term investments.

Rising costs for memory and constrained supply have been key factors hindering AI development, as well as affecting other sectors, such as gaming. This phenomenon has been referred to as “RAMmageddon” and is anticipated to persist until at least 2027, as reported by Nature.

Only time will reveal if this bleak prediction materializes. Tech giants are pursuing solutions to RAMmageddon beyond just increased manufacturing. For example, Google unveiled a new technology this week called TurboQuant, an ultra-efficient AI memory compression algorithm, which significantly enhances AI’s efficiency in memory usage.

Nevertheless, the indications suggest that additional memory production will be necessary. SK hynix is preparing for an influx of capital-intensive initiatives, intending to invest roughly $400 billion by 2050 to establish a semiconductor cluster in Yongin, South Korea. The company is also in the process of developing new facilities in South Korea and Indiana, with intended investments of about $25 billion and $3.3 billion, respectively, highlighting the extensive capital needed.

The chipmaker announced this week that it plans to acquire advanced extreme ultraviolet (EUV) lithography scanners from ASML by 2027 in a deal valued at $7.9 billion, aimed at boosting HBM production for AI.

All of this would be facilitated by a remarkable U.S. IPO. This could inspire other Korean chip manufacturers to follow suit.