On Friday, Delaware’s highest court overturned a 2024 ruling by the Court of Chancery that had voided Musk’s $56 billion, milestone-based pay plan, calling the lower court’s decision to rescind the package an excessively harsh remedy.

The justices said Tesla was never given a proper opportunity to demonstrate what a fair level of compensation for Musk might be, and awarded only $1 in nominal damages to the shareholder who brought the suit.

The decision effectively ends the long-running Tornetta v. Musk case and restores what remains the largest CEO pay package in corporate history. But its implications extend well beyond settling an old dispute.

At the heart of the ruling is a clear signal from the Delaware Supreme Court that even where flaws exist in a board’s process, courts should be cautious about unwinding compensation agreements that have already been approved by shareholders and executed over time. That principle is now being closely watched in light of Tesla’s latest and far more controversial move: a new compensation framework presented to Musk in 2025 that could, under optimistic scenarios, be worth around $1 trillion.

Background: the 2018 deal and the legal fight

Musk’s original 2018 package was revolutionary. He agreed to take no salary or cash bonus, instead tying his pay entirely to Tesla’s performance across 12 tranches of stock options. Each tranche was linked to aggressive targets for market capitalization, revenue, and profitability. If Tesla failed, Musk earned nothing. If it succeeded, shareholders would benefit alongside him.

Tesla’s explosive growth meant the plan was fully vested, turning the package into a symbol of Silicon Valley excess — and a magnet for legal scrutiny.

Shareholder Richard Tornetta sued, accusing Musk and Tesla’s board of breaching fiduciary duties by approving the deal through what he described as a conflicted and opaque process. In January 2024, Chancellor Kathaleen McCormick sided with Tornetta, ruling that Musk “controlled Tesla,” that the board was insufficiently independent, and that shareholders were not given all material information before voting. She ordered the pay package rescinded outright.

That ruling triggered a fierce backlash. Musk publicly attacked the decision, moved Tesla’s incorporation out of Delaware, and encouraged other companies to follow. Tesla also attempted to ratify the 2018 package through a second shareholder vote in 2024, while appealing the ruling.

The Supreme Court’s reversal

In overturning the Chancery Court, the Delaware Supreme Court did not fully vindicate Tesla’s process, but it rejected rescission as an appropriate solution. The justices said the lower court failed to consider alternative remedies and denied Tesla the chance to argue what compensation would be reasonable, given Musk’s role and Tesla’s performance.

By restoring the pay package, the court reinforced the idea that shareholder-approved compensation, especially when tied to extraordinary corporate outcomes, should not be easily undone after the fact.

A precedent for the $1 trillion question

That reasoning is now central to the debate around Tesla’s newest compensation proposal for Musk.

Earlier this year, Tesla board unveiled a new long-term incentive framework that would again rely heavily on equity awards tied to ambitious valuation and operational milestones. While the exact payout would depend on Tesla’s future performance, analysts estimate that if the company meets its most aggressive targets, the package could ultimately be worth close to $1 trillion — a figure that has stunned even seasoned observers of executive pay.

Critics have already warned that the proposal risks repeating the governance concerns raised in the 2018 case, particularly around board independence and Musk’s outsized influence. Supporters counter that Musk remains inseparable from Tesla’s value proposition and that shareholders should be free to reward him accordingly if he delivers exceptional results.

The Delaware Supreme Court’s ruling strengthens Tesla’s hand. By making clear that rescinding a shareholder-approved pay deal is an extraordinary step, the court has raised the legal threshold for successfully challenging future Musk compensation packages. While any new plan could still face lawsuits, legal experts say plaintiffs may find it harder to persuade courts to nullify such agreements outright, especially if Tesla improves its disclosure and approval processes.

The decision also lands amid a broader rethinking of Delaware’s role in corporate America. Tesla’s legal battle coincided with legislative changes to Delaware corporate law earlier this year, changes that were supported by firms representing major companies and aimed at reducing litigation risk for boards and executives.

In a bold pivot from previous U.S. export restrictions, President Donald Trump’s administration has initiated an inter-agency review that could greenlight the first shipments of Nvidia’s high-powered H200 AI chips to China, according to five sources familiar with the matter who spoke to Reuters.

The move fulfills Trump’s recent pledge to permit such sales while imposing a 25% fee on the U.S. government, but it has ignited fierce backlash from critics who warn it could bolster Beijing’s military capabilities and undermine America’s AI edge.

The review, launched by the U.S. Commerce Department—which oversees export controls—has forwarded license applications for the chip sales to the State, Energy, and Defense Departments for input, the sources said, speaking on condition of anonymity due to the confidential nature of the process. Under federal export regulations, these agencies have 30 days to provide their assessments, after which the final decision rests with Trump himself.

One administration official stressed that the evaluation would be rigorous, describing it as “not some perfunctory box we are checking.” The Commerce Department and Nvidia declined to comment immediately, while a White House spokesperson avoided specifics on the review but affirmed the administration’s stance: “The Trump administration is committed to ensuring the dominance of the American tech stack – without compromising on national security.”

This development marks a stark reversal from the Biden era’s stringent bans on advanced AI chip exports to China, implemented over fears that such technology could fuel Beijing’s military ambitions or be rerouted through third countries. Trump’s first term had similarly cracked down on Chinese access to U.S. tech, spotlighting allegations of intellectual property theft and dual-use applications for military purposes—claims Beijing has consistently rejected.

Proponents within the Trump team, including White House AI czar David Sacks, contend that allowing sales of the H200 chips could strategically deter Chinese firms like Huawei from accelerating their own AI advancements. By satisfying demand with American products, the argument goes, it reduces incentives for Beijing to invest heavily in domestic alternatives, helping U.S. companies like Nvidia and AMD maintain their lead.

However, the proposal has drawn sharp condemnation from China hawks on both sides of the aisle. Chris McGuire, a former National Security Council official under Biden and now a senior fellow at the Council on Foreign Relations, called the potential exports “a significant strategic mistake.” He labeled the H200 chips as “the one thing holding China back in AI,” questioning how agencies could deem the sales aligned with U.S. national security interests.

“I cannot possibly fathom how the departments of Commerce, State, Energy, and Defense could certify that exporting these chips to China is in the U.S. national security interest,” McGuire said.

The H200, Nvidia’s second-most advanced AI chip and predecessor to its flagship Blackwell series, excels in many industry tasks but has never been cleared for sale in China. Reuters previously reported that Nvidia is eyeing a production ramp-up for the H200 after early Chinese orders exceeded capacity. Trump had initially floated sales of a downgraded Blackwell variant but retreated, opting instead for the H200 amid ongoing scrutiny.

Uncertainties linger over Beijing’s response, including whether Chinese companies would be permitted to buy the chips and how swiftly approvals might materialize. The inter-agency process, unreported until now, underscores the high-stakes balancing act between economic gains—bolstered by the proposed 25% tariff-like fee—and geopolitical risks in the escalating U.S.-China tech rivalry.

As the 30-day clock ticks, the outcome is expected to impact global AI supply chains, testing Trump’s vow to prioritize American innovation while protecting the U.S.’ national security interests.

Sam Altman is openly conflicted about the idea of taking OpenAI public, a hesitation that underscores the tension between the company’s explosive growth and the structural demands that come with operating at unprecedented scale.

Speaking on the Big Technology Podcast published on Thursday, the OpenAI chief executive made clear that while he sees advantages in an eventual stock market listing, the costs — particularly for leadership and long-term decision-making — loom large.

“Am I excited for OpenAI to be a public company? In some ways, I am, and in some ways I think it’d be really annoying,” Altman said.

On the prospect of personally running a listed company, he asked: “Am I excited to be a public company CEO?”

Those remarks come at a time when OpenAI’s trajectory increasingly resembles that of a mega-cap public company, even as it remains privately held. Founded in 2015 as a research-focused lab, OpenAI has been transformed since the launch of ChatGPT in late 2022 into one of the most influential forces in global technology. ChatGPT now has around 800 million weekly users, a scale that rivals or exceeds many established consumer platforms, and OpenAI has embedded its models across enterprise software, developer tools, and consumer applications.

The company has also entered into sweeping commercial and infrastructure arrangements with some of the biggest names in technology. Altman said OpenAI has inked deals worth about $1 trillion with partners including Oracle, Nvidia, and AMD, underscoring both the scale of its ambitions and the capital intensity of building and running frontier AI models. These agreements reflect OpenAI’s reliance on vast amounts of compute, energy, and specialized hardware, costs that continue to rise as models grow more capable and more widely deployed.

That capital intensity sits at the heart of the IPO debate. Altman acknowledged that OpenAI “needs lots of capital” and will eventually run into shareholder and regulatory limits that make staying private difficult. Private markets have so far been willing to supply funding at eye-watering valuations. In October, OpenAI was valued at about $500 billion in a secondary share sale, briefly overtaking Elon Musk’s SpaceX as the most valuable private company in the world before SpaceX reclaimed the top spot. This week, The Information reported that OpenAI is seeking to raise billions more at a valuation of roughly $750 billion.

Such figures point to extraordinary investor confidence, but they also raise questions about sustainability and expectations. OpenAI’s revenues, while growing rapidly, still lag far behind its valuation. The company is reported to be generating annual run-rate revenue of around $20 billion, driven by ChatGPT subscriptions, enterprise licensing, and API access. Against that, its spending runs into the tens of billions of dollars a year, much of it tied to inferencing costs, data center expansion, and long-term compute commitments that cannot be fully offset by cloud credits or partnerships. An IPO would expose that imbalance to the full glare of public markets, where tolerance for prolonged losses can shift quickly.

There are also governance considerations. OpenAI’s unusual structure — originally a nonprofit with a capped-profit arm — was designed to prioritize safety and long-term research over short-term financial returns. While that structure has already evolved significantly as the company has commercialized, becoming a public company would further tilt incentives toward quarterly performance and shareholder returns. For a firm racing rivals like Google, Anthropic, Meta, and xAI to develop ever more powerful models, that pressure could complicate decisions about research timelines, safety investments, and deployment speed.

Still, Altman does not dismiss the idea of going public outright. He said he likes the idea that “public markets get to participate in value creation,” framing an IPO as a way to broaden ownership and give ordinary investors access to the upside of AI development, rather than confining gains to venture capital and sovereign wealth funds.

Signs of preparation are already emerging. Reuters reported in October that OpenAI is considering filing with securities regulators as soon as the second half of 2026. That timeline would allow the company to further scale revenues, refine its corporate structure, and, potentially, narrow the gap between spending and income before subjecting itself to public scrutiny. Yet when asked directly on the podcast whether OpenAI would list next year, Altman replied, “I don’t know,” adding that if it does happen, “we will be very late to go public.”

However, OpenAI appears intent on buying time — raising ever-larger private rounds, deepening strategic partnerships, and pushing its technology into more products and markets for now.