The legal battle over whether sports prediction contracts are financial derivatives or a form of gambling has deepened after a federal appeals court ruled that Ohio and Tennessee can apply their state gambling laws to sports-related contracts offered by Kalshi.

The 6th U.S. Circuit Court of Appeals ruled unanimously on Friday that Kalshi had failed to establish that its sports-event contracts qualify as “swaps” under federal commodities law. The decision represents a significant setback for Kalshi and other prediction market platforms that have argued their event contracts fall exclusively under the jurisdiction of the Commodity Futures Trading Commission.

“We hold that Kalshi has not shown that its sports-event contracts satisfy the statutory definition of a ‘swap’ so as to fall within the scope of the CFTC’s ‘exclusive jurisdiction,’” the three-judge panel wrote.

The ruling adds a fresh layer of uncertainty to a rapidly expanding prediction markets industry that has increasingly moved into sports, politics and other events. It also creates a growing split among federal appellate courts over the extent of the CFTC’s authority, increasing the prospect that the U.S. Supreme Court could eventually be asked to settle the dispute.

Kalshi and other prediction market operators maintain that event contracts are financial products rather than wagers. Their argument is that the Commodity Exchange Act gives the CFTC exclusive authority over swaps, preventing individual states from applying their own gambling laws.

States have taken the opposite position, arguing that sports contracts offered by prediction platforms are effectively sports bets and should therefore be subject to state gambling regulations.

The development has major implications for the industry. If the states prevail, prediction exchanges could face a patchwork of licensing requirements, restrictions, and enforcement actions across the country. If the CFTC’s position prevails, operators could potentially offer sports-related contracts under a federal regulatory framework without obtaining approval from individual states.

Sixth Circuit Rejects Federal Preemption Argument

The 6th Circuit’s decision addressed both parts of Kalshi’s argument. First, the judges found that Kalshi had not demonstrated that its sports contracts meet the statutory definition of a swap. Second, the court said that even if the contracts were considered swaps, federal law would not prevent Ohio and Tennessee from enforcing their gambling laws.

“Even assuming that Kalshi’s sports-event contracts are swaps, we alternatively hold that the CEA neither expressly nor impliedly preempts Ohio’s or Tennessee’s gambling laws,” the court said.

The decision overturns a federal district court ruling in Tennessee that had sided with Kalshi and reaffirms an earlier federal district court ruling in Ohio that supported the states’ position.

The case is part of a much broader legal confrontation involving prediction market platforms, state regulators and the CFTC.

The commission has sued nine states in an effort to defend what it considers its exclusive authority over event contracts under the Commodity Exchange Act. The 6th Circuit’s ruling rejects the central premise of that federal position, at least in the context presented by Kalshi’s sports contracts.

A Growing Split Among Federal Courts

The 9th U.S. Circuit Court of Appeals ruled last month that Nevada could regulate sports-related event contracts, finding that the contracts were sports bets rather than swaps.

The 3rd U.S. Circuit Court of Appeals reached a different conclusion in April in a case involving New Jersey. That court ruled against the state and held that the CFTC has exclusive authority to regulate swaps regardless of the type of underlying contract.

New Jersey has asked the Supreme Court to review that decision, filing its petition earlier this month.

The emerging division among the appellate courts increases the possibility of Supreme Court involvement, although it remains uncertain whether the justices will take up the New Jersey case now or wait for additional appellate decisions.

For the prediction market industry, the legal question is broader than Kalshi’s ability to offer sports contracts in a handful of states. Prediction markets have expanded rapidly by presenting contracts on future events as tradable financial instruments. Sports contracts have become an especially important part of that expansion because they create a potentially enormous market tied to games, leagues and individual sporting outcomes.

That growth has brought the platforms into direct competition with traditional sports-betting businesses while simultaneously placing them under a regulatory framework designed for derivatives markets.

The dispute therefore involves two different regulatory models. State gambling authorities generally regulate betting based on factors such as licensing, consumer protection and the legality of particular forms of wagering. The CFTC, by contrast, oversees derivatives markets and focuses on financial-market integrity, trading practices and systemic risks.

How courts classify sports-event contracts determines which regulatory system applies.

Supreme Court Question Comes Into Play

The conflicting appellate decisions make the Supreme Court relevant to the industry’s future regulatory structure. The 6th Circuit has now joined the 9th Circuit in allowing states to regulate sports-event contracts, while the 3rd Circuit has backed the CFTC’s broader jurisdiction over swaps. That means the legal status of similar products can depend on which federal appellate jurisdiction they are operating in.

The uncertainty could have practical consequences for prediction exchanges seeking to expand their sports offerings nationwide. Platforms may have to account for different legal environments while courts continue to determine whether the contracts constitute financial derivatives or gambling products.

The dispute also raises a larger question about the boundary between federal financial regulation and state gambling authority. The CFTC’s position is based on the Commodity Exchange Act and its claim to exclusive jurisdiction over swaps. The states’ position is that federal derivatives law does not prevent them from regulating contracts that function as sports wagers under their own laws.

The 6th Circuit’s unanimous ruling gives the states another significant legal victory and puts additional pressure on the prediction market industry’s federal regulatory theory. The Supreme Court may ultimately be asked to resolve the conflict, but there is no guarantee that it will intervene immediately.

President Donald Trump and House Speaker Mike Johnson are due to meet technology executives on September 29 for discussions on artificial intelligence, bringing the White House and congressional leadership into a debate that has intensified over the pace of AI development, safety, and regulation.

A source familiar with the plans confirmed the meeting, which was first reported by Axios. Further details, including the executives expected to attend and the specific agenda, have not been disclosed.

The meeting comes as the US faces a widening policy debate over how to manage sophisticated AI systems without undermining the country’s position in a technology race with China.

The move comes as major AI companies have recently increased their warnings about the risks associated with more capable new AI models. Leaders including OpenAI CEO Sam Altman and Anthropic CEO Dario Amodei have called for greater attention to AI safety and for governments to play a role in managing the technology’s development.

At the same time, there is no consensus within the technology industry over how much government intervention is appropriate. Nvidia CEO Jensen Huang, among others, has pushed back against calls to slow AI development, while technology companies continue to invest heavily in computing infrastructure and increasingly capable models.

That tension is likely to shape the September 29 discussions.

The meeting also comes as Washington and Beijing pursue substantially different approaches to AI governance.

China has established rules covering areas including AI algorithms, training data and the labeling of AI-generated content. The US has generally relied more heavily on existing laws and sector-specific oversight, while the Trump administration has opposed the creation of broad new AI-specific regulations.

Trump reiterated that position at the United Nations General Assembly this week, telling world leaders that he “rejects any attempt to construct a globalist scheme to control” AI. He described the technology as powerful and said the US would proceed carefully.

The administration’s approach creates a policy tension. Washington wants American companies to remain at the forefront of AI development, particularly as China expands its own domestic AI capabilities. At the same time, autonomous AI systems are creating questions about cybersecurity, privacy, misinformation, and potential misuse that existing regulatory frameworks may not fully address.

The September 29 meeting therefore comes at a point when the US government must balance two competing policy objectives: maintaining technological leadership and determining what safeguards should accompany increasingly capable AI systems.

Regulation Versus Development

The debate becomes necessary as AI companies move beyond conventional chatbots toward agents capable of taking actions on behalf of users.

Recent incidents involving AI agents have heightened attention on whether developers can reliably control systems once they are given access to external websites, software, and sensitive information. At the same time, companies are racing to deploy these systems commercially, creating pressure to develop products quickly.

This has produced a notable divide within the industry.

Some AI executives have argued that safety measures need to keep pace with advances in model capability. Others have warned that excessive regulation could slow American innovation and give Chinese competitors an advantage. The disagreement is not simply about whether AI should be regulated. It also concerns who should establish the rules, what activities should be covered, and whether regulation should focus on the technology itself or on specific risks and applications.

The debate is increasingly taken into consideration as Congress considers legislation aimed at AI safety and as the administration develops its own policy framework.

Competition with China provides another dimension to the meeting. The US and China are the two leading centers of AI development, with both governments treating advanced computing and artificial intelligence as strategically important technologies.

China’s regulatory approach has involved direct government rules over aspects of AI development and deployment. Washington has instead placed greater emphasis on maintaining American technological leadership while resisting proposals for comprehensive international controls.

That difference has implications beyond regulation. Rules governing advanced AI could affect where companies build data centers, how they train models, how researchers exchange information, and how AI products are deployed internationally.

For American technology companies, the policy environment will therefore influence both compliance costs and competitive strategy. The meeting with Trump and Johnson could provide an indication of whether Washington intends to maintain its relatively light-touch approach or whether pressure from lawmakers, safety researchers and parts of the technology industry will produce more targeted safeguards.

For now, the precise expected outcome is unclear. No detailed agenda or list of participants has been released. What is clear is however that AI policy has moved closer to the center of the US government’s economic and national-security agenda.

US Treasury yields held near multi-year highs on Friday as a combination of persistent inflation, elevated oil prices, stronger economic data and hawkish Federal Reserve signals pushed investors to demand higher returns on government debt.

The benchmark 10-year Treasury yield was little changed around 5.17%, after climbing to about 5.22% on Thursday, its highest level since 2007. The 30-year Treasury yield held near 5.46%, after reaching its highest level since 2004, while the two-year yield remained around 4.90%.

The move marks a sharp repricing in the US bond market. The 10-year yield has risen about 20 basis points in two sessions, its largest two-day increase since April 2025, while the 30-year yield has gained about 16 basis points over the same period. The bond selloff is not confined to the United States. Government bond yields have climbed across major developed markets, including Japan, Britain and the euro zone, as investors reassess the outlook for inflation and monetary policy.

The immediate pressure on Treasuries comes from a combination of factors. Federal Reserve officials have signaled that additional rate increases may be required, while oil prices remain above $100 a barrel and economic activity has remained stronger than expected.

Federal Reserve officials have become more concerned that inflation is proving more persistent than initially anticipated. St. Louis Fed President Alberto Musalem said this week that further monetary restraint would probably be required, explaining that both strong demand and supply pressures were keeping inflation risks elevated. The Fed’s preferred inflation gauge was running at 3.7% year over year in July, well above its 2% target.

The shift in expectations is considered relevant because the Fed raised interest rates by 25 basis points last week. Rate markets are now pricing substantial odds of another increase at the October meeting, extending a tightening cycle that investors had previously expected to be closer to its end.

That has changed the calculation for bond investors. Rather than assuming that weaker growth will eventually bring inflation down and allow the Fed to ease policy, markets are now pricing the possibility that interest rates will remain higher for longer.

The Long End is Becoming The Bigger Problem

The most consequential move has occurred at the long end of the Treasury curve. The 30-year yield’s rise toward 5.5% is important because long-term government borrowing costs feed directly into mortgage rates, corporate financing and the valuation of long-duration assets.

US mortgage rates have moved toward 7%, adding pressure to an already constrained housing market. Higher mortgage costs can reduce affordability, discourage transactions and weaken construction activity even if the wider economy remains resilient.

The rise in long-term yields also reflects concerns that extend beyond the Fed’s overnight policy rate. Investors are demanding more compensation for holding long-dated government debt as inflation risks, fiscal pressures and the supply of Treasury securities remain elevated.

That is why a Fed rate hike alone does not explain the current bond selloff. Even if the central bank stopped raising its policy rate, long-term yields could remain under pressure if investors believe the government will need to issue substantial amounts of debt or if inflation expectations remain elevated.

“The world’s bond markets are screaming, and ignoring it could prove very expensive,” said Nigel Green, CEO of deVere Group.

“Once risk-free rates sit above 5% in the world’s largest economy, every asset on the planet has to justify its price against that. Equities, property, private credit, emerging market debt, nothing’s immune.”

That repricing mechanism is becoming more relevant for financial markets. A 5%-plus risk-free yield provides investors with a much higher return alternative to equities and other assets than they had available during the ultra-low-rate era. It also raises the discount rate applied to future corporate earnings. Companies whose valuations depend heavily on profits expected years into the future are particularly sensitive to higher Treasury yields.

Oil is Keeping The Inflation Problem Alive

The bond selloff is also closely linked to the energy shock caused by the conflict involving the United States and Iran.

Oil prices eased on Friday as markets considered the possibility of a truce and a phased reopening of the Strait of Hormuz. Negotiators were reportedly exploring a pathway that could eventually restore shipping through the critical waterway. But crude remained above $100 a barrel, leaving a substantial inflation risk in place.

That decline in oil prices provided some relief to bond markets and helped support equities. But the market response remains cautious because a diplomatic breakthrough has not been secured.

“Markets tend to buy the rumor on signs of better news coming from the Middle East,” said Nordea chief market strategist Jan von Gerich. “But there is no quick resolution and the weekend is approaching so we could see some caution.”

The energy shock is an issue because it can complicate the Fed’s efforts to bring inflation down. Higher fuel and transportation costs can feed into a broad range of goods and services, while an oil shock can also weaken household purchasing power.

Fed officials have indicated that inflation is no longer simply an energy problem. Kashkari said price pressures remain elevated across the economy even after excluding food and energy, while Musalem said the inflation problem extends beyond oil into other commodity and import costs.

That makes the current environment more difficult for the central bank. If inflation were being driven exclusively by an oil shock, policymakers could potentially look through part of the increase. Persistent underlying inflation combined with strong demand gives them a stronger reason to maintain restrictive policy.

Global Bond Markets Are Sending the Same Signal

The Treasury selloff has occurred alongside a broader global bond repricing. Japan’s 10-year government bond yield touched 3.115%, its highest level since 1996. Eurozone yields eased on Friday but remained on course for another weekly increase, while other major sovereign bond markets have also experienced sharp increases in borrowing costs.

The global nature of the move is of the essence because investors allocate capital across markets. When yields rise in Japan, Europe and the US simultaneously, the relative attractiveness of risk assets changes across the global financial system.

Japan is a country of interest because its bond market has historically operated with much lower yields than the United States and Europe. A sustained rise in Japanese yields can influence domestic capital allocation and the attractiveness of overseas assets for Japanese investors.

Meanwhile, five of the Group of 10 major central banks have raised rates this month, according to the market assessment in the supplied data. Norway raised rates on Thursday, and Sweden’s Riksbank has indicated that it could follow with a hike before year-end.

The message from global policymakers is increasingly consistent: the inflation shock is proving persistent enough to require tighter monetary conditions.

Stocks Are Resisting The Bond-Market Pressure

Equities have so far shown greater resilience than might be expected from the scale of the bond selloff.

Global stocks were heading toward their strongest weekly performance since early August on Friday, supported by renewed enthusiasm around artificial intelligence and hopes that improved energy supplies from the Middle East could reduce inflationary pressure.

The divergence is striking. On one side, government bonds are pricing in a significantly more difficult inflation and fiscal environment. On the other, technology stocks and AI-related companies are attracting capital on expectations of continued investment and earnings growth.

However, analysts say that resilience may not last indefinitely if yields continue rising.

The 10-year Treasury is a foundational reference rate for valuations across financial markets. Higher yields raise financing costs for companies, increase mortgage rates for households, and provide investors with a more attractive alternative to risk assets.

This has become crucial for technology companies whose valuations have benefited from expectations of strong future cash flows. The recent AI rally has provided a powerful counterweight, but the higher the risk-free rate moves, the greater the earnings growth required to justify elevated valuations.

Therefore, the bond market is creating a higher hurdle for the equity market even as AI optimism continues to support technology shares.

The Dollar Benefits From Higher Rates

Higher US yields have also supported the dollar. The dollar index was slightly lower on Friday but remained on course for a second consecutive weekly gain after reaching its highest level since late July earlier in the week.

The currency has benefited from expectations of further Fed tightening because higher US interest rates can increase the relative return available on dollar-denominated assets.

The yen was an exception on Friday, with the dollar falling 0.4% against the Japanese currency to around 158.23. The move followed comments from Japanese Finance Minister Satsuki Katayama that Trump had raised concerns about yen weakness during a meeting with Japanese Prime Minister Sanae Takaichi.

The combination of rising US yields and a weak yen keeps pressure on Japanese policymakers, particularly because a wider interest-rate differential can encourage capital to flow toward dollar assets.

Markets Face A Difficult Policy Mix

The major problem for investors is that several forces are now pushing in the same direction. Oil prices are keeping inflation elevated. Strong US economic activity reduces the urgency for the Fed to support growth. Central bankers are signaling that additional rate increases may be necessary. Governments continue to carry substantial debt loads, increasing the sensitivity of fiscal accounts to higher borrowing costs.

At the same time, AI investment is supporting corporate earnings expectations and equity valuations, while hopes for improved Middle East energy supplies are preventing a more severe risk-off move.

ING analysts Padhraic Garvey and Benjamin Schroeder said they believed much of the rate-hike risk had already been priced into markets, but warned that government bond yields could remain under pressure because of debt dynamics, particularly around the 10-year area.

The Treasury buyback programme has provided some support to market functioning, but it has not reversed the broader upward pressure on yields.

The key question now is whether the recent surge in yields represents an overshoot or the beginning of a longer adjustment to a world of structurally higher inflation and government borrowing costs.

Nordea’s von Gerich said recent bond moves had gone further than economic conditions appeared to justify and saw room for yields to decline. That possibility depends heavily on the next pieces of economic and geopolitical information. A meaningful decline in oil prices, evidence that inflation is cooling, or a deterioration in economic activity could reduce expectations for additional Fed tightening and bring Treasury yields lower.

Conversely, another increase in energy prices, stronger-than-expected economic data, or further evidence of persistent inflation could reinforce the bond selloff.

Tesla is beginning deliveries of its Semi electric trucks to a new group of customers this week, executives said at the company’s Nevada factory, marking a major step for a vehicle that has taken almost nine years to move from unveiling to broader commercial production.

The Semi expands Tesla’s business beyond passenger cars and SUVs into heavy-duty freight, a market dominated by diesel-powered trucks. The development is taking place as US freight operators face higher fuel costs and growing pressure to reduce emissions, while competition in electric commercial vehicles continues to expand.

Tesla did not disclose production volumes or pricing for the Semi at an event at its Sparks, Nevada, plant on Thursday. The company reiterated its target of producing 50,000 trucks a year at the facility.

Chief Executive Elon Musk did not attend the event but appeared in a recorded message encouraging potential customers to place additional orders.

“I’d recommend placing more orders if you haven’t already, but the waiting list is already pretty significant,” Musk said.

Customers including PepsiCo, DHL and US Foods were represented at the event, with trucks bearing their logos displayed outside the factory.

The long-range version of the Semi is designed to travel 500 miles on a single charge when fully loaded, according to Tesla. A standard version has a claimed range of 325 miles.

“That’s 500 real-world miles. Our customers have validated it,” said Dan Priestley, Tesla’s director of Semi truck engineering.

The range is crucial for Tesla because heavy-duty trucking presents a much tougher electrification challenge than passenger vehicles. Long-haul operators need vehicles capable of carrying heavy loads over substantial distances while minimizing charging downtime. For fleet owners, the economics therefore depend not only on battery range but also on charging infrastructure, vehicle utilization, maintenance and the total cost of ownership.

Nine years from unveiling to scale

Tesla unveiled the Semi in 2017, initially targeting production in 2019. The vehicle eventually entered limited customer deliveries in late 2022, including early deployments with PepsiCo, but the company has struggled to move from limited production to the scale originally envisioned.

The first Semi intended for high-volume production rolled off the production line in April, while Tesla’s July shareholder letter said production would begin in 2026, removing an earlier expectation for volume production this year.

The delays have given established truck manufacturers and newer electric-truck companies additional time to develop competing products. Tesla is therefore entering a commercial market in which fleet operators have more electric options than they did when the Semi was first unveiled. The company also faces uncertainty over the trajectory of US electric-vehicle policy and incentives, which can influence the economics of electric truck purchases.

Still, recent orders suggest there is demand for the vehicle among large fleet operators.

Tesla this week received an order for 2,500 Semis from a coalition of major cargo-owning companies, including Microsoft and PepsiCo, according to clean transportation nonprofit Catalyst Mobility. The group said the order is nearly twice the size of the existing US fleet of electric Class 8 trucks.

Swedish freight technology company Einride also announced last month that it would add 500 Tesla Semis to its fleet. Those orders give Tesla a potentially important customer base as it attempts to establish the Semi as a commercially viable alternative to conventional heavy-duty trucks.

The Economics Will Matter More Than The Launch

The broader significance of the Semi is likely to depend on whether Tesla can demonstrate that electric freight can work economically at scale.

For fleet operators, fuel savings can be substantial when electricity replaces diesel, particularly when trucks travel high annual mileage. But those savings have to be weighed against the upfront cost of the vehicle, charging infrastructure, and the operational implications of charging heavy trucks.

Tesla’s ability to manufacture 50,000 Semis annually would also represent a substantial expansion of the US electric heavy-truck market. The company has not disclosed when it expects to reach that production rate.

The scale target is important because the Semi requires Tesla to solve a different manufacturing and supply-chain problem from the one associated with its passenger vehicles. Heavy trucks require large battery packs, while fleet customers generally expect high reliability and predictable operating costs.

The vehicle also places Tesla in direct competition for fleet budgets rather than individual consumers. A fleet operator can evaluate a truck using relatively straightforward measures such as cost per mile, payload, uptime, charging time, and residual value. Meeting those requirements consistently will determine whether early orders translate into repeat purchases.

The Semi’s delayed development has also changed the competitive landscape. When Tesla introduced the truck in 2017, it was making a relatively early bet on battery-electric long-haul transportation. Nearly a decade later, commercial fleet electrification has become a broader industry effort involving truck manufacturers, charging companies, and logistics operators.

That makes the current deliveries more than a product launch. They mark Tesla’s attempt to establish a position in a commercial vehicle market where operational performance and economics can matter more than the brand recognition that has helped drive its passenger-car business.

The biggest risk facing the AI industry may not be that artificial intelligence fails to live up to its promise, but that it becomes too cheap to support the extraordinary sums being spent to build it.

OpenAI and Anthropic are committing enormous amounts of capital to computing power, data centers, chips and other infrastructure in a race to remain at the frontier. At the same time, Chinese developers such as DeepSeek, Alibaba’s Qwen, Zhipu AI and Tencent are narrowing the performance gap while offering models at substantially lower costs.

That combination is raising a difficult question for investors: what happens to the economics of the frontier AI business if customers can obtain sophisticated models without paying the premium required to support the industry’s enormous infrastructure buildout?

Scott Wilson, whose early investment in SpaceX generated a multibillion-dollar windfall for Washington University’s endowment, believes the answer could be painful for the industry’s most highly valued companies.

“These trillion-dollar-plus frontier companies are not worth the liabilities that they signed up for,” Wilson said. “There’s going to be a ton of free alternatives.”

Wilson’s argument challenges one of the central assumptions behind the AI investment boom: that spending more than competitors on infrastructure will necessarily create a durable competitive advantage. His view is that the opposite could happen if advances in open-weight and lower-cost models make AI increasingly difficult to monetize.

The leading AI companies are spending hundreds of billions of dollars on computing capacity, data centers, chips, and related infrastructure in an effort to maintain an advantage over competitors. The underlying assumption is that superior models will command enough revenue to justify that investment.

But if increasingly capable models become available at dramatically lower prices, the economics could change. Customers may become less willing to pay premium prices for proprietary systems if open-weight or lower-cost alternatives deliver comparable results.

That is the scenario Wilson believes the market is underestimating.

The China Challenge Is Changing The Economics

Wilson’s skepticism has hardened as Chinese model developers have narrowed the performance gap with US frontier laboratories. Companies including DeepSeek, Alibaba’s Qwen, Zhipu AI and Tencent have emerged as credible alternatives, while open-weight models have made it easier for businesses to deploy AI without relying entirely on the largest US providers.

Wilson said conversations with colleagues operating in China have reinforced his view that progress in open-weight models is happening rapidly.

He has also been hearing similar signals from companies in Washington University’s investment portfolio.

“Whenever we talk to our portfolio companies, especially the ones who are heavy of AI, they are all moving aggressively towards open source,” he said. “It’s like any other high-cost U.S. good that has to compete with a low-cost import, particularly from China.”

That comparison captures the heart of Wilson’s thesis. If AI models become increasingly commoditized, the economics could begin to resemble other industries in which a high-cost producer struggles to maintain pricing power against a lower-cost competitor. The issue would not necessarily be whether US laboratories can build better models, but whether the incremental improvement is large enough for customers to justify paying substantially more.

OpenRouter data provide some evidence of the shift, although the platform represents only one slice of the AI market.

DeepSeek accounted for 25.3% of text-model requests on OpenRouter, compared with 18.6% for OpenAI and 2.9% for Anthropic.

The figures do not establish overall market share or revenue, and OpenRouter users are not necessarily representative of enterprise AI customers. They nevertheless illustrate the growing willingness of developers to experiment with alternatives to the dominant US laboratories.

That creates a difficult proposition for OpenAI and Anthropic. Their enormous capital requirements are based partly on the expectation that customers will continue to value frontier performance enough to sustain premium pricing.

If the performance gap narrows faster than the cost gap, that assumption becomes harder to defend.

Khosla Sees Infrastructure As The Moat

Vinod Khosla, the billionaire venture investor and early OpenAI backer, sees the economics very differently.

He believes that Wilson is focusing too heavily on the model itself and not enough on the infrastructure required to build, train, and operate it.

“People like that are silly, and they don’t understand how this works,” Khosla said when told of Wilson’s view. “They have this notion that the model is the value.”

Khosla’s argument is that a closed-model company can eventually control more of the technology stack, allowing it to lower its underlying costs even if competitors offer models for less.

That stack includes computing infrastructure, electricity, data centers, chips, software, and inference systems.

OpenAI, for example, is working on its own inference hardware, including its Jalapeo inference chip, which Khosla says could reduce the company’s dependence on Nvidia and third-party cloud providers.

“I’m not talking price, I’m talking about cost,” he said. “From power to data center to chips, to infrastructure software to inference models, the cost of the stack is almost certainly going to be lower in closed-source models than open-source.”

This is the central counterargument to Wilson’s thesis.

An AI company may be able to charge less for its models while still maintaining attractive margins if it can lower the cost of producing each response. Owning or co-designing critical infrastructure could give frontier laboratories advantages that are not visible when models are compared simply on headline API prices.

The difference is between price and unit economics.

A cheap open-weight model can be attractive to customers, but running that model may still require expensive computing infrastructure. A vertically integrated AI company could potentially offset higher model development costs by lowering its cost per inference through specialized chips, optimized software, and dedicated data centers.

If Khosla is right, today’s infrastructure spending is not simply an expense. It is the foundation of a cost advantage that could become more important as AI usage scales.

The Real Test Is Whether AI Becomes A Commodity

The disagreement exposes two competing visions of the AI industry. Wilson’s thesis assumes that model capabilities will converge quickly enough that the model itself becomes difficult to monetize. Under that scenario, customers will have little reason to remain locked into expensive proprietary systems when cheaper alternatives are available.

Khosla’s thesis assumes that the leading laboratories will maintain meaningful technological advantages while simultaneously gaining control over the infrastructure underneath their models. That could allow them to reduce costs, improve performance, and retain pricing power.

The outcome depends heavily on the pace of commoditization.

If models become interchangeable, the enormous capital commitments made by OpenAI and Anthropic could become problematic. The companies would have to generate enough revenue to service infrastructure commitments even as customers demand lower prices.

If frontier models remain materially better, however, the infrastructure race could reinforce their lead. Training sophisticated models requires enormous amounts of computing power, and the companies with access to the largest pools of capital could maintain advantages that smaller competitors cannot easily reproduce.

There is also a middle scenario in which the model layer becomes cheaper while value migrates elsewhere.

AI customers may ultimately care less about which laboratory produced the underlying model and more about applications, proprietary data, distribution, workflow integration, and reliability. In that world, model providers could face falling margins even as the broader AI economy continues to grow.

The implication is expected to be significant for companies valued on the assumption that model development itself will capture a large share of the industry’s eventual profits.

OpenAI And Anthropic Face A Different Test From Their Investors

The debate comes at a critical point for both companies. OpenAI has already demonstrated that investors are willing to assign enormous value to its position at the center of the AI ecosystem. Anthropic, meanwhile, is expected to go public next month, putting its financial model under much closer scrutiny from public-market investors.

Then, the question will not simply be whether the companies can produce increasingly capable models. It will be whether those models can generate enough recurring revenue and gross profit to justify the capital required to develop and operate them.

For now, the market has largely rewarded the spending race. The biggest laboratories have attracted extraordinary amounts of capital, while chipmakers, cloud providers and data-center operators have benefited from the resulting infrastructure buildout.

But a large infrastructure footprint is only an advantage if demand and margins grow quickly enough to absorb it.

Wilson’s SpaceX investment exemplified the power of being early to a technological transformation. His current argument is effectively that investors may be making the opposite mistake with AI: paying enormous prices for companies before the economics of the industry have been established.

Khosla believes those economics are precisely what the infrastructure race will establish.