A further rise in long-term U.S. Treasury yields toward 5% could trigger a sharp de-risking across equities and send the S&P 500 down as much as 20%, according to Phillip Colmar, a partner at market research firm MRB Partners.

The warning comes after a turbulent week in bond markets, with investors increasingly focused on the combination of elevated inflation, strong economic growth expectations and rising U.S. government debt.

The benchmark 10-year Treasury yield climbed above 4.7% this week, while the 30-year yield moved above 5.2%, increasing pressure on equity valuations and raising concerns about how much higher borrowing costs could affect corporate investment and profits.

Colmar said stocks could be “on the brink” of a de-risking episode if long-term yields continue to rise.

“If it looks like it’s going up and it’s not going to be stopped, you could end up with a de-risking event that starts below 5%,” Colmar told Business Insider, referring to the 10-year Treasury yield.

He estimated that a move toward 5% could ultimately result in a 15% to 20% decline in the S&P 500, although he did not make a specific forecast for where Treasury yields will settle.

The warning comes at a particularly sensitive point for equity markets. Investors have spent much of the past year placing large bets on artificial intelligence, expecting rapid productivity gains and strong earnings growth from companies building AI infrastructure and applications.

Higher long-term yields threaten that trade by changing the relative attractiveness of stocks. Treasury securities are generally viewed as the benchmark risk-free asset, so when their yields rise, investors can demand higher expected returns from equities to compensate for taking additional risk.

The effect weighs largely on growth stocks, whose valuations depend heavily on earnings expected several years into the future. Higher discount rates reduce the present value of those future earnings, putting pressure on companies with high valuations and long-duration growth expectations.

The AI sector is especially exposed because the industry’s expansion requires enormous amounts of capital.

Technology companies are spending heavily on data centers, computing infrastructure, chips, and electricity capacity to support sophisticated AI systems. Much of that investment is expected to generate returns over several years rather than immediately.

Higher financing costs could therefore reduce returns on invested capital and make it more difficult for companies to justify the scale of their spending.

Colmar said the combination of elevated expectations and rising borrowing costs could create an “air pocket” in the market.

“You just end up with an air pocket between what might be a decent theme, but expectations were just too high and they can’t be met now,” he said.

That dynamic could be especially damaging if investors begin lowering earnings forecasts for AI companies at the same time that they demand greater discipline over capital spending.

The concern is not necessarily that the AI investment cycle will collapse. Rather, higher rates could force investors to reassess how quickly companies can turn enormous infrastructure spending into revenue and profits.

The Treasury’s response to rising long-term yields could also become a source of market uncertainty.

Treasury Secretary Scott Bessent said this week that the department would increase its bond buybacks in an effort to reduce the supply of longer-dated securities available in the market. Lower supply can support bond prices and, in turn, put downward pressure on yields.

Colmar warned, however, that the policy could have the opposite psychological effect if investors interpret the move as an attempt to suppress yields without addressing the underlying reasons for the increase.

He also pointed to the lack of coordination with the Federal Reserve, which has been allowing Treasury securities to mature and roll off its balance sheet as part of its quantitative-tightening process.

If investors conclude that the Treasury is becoming increasingly concerned about the level of long-term yields, the intervention could undermine confidence rather than restore it.

“The market sniffs out the panic,” Colmar said.

The underlying drivers of higher yields are considered important. Rising government borrowing needs increase the amount of debt that investors must absorb, while expectations for stronger economic growth can push yields higher by reducing expectations for monetary easing.

Persistent inflation creates another challenge because investors demand greater compensation for holding long-duration bonds when they believe purchasing power will erode more rapidly.

The U.S.-Iran war has added another inflationary dimension by increasing uncertainty around energy prices and supply chains, further complicating the outlook for inflation and interest rates.

A sustained rise in long-term yields would therefore create a difficult environment for both bonds and equities. Higher yields can attract capital away from stocks while simultaneously increasing the discount rate used to value companies and raising financing costs across the economy.

Colmar remains positioned for economic growth rather than an immediate recession, but he recommends investors reduce exposure to areas most vulnerable to higher rates if Treasury yields continue to rise.

One strategy is to trim exposure to AI stocks and increase allocations to defensive sectors, with healthcare among his preferred areas because of its relatively lower sensitivity to interest rates. Financial stocks could also benefit from a higher-rate environment, depending on the shape of the yield curve and the effect of borrowing costs on banks’ net interest margins and credit quality.

Exchange-traded funds such as the Vanguard Health Care ETF and Financial Select Sector SPDR Fund provide exposure to those sectors.

The broader message from the bond market is that the next major risk to equities may not come from an abrupt deterioration in economic growth. It could instead come from a gradual repricing of the cost of capital.

That matters for investors because a strong economy can coexist with falling stocks if Treasury yields rise quickly enough. In such an environment, corporate earnings may continue to grow while equity valuations contract as investors demand greater returns to compensate for higher interest rates.

The AI sector has benefited from expectations of enormous future earnings, but those expectations are increasingly being tested against the cost of building the infrastructure needed to produce them.

If the 10-year Treasury yield approaches 5%, investors may begin asking whether projected AI returns are large enough, and arrive quickly enough, to justify current valuations and capital spending. That could turn higher bond yields from a macroeconomic concern into a direct threat to one of the market’s most crowded investment themes.

A mysterious artificial intelligence model called Ox Alpha has triggered intense speculation among developers and AI researchers after appearing on OpenRouter without revealing the company or team behind it.

The free model was introduced on OpenRouter on Thursday as a “stealth model” designed for coding, long-running agentic work and production workloads. OpenRouter’s listing says it is developed and operated by an anonymous third-party provider and makes clear that OpenRouter itself is not the model’s developer, owner or provider.

The model’s emergence has attracted attention partly because of its reported capabilities. OpenRouter describes Ox Alpha as suited to long-horizon software engineering, complex reasoning and workflows that combine text with visual context. It has a context window of about 1.05 million tokens and is currently offered at no charge through the platform.

Stripe CEO Patrick Collison, whose company recently agreed to acquire OpenRouter, called Ox Alpha “very impressive.” Stripe’s acquisition of OpenRouter, reportedly valued at more than $8 billion, gives the payments company a direct position in the infrastructure used by developers to access and route requests among competing AI models.

But the technical performance of Ox Alpha is only part of the story. The absence of a disclosed developer has turned its launch into an exercise in AI model forensics, with researchers comparing its behavior, API characteristics and capabilities against known systems in an attempt to identify its origin.

Much of the early speculation has centered on China, particularly Z.ai, the Chinese developer behind the GLM family of models.

AI analyst Andrew Curran said the initial discussion focused heavily on whether Ox Alpha was related to Z.ai’s GLM models, although he later said confidence in that theory had weakened as more information emerged.

There are reasons for the comparison. Ox Alpha and Z.ai’s GLM models share several technical characteristics, including a 1.05 million-token context window and a focus on complex reasoning, coding, and long-running agentic workloads. OpenRouter’s own comparison pages show that Ox Alpha and GLM 5.2 and 5.3 have identical context-window lengths, although Ox Alpha remains officially attributed only to the anonymous “Stealth” provider.

That similarity, however, is not proof of common ownership.

The public GLM 5.3 model is listed by OpenRouter as a Z.ai model designed for complex software engineering and long-horizon agent tasks. Ox Alpha has a similar positioning but also supports workflows involving visual context, while the publicly listed GLM 5.3 is described as text-only.

That has led some researchers to speculate that Ox Alpha could be an unreleased or experimental model related to the GLM family rather than a publicly available Z.ai model. Other theories have pointed toward models being developed by U.S. technology companies, including Microsoft’s MAI family, but there is currently no public confirmation establishing any of these connections.

Some developers have attempted to identify the model by examining its API configuration and behavior. Community researchers have pointed to similarities between Ox Alpha’s interface and GLM models, while anecdotal reports from users have also fueled speculation that it may identify itself as a Z.ai or GLM system in certain interactions.

Such behavioral evidence should be treated cautiously. A model’s response about its own identity is not reliable evidence of who trained or operates it because that information can be influenced by system instructions, training data, or provider configuration.

The anonymity is notable because OpenRouter says prompts and completions submitted to the Stealth provider are retained by that provider. OpenRouter also says those prompts and completions are not used for training, while other uses are governed by the Stealth Model Terms.

That creates a separate issue for developers considering the model for production workloads. A free, highly capable model can be attractive for coding agents and other applications that consume large amounts of inference, but the identity of the provider and the handling of submitted data are important considerations when applications process proprietary code or sensitive business information.

Ox Alpha’s arrival also comes at an important moment for the AI model market. Developers are now using platforms such as OpenRouter to switch between models based on performance, cost, availability, and task requirements rather than committing their applications to a single AI provider.

The model’s free availability gives it another advantage during the preview period. Reports indicate that the provider is offering substantial capacity, allowing developers to test the system at a scale that would normally carry significant inference costs. That has helped accelerate experimentation and, in turn, the effort to determine where the model came from.

The mystery shows that it has become difficult to determine the provenance of frontier AI systems simply by looking at their performance. Model providers can release systems anonymously, operate them through third-party infrastructure and expose only an API, leaving researchers to infer their origins from technical fingerprints.

For Z.ai, if the speculation ultimately proves correct, the appearance of a powerful unreleased or experimental GLM-related system would provide another indication of the pace at which Chinese AI developers are competing with U.S. companies in advanced reasoning and agentic software.

But that conclusion remains unconfirmed.

That means the only verified information is that Ox Alpha is a powerful, free reasoning model available through OpenRouter, that its provider has deliberately chosen anonymity during the preview, and that OpenRouter says it does not own or operate the model.

The uncertainty itself has become part of the model’s appeal. In a market where OpenAI, Anthropic, Google, Meta, and Chinese developers such as Z.ai openly attach their names to capable systems, Ox Alpha has generated attention by doing the opposite: releasing a capable model first and leaving the industry to guess who built it.