Chinas J-35 Ski-Jump Flights: USS Ford Stealth Gap

A single photo viral on social media, just turned a rumor into fact. The image captures a J-35, afterburners blazing, tail nozzle spitting a long tongue of flame. At the runway's end, the deck kicks up sharply. That's the unmistakable profile of a ski-jump carrier. Is it the Liaoning or the Shandong? The photo alone can't tell us.

Afterburners blazing, a J-35 roars toward a ski-jump deck under cover of darkness.

But look closer. A nighttime launch. An empty sea. A ski-jump deck. Those details scream one thing: this stealth fighter is already combat-deployed. You don't pick a dark night for a test flight.

For a long time, skeptics doubted the J-35 could ever fly from a ski-jump deck. The problem? Catapult launches and ski-jump takeoffs punish an airframe in completely different ways. An electromagnetic catapult yanks the aircraft forward along its spine. That brutal tug demands a beefed-up landing gear and a reinforced fuselage. The J-35 has already proven it can handle that. It's been flung off the Fujian's catapults repeatedly. So that box is checked.

But a ski-jump takeoff hits differently. The shock doesn't run lengthwise. It slams along the X-axis. As the deck ramps upward, the jet gets a violent nose-up jolt right at the point of departure. Surviving a catapult shot doesn't guarantee you can handle a ski-jump. Basic aeronautical engineering.

Now, you might ask: what about landing? A J-35 slamming onto the deck hits harder and faster than any takeoff. Isn't that a tougher test than the X-axis jolt? Fair point. But takeoff and landing loads are different beasts. Each scenario demands its own round of testing, one by one.

And passing tests is just step one. It doesn't mean the jet is ready to move in. The J-35 is a fifth-generation stealth fighter. Its skin coatings and onboard systems need a whole different level of care compared to the old J-15. The carrier's hangar and maintenance gear must be ripped out and rebuilt. Only after those modifications are done, and the new maintenance routines are proven, can the fighter truly call the ship home.

All three Chinese carriers are now ready to launch the stealthy J-35.

So where do things stand? All three of China's carriers can now field the J-35. The ski-jump deck is no longer a barrier. The only bottleneck is the hangar and support system overhaul. For the Dalian Shipyard, that's not a technical headache. The real question is whether the navy has the appetite to do it.

The US Navy's Awkward Reality

Now turn that same question on the US Navy. The picture gets awkward fast. The USS Gerald R. Ford has long insisted it cannot carry the F-35C stealth fighter. The ship was simply not designed for it.

The gap in stealth fighter deployment? It's widening, and not in the US Navy's favor.

The Navy's earlier explanation: the Ford's design was frozen before the F-35C entered service. You can't build custom maintenance facilities for a jet that doesn't exist yet. That sounds reasonable. But it falls apart under scrutiny.

First, the Ford is the US Navy's next-generation flagship carrier. Its designer, Newport News Shipbuilding, handled everything: detailed 3D engineering, block construction, final assembly, mooring trials. All of it.

And yet, this is a shipyard with a century of experience. It failed to make any forward-looking provisions for the next-generation carrier aircraft. Not even a reserved maintenance workshop. That is hard to swallow.

Second, it's not that US carriers can't handle the F-35C. It's that the Ford can't. The USS Carl Vinson, commissioned in the 1980s, operates the jet just fine. The Ford, built in the 2010s, cannot. That contrast is striking.

Whether the Ford's electromagnetic catapult system is incompatible with the F-35C is a separate question.  On stealth fighter deployment, the gap between the Chinese and US navies is widening. The J-35 is now the world's first stealth fifth-generation fighter to complete an electromagnetic catapult launch. And every Chinese carrier can handle it.

Meanwhile, the US Navy still relies on the F/A-18E/F as its workhorse. That jet's basic design goes back to the 1970s. In any future encounter at sea, it would be a case of China using stealth fighters to bully older-generation jets.

More than just carriers and jets

China's stealth aviation push is systematic. The entire fleet's combat capability is on the same trajectory. Compatibility for catapult-equipped carriers solved. The needs of ski-jump carriers did not go ignored.  All three carrier lines are advancing in parallel.

On the US side, the Ford has stumbled repeatedly on new systems: electromagnetic catapults, advanced arresting gear. It still hasn't hit its original combat capability targets. The plan to deploy stealth fighters has been delayed again and again.

The contrast answers a clear question: which navy adapts better to a changing battlefield? China's or America's?

One Navy Adapts, the Other Stumbles

Talk of a changing battlefield? It really tests one thing: how a navy uses its carrier aircraft. What a carrier flies dictates how far and how stealthily it can project power.

One navy adapts. The other stumbles. It's clear who's quicker on their feet.

Now put a J-35 on a ski-jump carrier. You get a leap in stealth strike capability. A platform once written off as fit only for conventional fighters suddenly joins the fifth-generation package.

For a navy operating three carriers built along different technical paths, a single stealth fighter model changes everything. Training, maintenance, and talent pipelines merge into one. The combat payoff is not just additive. It is multiplicative.

Now look at the United States. The Ford's journey from launch to commissioning was tortuous. And integrating a stealth fighter? That adds another layer of uncertainty. A platform purpose-built for a next-generation carrier is proving slow to actually host the next generation of aircraft. That gap is no accident. It reflects real shortcomings in design and construction.

The Chinese Navy's method: get the platform working step by step, then tailor maintenance to the aircraft's quirks. The US Navy's method: burn time in endless integration struggles between systems and the platform. Which path is more pragmatic? Time has already handed down a partial verdict.

The J-35 on a ski-jump deck: just one milestone in China's stealth transformation.

Putting J-35 on a ski-jump carrier is just the beginning of China's drive toward a fully stealth navy. The real story is the diffusion of stealth aviation power: from individual platforms to the entire system. Once Liaoning, Shandong, and Fujian all fly the same stealth fighter, China's navy will have far more options for projecting power on the high seas.

For the US military, the problem is stark: how to get the Ford to quickly accommodate the F-35C. That remains a hard problem for its navy. One step forward, one step behind. That neatly captures the different stages these two navies now occupy.

The Art of Stopping War

** 博客文章文責自負,不代表本公司立場 **

The Botched AI Report That Almost Went Hot

CNN reported that an AI-generated intelligence report nearly led the U.S. military to take military action against a Chinese ship.

CNN cited four sources. They said the report circulated inside the U.S. military this spring, during a conflict with Iran, and triggered heightened alert. It claimed the Chinese ship was carrying components related to a nuclear weapons research and development program in the Middle East.

The U.S. military immediately launched an operation to intercept the vessel. Armed personnel were preparing to board, and U.S. military aircraft had already been scrambled.

At one point, U.S. military armed personnel planned to board and intercept the Chinese vessel involved.

Just as the operation was about to begin, U.S. officials dug into the report. It had been compiled by an analyst at U.S. Special Operations Command. They discovered it had been generated with the help of artificial intelligence. The analyst's chatbot had misidentified the cargo.

CNN noted it could not confirm what the misidentified cargo actually was. Sources bluntly described the report as "completely wrong" and said the episode "almost started a war."

U.S. military analysts now routinely use artificial intelligence to blend different types of intelligence and generate reports.

The sequence of events is not complicated. But the deeper problem it exposes deserves serious reflection.

According to CNN, the analyst first queried a chatbot about cargo manifest intelligence from U.S. Pacific Command. The chatbot then merged open-source intelligence with classified signals intelligence from inside the U.S. government and reached a conclusion about the cargo.

The analyst used artificial intelligence again to package the finding into a standard-format intelligence report and sent it out. These reports carry a high degree of trust in the U.S. military command system. It quickly entered the operational process.

The report did not specify whether the chatbot was a commercial product or a U.S. government-developed system. But a former senior official familiar with AI in the U.S. military and intelligence agencies revealed that many tools inside the U.S. military are essentially commercial products that have been repackaged.

U.S. Defense Secretary Pete Hegseth is pushing the Artificial Intelligence Acceleration Strategy to speed up the military's adoption of AI.

The Pentagon's AI Sprint Outruns Its Checks

That remark lands at the heart of the problem. U.S. military and intelligence officials are sprinting to push AI into every operational link. The work spans raw intelligence analysis, strike target screening, and the unglamorous back office: budgets, logistics, supply chains.

Defense Secretary Hegseth gave the push a name in January. His Artificial Intelligence Acceleration Strategy calls for loosening experiments, clearing bureaucracy, and concentrating resources to accelerate military AI. A memo alongside it says the framework will spread AI across every part of the force.

The acceleration is clear. The review and verification mechanisms are clearly out of step. Multiple U.S. officials familiar with the matter admit that agencies use different tools, follow incompatible instructions and security standards, and have yet to build a unified system to verify the intelligence those tools produce.

The result is a patchwork. Different military and intelligence units deploy their own AI systems, so reliability and computing performance vary enormously.

Without unified standards, the danger writes itself. A model prone to "hallucination" can push a seriously inaccurate report to the edge of military operations with no check at all.

It is not a hypothetical. Multiple sources say that with such tools becoming common inside the U.S. government, the model hallucination that hit this analyst has long been routine across the intelligence community. Young analysts lean on these tools hardest, and they are more likely to accept AI output without scrutiny.

One source put it bluntly: "AI allows you to get to a bad idea faster."

Dismissing this as a mere technical fault would badly understate the geopolitical risk. CNN noted the episode shows the immense danger of using a powerful, emerging, and not fully mastered technology directly for target identification and strike targeting in moments of heightened combat tension.

Military analysts have long feared this. If command systems rely on poor-quality or contaminated data, AI could trigger catastrophic miscalculations. It could even lead U.S. forces to fire directly on Chinese vessels based on false information.

Pay attention to what was on the other end. Throughout this affair, the designated target was a lawful Chinese vessel. The basis for that judgment was a report that nobody cross-verified from start to finish.

U.S. special operations forces stood ready. Warplanes roared skyward. The situation was a hair's breadth from a high-seas clash. And what stopped it was not an institutional intelligence review inside the U.S. military. It was one official's on-the-spot question before the operation launched.

Gaps, Sourcing, and a Warning at Sea

Other sources say AI is reaching deeper into U.S. military target locking. Yet Washington still has no clear rules for when human commanders should step in to prevent civilian casualties or friendly fire.

The result is a glaring contradiction: the United States insists it cannot fall behind in military AI integration, or it risks losing its edge in any future confrontation. But the data chains and verification mechanisms behind its operational decisions lack the most basic checks. That decision-making loophole, with its priorities reversed, is more worrying than any single algorithmic bias.

The CNN report rests on four anonymous insiders. Even the report itself admits it cannot verify what the cargo actually was.

No named official vouched for the story. The timing is vague, placed only "during the conflict with Iran." Specific details lack corroboration from public records.

When foreign media disclose military inside information by citing anonymous officials, the outside world must closely follow the strategic implications. But it must also examine the claims with an objective and prudent attitude.

One point is beyond doubt. When a country hands a military strike process over to a flawed system that lacks regulatory checks, the security risks will inevitably spill across borders. That process involves other nations' navigation rights and the safety of human lives.

This time a Chinese merchant ship was locked onto because of false intelligence. What will face the threat of miscalculation on the high seas next? The United States Armed Forces clearly cannot offer any guarantee.

The safety of navigation on international high seas and geopolitical stability must never be reduced to a risky bargaining chip under the vulnerabilities of AI algorithms.