Via Middle East Eye

Saudi Arabia wants to replace Israel with Syria as the transit country for a fiber-optic cable designed to connect the kingdom to Greece through the Mediterranean Sea, two regional officials familiar with the project told Middle East Eye.

Saudi Arabia’s insistence that it be connected to Greece through Syria, and not Israel, as previously discussed, underscores how regional alignments are shifting as Riyadh looks to bolster Damascus’s standing in the region and potentially isolate Israel.

Saudi Arabian Crown Prince Mohammed bin Salman has publicly accused Israel of committing genocide in Gaza, where over 72,000 Palestinians have been killed. Riyadh is also at odds with the UAE, Israel’s closest Arab partner, in Yemen, Sudan, and the Red Sea.

Athens is trying to position itself as a hub between Europe and the Middle East for energy, real estate and Artificial Intelligence (AI).

Greece has courted Qatar, the UAE and Saudi Arabia for investment, but it is particularly close to Israel, which policymakers in Athens view as an ally against Turkey, and an insurance policy to keep the US engaged in the Eastern Mediterranean.

Saudi Arabia’s shift on the project could throw a wrench in Greece’s relationship with Israel, if it is indeed snubbed in the new route.

Fiber-optic cables carry essential digital services from country to country in milliseconds using pulses of light. Their importance is growing as Gulf states position themselves as exporters of AI, seeking to send data to Europe. 

Greece and Saudi Arabia announced the East to Med data Corridor, or the EMC project, in 2022. It is a joint partnership between Saudi Telecom (STC), the Greek electricity provider PPC, Greek telecoms and the satellite applications company, TTSA.

At the time, Saudi Arabia was in talks with the US on a deal that would see them normalize relations with Israel. Those negotiations were derailed by the Hamas-led 7 October 2023 attacks, which Israel retaliated against by launching an offensive on Gaza that the United Nations and human rights groups have deemed a genocide.

Israel also attacked Lebanon, Syria and Iran. “There were a number of projects that planned to go through Saudi Arabia, Jordan and Israel – this was one of them,” Julian Rawle, a US-based submarine fibre-optic cable consultant, told MEE.

“Saudi Arabia asking for transit through Syria is new. People are looking for additional terrestrial routes between the Indian Ocean and the Mediterranean. Syria is another option, if people feel comfortable with the evolving political situation there,” he added.

A presentation by Greece’s PPC dated November 2025, obtained by MEE, does not show Syria linked up to the EMC network. The corridor appears to move through Israel and its offshore waters. 

In addition, another regional official told MEE that Saudi Arabia envisions an electrical cable project with Greece bypassing Israel in favour of Syria. This project would link the Gulf state and Europe via a High Voltage Direct Current (HVDC) interconnection. 

‘Shift in attitude’

Saudi Arabia’s attempt to bring Syria into the projects underscores how it is using its wealth to bolster regional allies at a time when it is challenging the UAE and Israel in the region. It also hints at Riyadh’s broader vision for the region. 

“For Saudi Arabia, Damascus is at the heart of regional connectivity,” a western official familiar with Riyadh’s investment drive told MEE. “The Saudis want the roads, cables and trains to go through Syria”.  

Saudi Arabia’s STC announced in February that it will invest about $800m in Syria’s telecommunications infrastructure. The kingdom’s state news agency said the plan is to “connect Syria regionally and internationally through a fibre-optic network extending over more than 4,500 kilometres”.

Kristian Coates Ulrichsen, a Gulf expert at Rice University’s Baker Institute, said Saudi Arabia’s bid to include Syria at the expense of Israel reveals how much the region has reordered itself.

“A project like this is consistent with Saudi attempts to reintegrate Syria to the regional fold and play down any tangible links with Israel,” he added.

The year “2022 was the height of talk about normalization between Saudi Arabia-Israel normalisation. This is indicative of the shift in Riyadh’s attitude,” he told MEE.

Europe’s ports of entry shifting east

Greece envisions itself serving as a hub for multiple cable routes, as the Gulf states boost their investments in AI data centres and link up to East Asian business capitals such as Singapore.

Originally, the port cities of Marseille and Genoa were the embarkation points for fibre-optic cables arriving in Europe. But the industry wants to diversify routes, and ports of entry to Europe have been shifting further east, putting Greece and Turkey on the map.

The Eastern Mediterranean is littered with the corpses of grand infrastructure projects dreamed up by regional leaders and by Washington-based think tanks. A gas pipeline to connect Greece, Cyprus and Israel never materialized. Likewise, the Great Sea Interconnector cable, envisioned to link Greece, Cyprus and Israel, has faced multiple delays.

Turkey, which lays claim to a wide swath of the Eastern Mediterranean disputed by Greece, has opposed the projects. A trade corridor under discussion also aims to link India to Greece, Israel and the UAE.

But Rawle told MEE that the East to Med data Corridor, or EMC West, is one of the more viable projects. He said that making a down payment to the system supplier is a key hurdle that the industry watches as a marker for progress.

Greek and Saudi banks signed an agreement to finance 60 percent of the project. In 2023, EMC signed a supply contract with Alcatel Submarine Networks to construct two subsea and terrestrial data cables.

Iran’s black market for Starlink terminals has rapidly repriced, with reported street quotes of $4,000 per terminal, up from $700 to $1,000 last summer, as war risks surge and renewed fears grow over stricter internet censorship and another blackout.

Bloomberg spoke with Starlink terminal sellers and human rights groups that said satellite internet via Starlink terminals is one of the only ways to stay connected to the rest of the world, as the Iranian government has cut access or forced traffic onto a highly monitored national intranet amid social uprisings.

Now the terminals cost upwards of $4,000 per unit as war risks rise, and President Donald Trump has set a deadline for Iran to reach a “meaningful deal” on its nuclear program or face conflict with the US.

Last month, the Iranian government implemented a two-week full-scale internet blackout to quell unrest, and connectivity to the outside world remains heavily throttled. There are about 50,000 terminals in the country, according to digital rights group Holistic Resilience.

A Starlink seller on Telegram told the media outlet that units were extremely expensive in Iran, despite retailing in many parts of the world for a fraction of the price. In the US, Starlink terminals retail for several hundred dollars.

SpaceX waived Starlink subscription fees across Iran, and millions of Iranians have used U.S.-funded VPNs.

The Wall Street Journal reported last week that the Trump administration covertly sent 6,000 Starlink terminals into Iran.

🇺🇸🇮🇷 US smuggled 6000 Starlink terminals into Iran last month in an effort to keep protesters online after Iran cut internet access – WSJ pic.twitter.com/8cRoa8oCEc

— Lord Bebo (@MyLordBebo) February 12, 2026

Anyone using or distributing a Starlink terminal in Iran risks a lengthy prison sentence.

Authored by Jeff Dornik via American Greatness,

We were told this time would be different. We were told that a second Trump administration would not repeat the mistakes of the first, that hard lessons had been learned, and that the Deep State would finally be confronted rather than tolerated. One year into President Trump’s second term, it is both fair and necessary to ask whether those assurances are being honored—not from hostility but from a sincere desire to see the America First agenda succeed, endure, and become irreversible.

During President Trump’s first term, Congress squandered its moment. The first two years were consumed by infighting, hesitation, and internal paralysis, even with Republican control. Then came the midterms, control was lost, and meaningful legislative progress effectively ended. What followed were impeachment spectacles and relentless political warfare, while entrenched corruption inside the federal government remained untouched. Now, just past the first year of President Trump’s second term, the pattern feels disturbingly familiar. The urgency voters demanded is not being matched by the actions of those entrusted to deliver it.

The question that must be asked plainly is this: when is the Trump administration actually going to root out the Deep State?

Executive Orders are being signed at a rapid pace, but Executive Orders are not reform. They are temporary directives that can be erased with a single signature the moment someone like Gavin Newsom takes office. Without legislation, without prosecutions, and without accountability, nothing is secured. Power is being exercised, but it is not being anchored, and lasting change is never achieved that way.

Kash Patel built his credibility by telling the truth about corruption in Washington. His book and documentary, Government Gangsters, documented in detail how entrenched bureaucrats and intelligence officials worked against President Trump from within the federal government. He even came on my show and spoke openly about this corruption, and he stated repeatedly across multiple platforms that the FBI, particularly at its highest levels, was deeply compromised and required fundamental reform. He did not argue that the Bureau should be abandoned, but that it could not be trusted without aggressive leadership, restructuring, and accountability for the Deep State operatives within the bureau. He warned that the Deep State would never reform itself and would have to be confronted directly. He also told Glenn Beck that the head of the FBI possessed Jeffrey Epstein’s client list. These were not casual remarks. They were core assertions made publicly and repeatedly.

Now Kash Patel is the head of the FBI, and the public posture has shifted dramatically. The same institution he once described as captured is now treated as credible and restrained. The Epstein client list, once discussed as a known reality, is now dismissed as conspiracy, even as new Epstein-related documents continue to be released to the public over the protest of the Trump administration. Each document release raises more questions, not fewer, and every delay from federal law enforcement deepens public distrust rather than restoring confidence. A reversal this significant demands explanation. Trust is not rebuilt through silence, and credibility is not preserved by pretending prior statements were never made.

These questions extend far beyond the FBI and land squarely on the Department of Justice, where accountability appears to collapse the moment it threatens entrenched power. The removal of Ed Martin from his role inside the DOJ is not just a minor personnel decision; it appears to be a clear signal that real investigations into weaponization and lawfare are not being tolerated. Ed Martin was positioned to expose how the Biden Department of Justice targeted Americans, abused prosecutorial authority, and used federal power as a political weapon. According to Emerald Robinson, whose reporting has repeatedly exposed corruption others refuse to confront, Martin was removed from his position by the same people who refer to parents as terrorists: “Vance Day, senior counsel for Todd Blanche, refers to parents targeted by Biden DOJ as ‘terrorists’ in recent meeting with one parent asking for accountability. Blanche’s office also removed Ed Martin from his role at the DOJ.” That disclosure alone should alarm every American paying attention.

Parents who were targeted and persecuted by the Biden Department of Justice are now being labeled terrorists by senior DOJ leadership, while the man tasked with investigating that persecution is sidelined. Whether this is described as a firing or a demotion is irrelevant, because the outcome is the same. Another one of the good guys has been removed from doing the work voters were promised would finally drain the swamp. This is not an isolated incident or a misunderstanding but a pattern that repeats with disturbing consistency. Every time someone begins making real progress against the Deep State, authority is stripped, investigations are stalled, and momentum is deliberately crushed before accountability can be delivered.

So the questions must be asked:

Where are the arrests?

Where are the prosecutions?

Why has Attorney General Pam Bondi not brought cases against members of the January 6 Committee despite documented misconduct and destroyed records?

Why has the Department of Justice taken no action against Anthony Fauci even after Sen. Rand Paul issued criminal referrals? 

Why is the DOJ actively fighting to shut down Brook Jackson’s case against Pfizer instead of allowing it to proceed and standing with a whistleblower who exposed documented fraud?

Why do Epstein-related documents continue to surface while no meaningful accountability follows?

What happened to transparency, and what happened to equal justice under the law?

Congress bears equal responsibility for this failure. DOGE exposed widespread corruption, fraud, and waste throughout the federal government, yet Republican spending bills continue to fund the very programs DOGE identified. If fraud has been uncovered, there is no justification for continuing to finance it. Where is the legislation to codify President Trump’s Executive Orders so they cannot be casually undone by the next administration? Where is the structural reform that fixes what is broken instead of managing its decay?

It increasingly appears that Republican leadership is content to run out the clock while President Trump absorbs the political risk. That approach is not strategic. It is reckless. We are less than one year away from the midterms, and the possibility of losing congressional control is very real. If that happens before reforms are locked into law, the opportunity may be lost entirely.

So what is the agenda for this year? What are we fighting to accomplish while the window is still open? America First cannot remain a slogan. It must become law, policy, and precedent. And by America, I mean Americans first, not institutions, not bureaucracies, and not federal agencies that operate without consequence.

The stakes are higher than many are willing to acknowledge. A future Democrat administration, particularly one led by Gavin Newsom, would inherit not only an unrestrained administrative state but a vastly expanded and centralized federal system increasingly powered by artificial intelligence. The Trump administration is already integrating AI throughout government. Without moral clarity, legal restraint, and decentralization, that power will inevitably be turned against the people it claims to serve.

This is not an attack on President Trump. It is a call to fulfill the mandate voters gave him. The American people did not vote for symbolism. We voted for accountability. We voted for justice. We voted for a government that serves the people rather than ruling over them.

Time is running out. If decisive action is not taken this year, there is a real chance it never will be. If that happens, no one should be surprised when everything collapses the moment power changes hands. Truth demands courage, and courage demands action.

The question is no longer whether the Deep State exists, but whether those in power are willing to confront it while they still can.

Views expressed in this article are opinions of the author and do not necessarily reflect the views of ZeroHedge.

Margaret Thatcher once said, “The trouble with Socialism is that eventually you run out of other people’s money,” and New York City’s new socialist mayor, Zohran Mamdani, is learning just how right she was, and New Yorkers are going to pay a hefty price for it.

On Tuesday, a mere two months after declaring he would “replace the frigidity of rugged individualism with the warmth of collectivism,” Mamdani announced a $127 billion preliminary budget for fiscal year 2027, a $5 billion increase from the prior year, while simultaneously warning residents of “painful” tax hikes if state officials refused to bail him out to cover his socialist policies. 

“That’s a city budget bigger than the state budgets of 47 states. Even the state government of Florida (population 23 million) spends less than New York City’s,” explains The Washington Post editorial board. “And the state still managed to attract hundreds of thousands of New Yorkers in recent years.”

“The reality is that The reality is that Americans may like the idea of ‘free’ stuff — it’s how socialists win elections — but they are less excited about having to pay for it” they continued. “They’re even less excited when they live in a state that ranks at the very bottom of the Tax Foundation’s State Tax Competitiveness Index.”

During a press conference earlier this week, Mamdani called on New York Gov. Kathy Hochul to raise income taxes on the “ultra-wealthy” help fund his budget for New York City.

“The onus for resolving this crisis should not be placed on the backs of working and middle-class New Yorkers,” Mamdani said. “If we do not fix this structural imbalance and do not heed the calls of New Yorkers to raise taxes on the wealthy, this crisis will not disappear. It will simply return, year after year, forcing harder and harsher choices each time. And if we do not go down the first path, the city will be forced down a second, more harmful path. Faced with no other choice, the city would have to exercise the only revenue lever fully within our own control. We would have to raise property taxes.”

Mamdani says if Hochul doesn’t allow NYC to raise taxes on corporations and highest earners, he will raise property taxes on New York City homeowners. pic.twitter.com/Bm07LxS9OH

— The Post Millennial (@TPostMillennial) February 17, 2026

Hochul rejected the tax hike demand without hesitation, telling Mamdani to expand his “ridiculously low” proposed spending cuts instead. 

Mamdani has claimed his administration identified $1.7 billion in cuts. The Post’s editorial board was not impressed, calling it a “laughable number.”

“The reality is that Mamdani is trying to expand a city government that already does way too much,” they argued. “ The city should provide basic services, such as law and order, but instead it pours billions into social spending like housing and health care.”

They even cited California as a cautionary tale, warning that in the Golden State, “a slew of billionaires are fleeing at the mere possibility of a wealth tax. They’ll avoid the wealth tax — and California will miss out on the billions that these individuals otherwise would have contributed before a wealth tax was even imposed.”

More experienced Democrats in New York understand this. Gov. Kathy Hochul, no one’s idea of a fiscal hawk, nevertheless instigated Mamdani’s tantrum by refusing to go along with more tax hikes. The city council speaker and comptroller also have sway and are skeptical of new taxes.

This week, it was revealed that acclaimed director and filmmaker Steven Spielberg officially became a New York resident on January 1, effectively avoiding the billionaire tax—though a representative for Spielberg and his wife Cate Capshaw claimed the move was to be closer to family.

Mamdani’s pre-election promises — free buses, expanded child care, cash assistance, rental aid, and smaller class sizes for teachers’ unions — were crowd-pleasers that earned him “tax the rich” chants at campaign rallies. The problem is that governing a city with a structural deficit requires something more than slogans. His preliminary budget now acknowledges a $5.4 billion shortfall for the current fiscal year, with projections that worsen over time. 

“No one in New York is ambitious enough to dramatically reshape city government, and residents either vote for class warfare or vote with their feet. A reckoning will have to come eventually. The question is how bad it gets before reality sets in,” the board concluded.

Ouch.

Via The Cradle

Saudi crude shipments to India are set to reach their highest level since 2020 this month, narrowing the gap with Russian supplies as New Delhi faces ongoing US pressure to reduce its purchases of Russian oil, according to data from Kpler cited by Bloomberg on Friday.

Flows from Saudi Arabia are projected at between 1 million and 1.1 million barrels per day (bpd), Kpler lead research analyst Sumit Ritolia said, describing the volumes as the strongest since November 2019, bringing Saudi exports broadly in line with Russian deliveries.

Kpler estimates Russian crude shipments to India at around 1.2 million bpd this month, which would keep Moscow as India’s largest supplier, though at levels well below the more than 2 million bpd seen at peak periods over the past three years.

For March, Kpler forecasts Russian flows easing further to between 800,000 and 1 million bpd, noting that such a drop could allow Saudi Arabia to retake the top spot if Indian refiners keep replacing Russian crude with Saudi shipments.

In early February, US President Donald Trump announced a trade deal with India that linked tariff reductions to India halting purchases of Russian crude oil. 

He said he would cut punitive tariffs on Indian goods in exchange for Prime Minister Narendra Modi’s agreement that India would stop buying Russian oil. He even suggested New Delhi would increase purchases of US – and potentially Venezuelan – oil.

Current import levels indicate that Indian buyers have not moved abruptly to cut Russian intake.

India emerged as a major buyer of Russian crude after the 2022 invasion of Ukraine, when discounted barrels diverted from Europe were rerouted to Asian markets – meaning that any loss of market share in India would shrink one of Moscow’s main export outlets.

Meanwhile, as Indian spot purchases soften, China’s imports of Russian oil are on track for a record month. 

Russian share of India’s January oil imports lowest since late 2022, data shows
https://t.co/w6wClXRX1Z

— Economic Times (@EconomicTimes) February 18, 2026

Data from Vortexa and Kpler, cited by Reuters, shows Chinese inflows at roughly 2.07–2.08 million bpd in February, underscoring a shift in trade patterns rather than a collapse in overall Russian exports.

Another ‘win’ for America’s disastrous Syria policy, long predicated on overthrowing the Assad government and installing a ‘moderate’ Sunni regime – though it turns out Jolani’s bearded Hayat Tahrir al-Sham (HTS) militants are anything but…

“U.S. intelligence agencies have concluded that 15,000 to 20,000 people, including Islamic State affiliates are now at large in Syria, after an exodus from a camp that held jihadists’ families, U.S. officials familiar with the estimate said,” The Wall Street Journal reports Friday.

Who could have predicted that chaos, instability, and terrorism would come out of the CIA’s Operation Timber Sycamore? Well, we did, and every rational observer of the Syria situation.

A billion plus dollars and hundreds of thousands of lives after the decade-long proxy war, and this is all Washington has to show for it:  

Security experts have long warned that the wives of Islamic State fighters were effectively raising the next generation of militants at the sprawling Al-Hol facility. Security at the camp fell apart in recent weeks after Syria’s government routed the U.S.-backed Syrian Democratic Forces, which had guarded Al-Hol for years, raising concerns about the release of people who might have become radicalized during the years held behind the razor wire.

The size of a small city, the camp in Syria’s eastern desert at one point held more than 70,000 people after U.S.-backed forces destroyed what remained of Islamic State’s self-proclaimed caliphate in Syria in 2019. At the end of 2025, more than 23,000 people were there, according to a report this week from the Pentagon’s Inspector General.

The US military is rapidly backing out of this region after the years-long occupation, effectively throwing the Kurds (SDF) under the bus, as HTS radicals move in and take control. 

Given many analysts have pointed to HTS being ‘ISIS-lite’ to begin with, the following WSJ note is no surprise: “The vast majority have left Al-Hol after the Syrian government took control last month. Western diplomats in Damascus assessed that more than 20,000 people fled the camp in a matter of days earlier amid rioting and a surge of escape attempts.”

There were even reports that the ISIS prisoners greeted the government HTS troops rolling in as ‘liberators’. The new government is certainly not “fighting” Islamic State cells… quite the contrary:

As the Asayish is set to re-take the control of the Al hol camp and region (including everything north of it) – the Syrian Government Forces currently in charge of Al Hol Camp have opened the gates of the camp and are releasing+transporting as many ISIS members and families out… pic.twitter.com/2jQOR4hisI

— ScharoMaroof (@ScharoMaroof) February 13, 2026

And now the Washington blob is simply moving on to the next regime change operation, this time a little further east in Iran, which it turns out was a key Assad ally.

So in place of the secular nationalist Baath party (under the Assad family), the West has the literal founder of Syrian al-Qaeda as president of Damascus, letting ISIS prisoners and affiliates walk free.

Authored by Zachary Stieber via The Epoch Times (emphasis ours),

The Food and Drug Administration (FDA) will approve many new drugs based on one trial moving forward, agency leaders have said.

The FDA has typically required two studies from companies seeking approval for most new drugs, although in recent years it has approved some drugs based on a single well-run trial.

“The FDA has demonstrated disease-by-disease flexibility and has granted approvals based on a single premarket study with confirmatory evidence. In some fields, such as oncology, single trials have supported the majority of drug approvals,” Dr. Marty Makary, the FDA’s commissioner, and Dr. Vinay Prasad, head of the FDA’s Center for Biologics Evaluation and Research, said in an article published on Feb. 18 by the New England Journal of Medicine.

“However, although we have exercised flexibility in the past, there remains confusion from manufacturers regarding settings in which a single trial will be accepted. Moving forward, we are announcing that a one-trial requirement will be the FDA’s new default standard. This reform is being rolled out synchronously with the agency’s postmarket initiative to collect robust data on all drugs and devices.”

The two-study standard for drugs dates to the early 1960s, when Congress passed a law requiring the FDA to review data from “adequate and well-controlled investigations” before clearing new medications. For decades, the agency interpreted that requirement as meaning at least two studies, preferably with a large number of patients and significant follow-up time.

The second study would, in theory, confirm that the first trial’s results weren’t a fluke and could be reproduced.

Beginning in the 1990s, the FDA increasingly began accepting single studies for the approval of treatments for rare or fatal diseases that companies often struggle to test in large numbers of patients. Over the past five years, roughly 60 percent of first-of-a-kind drugs approved each year have been cleared based on a single study.

Makary and Prasad said that the historical reliance on multiple studies “was intended to provide credible causal evidence that a therapy could improve clinical outcomes with acceptable safety in a world where biologic understanding was more limited than it is today.”

They later added: “In the modern world, as drug discovery becomes increasingly precise and scientific, the FDA considers not just effects on survival, but biochemical and intermediate changes that tell a complete biologic story: does this drug actually work? In this setting, overreliance on two trials no longer makes sense.”

The change will save drug developers money and reduce the time it takes to get drugs to market, the officials said. They expect more drug development in response.

Dr. Janet Woodcock, the FDA director who led the agency’s drug center for about 20 years before retiring in 2024, said the change makes sense and reflects the FDA’s decades-long move toward relying on one trial, combined with supporting evidence, for various life-threatening diseases, including cancer.

“The scientific point is well taken that as we move toward greater understanding of biology and disease we don’t need to do two trials all the time,” Woodcock said.

Dr. Reshma Ramachandran, assistant professor of medicine at the Yale School of Medicine, said in a post on X that it’s true most FDA approvals in recent years have been based on single, strong trials.

“But as the authors noted (& have for years!), patients are increasingly left with uncertainty of their effectiveness,” she wrote, “so why set a standard continuing the (bad) same old instead of demanding more?”

Makary and Prasad said that they reserve the right to demand additional testing if a trial has limitations or deficiencies.

“Instead of prioritizing finite reviewer time reading and assessing two or more pivotal trials, we will focus our energies on ensuring that the one clinical trial we require provides the most up-to-date and useful information for American patients,” they wrote.

The Associated Press contributed to this report.

Amazon’s cloud-computing arm suffered at least two recent service interruptions linked to the use of its own artificial intelligence coding assistants, prompting some internal concerns about the company’s rapid deployment of autonomous software agents inside production environments.

In mid-December, Amazon engineers allowed the company’s Kiro AI coding tool to implement system changes that ultimately led to a roughly 13-hour disruption affecting one of the systems customers use to analyze the cost of AWS services, people familiar with the matter told the Financial Times. 

The agentic tool – which is capable of taking autonomous actions on behalf of users – reportedly determined that the optimal remediation step was to delete and recreate a computing environment. AWS later circulated an internal postmortem examining the outage.

Employees said the December incident marked the second time in recent months that one of Amazon’s internally deployed AI development tools had played a central role in a service disruption. In both cases, engineers permitted the software agent to execute changes without requiring secondary approval, a safeguard typically mandated for manual interventions in production systems.

AWS accounts for roughly 60% of Amazon’s operating profit and is investing heavily in artificial intelligence tools designed to function as independent “agents” capable of carrying out tasks based on high-level human instructions. The company – along with other large technology firms – is also positioning such tools for sale to external enterprise customers.

Amazon said it was a coincidence that AI tools were involved in the disruptions and maintained that the same outcome could have resulted from conventional development software or manual intervention.

“In both instances, this was user error, not AI error,” the company said, adding that it had found no evidence that mistakes occur more frequently when AI tools are involved.

The company described the December interruption as an “extremely limited event” affecting a single service in parts of mainland China and said the second incident did not impact a customer-facing AWS system.

Neither disruption approached the scale of a broader AWS outage in October 2025 that lasted approximately 15 hours and temporarily took multiple customers’ applications offline – including services operated by OpenAI.

Employees said the company’s AI development tools are often treated as operational extensions of human engineers and are granted comparable system permissions. In the December case, the engineer involved had broader access than anticipated – a user access-control issue that Amazon said allowed the changes to proceed without appropriate review.

AWS introduced Kiro in July as a next-generation coding assistant designed to go beyond so-called “vibe coding,” in which developers rapidly assemble applications using AI-generated suggestions. Instead, Kiro was intended to produce code directly from formal specifications.

Prior to Kiro’s launch, AWS engineers relied on Amazon Q Developer, an AI-powered chatbot designed to assist with software development. Employees said that tool was involved in an earlier outage.

Some staff members said they remain skeptical about the reliability of AI-assisted coding for mission-critical tasks, particularly as Amazon has set internal targets encouraging 80% of developers to use AI tools for coding at least once per week. The company is said to be closely monitoring adoption rates.

Amazon said customer uptake of Kiro has been strong and that it wants both clients and employees to benefit from efficiency gains. Following the December incident, AWS implemented additional safeguards, including mandatory peer review procedures and expanded staff training.

Authored by Autumn Spredemann via The Epoch Times (emphasis ours),

Artificial intelligence (AI) is often framed as a force multiplier that can accelerate decision-making and produce valuable information. Meanwhile, AI deployment exercises have yielded mixed results, highlighting challenges such as systems stalling and unpredictable software outside controlled environments.

Some defense insiders believe that AI tools also introduce new safety and escalation risks if not developed, evaluated, and trained correctly.

Over the past year, U.S. military testing has demonstrated that some AI systems are failing in the field. In May 2025, Anduril Industries worked with the U.S. Navy on the launch of 30 AI drone boats, all of which ended up stuck idling in the water after the systems rejected their inputs.

A similar setback occurred in August 2025 during the company’s test of its Anvil counterdrone system. The resultant mechanical failure caused a 22-acre fire in Oregon, according to a Wall Street Journal report.

Anduril responded to the reported AI test failures, calling them “a small handful of alleged setbacks at government experimentation, testing, and integration events.”

“Modern defense technology emerges through relentless testing, rapid iteration, and disciplined risk-taking,” Anduril stated on its website. “Systems break. Software crashes. Hardware fails under stress. Finding these failures in controlled environments is the entire point.”

But some say the challenges AI faces in the national security landscape should not be taken lightly. Problems such as brittle AI models and building on the wrong kind of training data can create systems that do not perform as expected in a battlefield scenario.

“This is why military-grade AI, purpose-built for national security use cases and the warfighter, is critical,” Tyler Saltsman, founder of EdgeRunner AI, told The Epoch Times.

Saltsman’s company has active research and development contracts with the U.S. military. He said AI systems are not typically designed for warfighting.

“[AI models] may choose to refuse or deflect certain questions or tasks if those requests do not comply with the AI system’s own rules,” Saltsman said. “A model refusing to provide guidance to a soldier in combat or giving biased responses rather than operationally relevant responses can have life-or-death implications.”

Scenarios such as the one Saltsman described can start with the wrong kind of training data.

Data Dilemma

Jeff Stollman, who has worked with defense contractors as an independent consultant and is familiar with a range of products and services used by the military and intelligence communities, said much of “the data needed has not been collected historically.”

“And because internet data is typically of limited value and internet-based models can’t be run on isolated classified networks, military and intelligence users will need to collect their own new data,” Stollman told The Epoch Times.

He said there are three categories of training data used by the defense and armed forces communities, all of which have different hurdles.

Offering an example of a sustainment—or maintenance—data challenge, Stollman said that collecting this type of information typically requires adding sensors that can record the data needed to predict malfunctions and failures.

“This includes measuring temperature, vibration, friction, the amount of wear on various parts,” he said. “This is an expensive undertaking. Sensors aren’t free. They add weight and volume to space and weight-constrained platforms such as aircraft and spacecraft.”

This type of data collection is offloaded to a database because of limited onboard computer resources. Although that sounds logical at first, the problem is the time it can take.

“For platforms like ships and submarines, windows for transmission of such data, which might give away the position of the platform, are limited,” Stollman said. “As a result, data may not be accessible for months at a time.”

Another challenge of AI integration is reliability. Issues such as AI “hallucinations” and poor decisions can be amplified in adversarial environments.

“The most dangerous assumption is that AI can distinguish between legitimate inputs and adversarial manipulation,” Christopher Trocola, founder of ARC Defense Systems, told The Epoch Times.

He cited the July 2025 experiment in which AI-powered, cloud-based platform Replit’s “vibe coding” ended with an AI assistant panicking and trying to cover its tracks. The AI coding assistant reportedly deleted a live production database, fabricated thousands of fake records, and created misleading status messages.

“Military applications amplify these vulnerabilities catastrophically,” Trocola said.

He explained that three critical AI assumptions can fail under adversarial pressure: prompt injection resistance, hallucination control, and intent recognition.

This is when adversaries can manipulate AI through carefully crafted inputs designed to override instructions, generate false information, or indicate that malicious inputs are benign.

“This represents what’s known as distribution shift: AI trained in controlled environments failing catastrophically when deployed in real-world adversarial contexts,” Trocola said.

Saltsman said this highlights the importance of building AI models with military applications in mind.

“Most commercial AI systems are black boxes,” he said. “We don’t know what data trained the models. We don’t know what guardrails or biases were baked into the models. And we don’t know if our data is truly secure. All of this is highly problematic in national security settings.”

Risk Evaluation

Stollman noted that generative AI—which is already used in U.S. intelligence and defense—is “plagued” with problems such as hallucinations. However, it is also the most practical kind of AI for military operations.

“Generative AI is useful in areas such as reconnaissance, where it is necessary to identify installations and activities from data collected by various sensors: photos, radar, sonar, etc.,” Stollman said. “It can also be used to support decision-making.”

“For example, drones or missiles could be given autonomy of action to overcome signal jamming that prevents their being controlled remotely by humans,” he said. “But before such autonomy can be deployed, it is necessary to anticipate all the failure modes that could lead to undesirable consequences.”

Saltsman said he agrees that AI development and deployment must be carefully balanced with long-term risk evaluation.

“But make no mistake, we are in an AI war against China, and we must win the race,” he said.

He noted that if China’s AI models and hardware dominate the market, the United States could become dependent on the Asian nation for critical technologies.

“Therefore, it is a national security imperative that we accelerate the pace of AI development while also balancing the risks,” Saltsman said.

In 2025, the United Nations said that the use of AI in warfighting was no longer a hypothetical future scenario. The U.N. also stressed the risks and consequences of AI system failures in this capacity.

“Without rigorous safeguards, it risks undermining international humanitarian law,” the agency stated.

“Complex battlefields already test human judgment in distinguishing between combatants and civilians; for machines, the challenge is even greater, particularly in urban settings where civilians and fighters often intermingle.”

Trocola said he shares concerns that AI deployment in the military and defense sectors is outpacing risk assessment.

“Documented patterns suggest this creates systematic vulnerabilities,” he said. “Industry data shows [70 percent to 80 percent] of AI projects fail due to organizational readiness gaps.”

The Department of War AI Acceleration Strategy launched in January, which emphasizes rapid deployment to counter strategic competitors.

Read the rest here…