Amazon Tightens Coding Guardrails After AI Tools Linked To Major Outages: Report

March 11, 2026, 10:47 pm

Amazon AI coding outages have been a concern for Amazon, and there has been a series of outages for their e-commerce and cloud services. Apparently, this has been caused by the coding assistants used by artificial intelligence. After this, Amazon is expected to implement new measures and documentation regarding their coding process. This follows a series of outages experienced by Amazon in the third quarter of 2025 and significant outages that took place in early March 2026, resulting in massive order losses due to the company’s failure to meet user demand.

Amazon Moves to Strengthen Coding Safeguards

Amazon is planning to tighten its internal guardrails for its coding process after facing a series of outages in its e-commerce business in recent times. Amazon is of the view that it can minimize disruptions in its e-commerce business by tightening its internal guardrails for its coding process.

Dave Treadwell, who is Amazon’s Senior Vice President of e-commerce services, talked to its employees at an internal meeting that was held on March 10. According to reports, Dave told his employees that Amazon would implement temporary safety practices for its retail business.

He claimed that this would introduce a “controlled friction” into the coding process. Such practices will slow down risky changes in the most important areas of the retail experience. At the same time, Amazon is also committed to long-term solutions that include both deterministic and agentic safety measures.

Internal Meeting Highlights Rising Incidents

Amazon arranged the meeting to discuss the rising “trend of incidents,” which has emerged over the third quarter of 2025. Some internal notes prior to the meeting pointed to several significant outages over the last few weeks.

The meeting was about the reasons for these outages and what should be done about them. Amazon executives think that the increasing incidents necessitate immediate policy changes.

Meanwhile, the meeting, according to an Amazon spokesperson, was just a weekly review. The meeting, the company explained, included the review of the availability of its website and app, which are part of the normal course of the company’s operations. The company also explained that only one of the incidents reviewed during the meeting involved artificial intelligence.

Stricter Code Reviews and Documentation

Furthermore, Amazon is also introducing more stringent processes for its engineers. Engineers will be required to document the changes they are making to the code more extensively. Engineers will also be required to obtain more approvals before they make the necessary updates.

Under the new policy, the engineers will be required to obtain reviews from at least two other colleagues before they are allowed to push the code into the production system. Engineers will also be required to use the company’s documentation and approval tool.

Moreover, Amazon will use its automated coding system that adheres to the company’s key reliability engineering principles. These steps will help avoid errors that may impact millions of customers.

March 2026 Outages Trigger Concern

The recent change in policy is a result of several disruptions that took place in Amazon’s retail platform in the past days.

On March 2, 2026, users experienced incorrect delivery time messages when adding items to their cart on Amazon’s website and mobile app. They also experienced challenges in completing their purchases. Some users were unable to view certain features such as account information and price of items.

The outages lasted for almost six hours. Within this period, Amazon reportedly lost around 120,000 orders. The internal system also experienced around 1.6 million website error messages.

Amazon’s internal review reportedly found that the AI coding tool known as Q contributed to the outages.

Another Outage Caused Massive Order Loss

Another major disruption happened just a few days later. On March 5, 2026, there was another outage of Amazon’s marketplace, which resulted in a dramatic decline in orders in the US.

According to reports, there was a decline of around 99 percent in orders in the US. As a result of this, Amazon lost around 6.3 million orders.

Later, it was found that engineers made a change to production without approval, which resulted in the outage and a decline in orders.

AWS Disruptions Linked to AI Coding Tools

The cloud computing segment of Amazon has also experienced similar problems. It has been reported that at least two cloud computing outages experienced by the Amazon Web Services segment were caused by the use of AI coding assistants.

One of the major outages faced by the company happened in December 2025 and lasted for around 13 hours. The cost calculator of AWS clients of the company was suspended for a short period of time.

The problem had been caused by alterations to the system that had been implemented by engineers using an AI coding assistant tool provided by the company, called Kiro. The tool had decided to “delete and recreate the environment.”

AI Coding Creates New Challenges for Tech Companies

The incidents reveal an emerging trend of tension within the technology sector. There is an increase in the use of AI programs to produce software codes. Yet, many firms still do not have effective systems to review and verify vast amounts of AI-produced software codes.

Today, software engineers face pressure to create software quickly with the help of artificial intelligence. Yet, review systems cannot keep up with the rapid growth of automated coding.

This is happening at a time when there is a large reduction of workforce in the technology sector. Technology firms still lay off large numbers of employees while using AI programs to cut costs.

Layoffs Add Pressure on Engineering Teams

The online retailer announced in January 2026 that it would cut more than 16,000 corporate jobs. The company has said that the layoffs are not directly related to the recent service issues.

According to reports, some of the engineers who work in the online retailer claimed that they are currently handling a number of critical incidents compared to before. These critical incidents are referred to as “Sev2s” and require to be resolved in a timely fashion to prevent failures in a number of products.

Tier-1 Systems Now Under Audit

Another initiative taken by Amazon is to conduct a thorough review process to prevent such outages from happening again. The company has asked the owners of 335 Tier-1 systems to audit all the changes made to the production codes.

Tier-1 systems are those critical infrastructure pieces that directly impact the consumers. So, Amazon wants these systems to be monitored by the directors and vice-presidents.

Through these measures, Amazon is trying to mitigate the risks associated with rapid AI-driven software development.

Today Tech Industry News | Latest Tech Industry News | Breaking Tech Industry News in English | Latest Tech Industry News Headlines -