Amazon Web Services on Wednesday introduced Kiro Powers, a system that allows software developers to give their AI coding assistants instant, specialized expertise in specific tools and workflows — addressing what the company says is a fundamental hurdle in operating artificial intelligence agents today.
AWS made the announcement at its annual Re:Invent conference in Las Vegas. This capability is different from the way most AI coding tools work today. Typically, these tools preload every possible capability into memory – a process that burns through computational resources and can overwhelm the AI with irrelevant information. Kiro Powers takes the opposite approach, activating specialized knowledge only when a developer really needs it.
"Our goal is to give the agent specific context so that he can reach the right result faster – and thus reduce costs," Deepak Singh, vice president of developer agents and experiences at Amazon, said in an exclusive interview with VentureBeat.
The launch includes partnerships with nine technology companies: Datadog, Dynatrace, Figma, Neon, Netlify, Postman, Stripe, SupaBase, and AWS’s own services. Developers can also create their own powers and share them with the community.
Why do AI coding assistants shut down when developers connect too many tools?
To understand why KIRO powers matter, it helps to understand the growing tension in the AI development tools market.
Modern AI coding assistants rely on something called Model Context Protocol, or MCP, to connect to external devices and services. When a developer wants their AI assistant to work with Stripe for payments, Figma for design, and Supabase for databases, they connect MCP servers for each service.
The problem: Each connection loads dozens of tool definitions into the AI’s working memory before a single line of code can be written. According to AWS documentation, connecting just five MCP servers can consume more than 50,000 tokens — about 40 percent of the AI model’s context window — before the developer can type in his first request.
Developers have become increasingly vocal on the issue. Many people complain that they don’t want to waste their token allocation just to have an AI agent figure out which tools are relevant to a specific task. They want to get to their workflow instantly – not see an overloaded agent struggling to resolve irrelevant context.
This phenomenon, which some in the industry call "reference rot," Leads to slow responses, low-quality output, and significantly higher costs – as AI services typically charge by token.
Inside the technology that offloads AI expertise on demand
Kiro Powers solves this by packing three components into a single, dynamically loaded bundle.
The first component is a steering file called POWER.md, which serves as an onboarding manual for the AI agent. This tells the agent what tools are available and, importantly, when to use them. The second component is the MCP server configuration itself – the actual connection to external services. The third includes optional hooks and automations that trigger specific actions.
When a developer mentions "Payment" Or "check out" In their interaction with Kiro, the system automatically activates Stripe Power, loading its tools and best practices into context. When the developer transitions to database work, SupaBase becomes active while Stripe becomes inactive. The baseline reference used approaches zero when no power is activated.
"You click a button and it loads automatically," Singh said. "Once powered up, developers simply select ‘Open in Kiro’ and it launches the IDE with everything ready."
How AWS is bringing specialized developer technologies to the masses
Singh framed KIRO powers as the democratization of advanced development practices. Before this capability, only the most sophisticated developers knew how to properly configure their AI agents with particular context – writing custom steering files, crafting precise signals, and manually managing which devices were active at any given time.
"We found that our developers were adding capabilities to make their agents more unique," Singh said. "They wanted to give the agent some special powers to solve a specific problem. For example, they wanted their own front end developers, and they wanted agents to specialize in the backend as a service."
An important insight came from this observation: if SupaBase or Stripe could build an optimal reference configuration once, every developer using those services could benefit.
"Chiro Powers formalizes it – things that people, only the most advanced people, were doing – and allows anyone to have those kinds of skills," Singh said.
Why dynamic loading beats fine-tuning for most AI coding use cases
The announcement also positions Kiro Powers as a more affordable alternative to fine-tuning, the process of training AI models on specific data to improve their performance in specific domains.
"It’s very cheap," When asked how the powers compare to fine-tuning, Singh said. "Fine-tuning is very expensive, and you can’t fine-tune most marginal models."
This is an important thing. The most capable AI models are generally from Anthropic, OpenAI and Google. "closed source," Meaning developers cannot modify their underlying training. They can influence the behavior of models only through the signals and context they provide.
"Most people are already using powerful models like Sonnet 4.5 or Opus 4.5," Singh said. "Those models need to point in the right direction."
Dynamic loading mechanism also reduces running costs. Because powers are only activated when relevant, developers are not paying for token usage on tools they are not currently using.
Where Kiro powers fit into Amazon’s big bet on autonomous AI agents
Kiro Powers’ arrival comes as part of a broader effort by AWS in what the company calls "agentic ai" – Artificial intelligence systems that can operate autonomously for extended periods.
Earlier at re:Invent, AWS announced three "frontier agent" Designed to work for hours or days without human intervention: Kiro Autonomous Agent for software development, AWS Security Agent, and AWS DevOps Agent. These represent a different approach from KIRO powers – dealing with large, vague problems rather than providing specialized expertise for specific tasks.
Both approaches are complementary. Frontier agents handle complex, multi-day projects that require autonomous decision making across multiple codebases. In contrast, Kiro powers developers to provide precise, efficient tools for everyday development tasks where speed and token efficiency matter most.
The company is betting that developers need both ends of this spectrum to be productive.
What Kiro Powers tells us about the future of AI-assisted software development
This launch reflects a maturing market for AI development tools. GitHub Copilot, which Microsoft launched in 2021, introduced millions of developers to AI-assisted coding. Since then, a proliferation of tools including cursors, cline, and cloud code have competed for the attention of developers.
But as these devices have become more capable, they have also become more complex. The Model Context Protocol, which Anthropic open-sourced last year, created a standard for connecting AI agents to external services. This solved one problem while solving another: context overload which Kiro now addresses.
AWS is positioning itself as a company that understands large-scale production software development. Singh emphasized that Amazon’s experience running AWS for 20 years, coupled with its own vast internal software engineering organization, gives it unique insight into how developers actually work.
"This is not something you would just use for your prototype or your toy application," Singh said of AWS’s AI development tools. "If you want to build production applications, there’s a lot of knowledge that we bring to AWS that applies here."
The way forward for KIRO powers and cross-platform compatibility
AWS indicated that Kiro Powers currently only works within the Kiro IDE, but the company is working toward cross-compatibility with other AI development tools, including command-line interfaces, cursors, queues, and cloud code. The company’s document describes a future where developers can "Create electricity once, use it anywhere" – although that vision still remains aspirational.
For technology partners launching Powers today, the appeal is straightforward: Instead of maintaining separate integration documentation for every AI tool on the market, they can create a single Powers that works everywhere on Kiro. As more AI coding assistants come to market, that type of proficiency becomes increasingly valuable.
Kiro Powers is now available at no additional charge beyond the standard Kiro subscription for developers using Kiro IDE version 0.7 or later.
The underlying bet is a familiar one in the history of computing: that the winners in AI-assisted development will not be the tools that try to do everything at once, but the ones that are smart enough to know what to forget.
<a href