In an announcement, Amazon CEO Andy Jassy announced the company has over 1,000 generative AI services and applications in development, while acknowledging he expects the technology will reshape the company’s workforce. “We will need fewer people doing some of the jobs that are being done today, and more people doing other types of jobs,” he said. He added that it would be “hard to know exactly where this nets out over time,” but noted that the company expected genAI to reduce its total corporate workforce. The company employs about 1.5 million people internationally.
Jassy, along with Meta CEO Mark Zuckerberg, have been more forthright in spelling out the stakes than some of their fellow CEO peers, who have maintained careful public optimism even as they execute similar plans. “Probably in 2025, we at Meta…are going to have an AI that can effectively be a sort of mid-level engineer,” Zuckerberg said in a January 2025 Joe Rogan podcast. He eventually pushed out the timeline to mid-2026 at April’s LlamaCon conference.
The numbers back up such big claims that genAI had reached the point of doing a fair share of high-level coding work. Microsoft reports 20-30% of its code is now AI-generated, with CEO Satya Nadella noting the quality varies dramatically—”fantastic” for Python but “not that great” for C++. Google exceeded 30% by April 2025, up from 25% just six months earlier. Most aggressive is Anthropic’s CEO prediction: 90% of code will be AI-written within six months, essentially all within a year. Between 2024-2025, Amazon, Meta, Microsoft, and Google collectively eliminated over tens of thousands positions, including many engineering and technical roles, while committing nearly hundreds of billions to AI infrastructure. Nvidia’s Jensen Huang warns: “You’re not going to lose your job to an AI, but you’re going to lose your job to someone who uses AI.” Meanwhile, tech companies are shelling out millions to hire elite AI workers. While most AI positions have hundreds of thousands of dollars in salaries and perks, some make considerably more. Meta is offering some talented AI engineers $2 million or more annually — as are other tech companies.
Getting more done with scrappier teams
Jassy also implicitly spelled out a potential survival guide for employees, recommending that they “be curious about AI, educate yourself, attend workshops and take trainings, use and experiment with AI whenever you can.” He noted a need to learn how to “get more done with scrappier teams.”
Those who embrace the change and become conversant in AI, Jassy suggested, “will be well-positioned to have high impact and help us reinvent the company.”
The message comes amid Amazon’s ongoing workforce transformation. Since 2022, the company has eliminated over 27,000 positions across various departments, with reported plans to cut approximately 14,000 managerial roles by early 2025. Recent reductions include about 100 jobs in May 2025 within its devices and services division, affecting teams like Alexa, Echo, Ring and Zoox. Anthropic warns of worse: its CEO predicts AI could eliminate half of entry-level white-collar jobs and spike unemployment to 20% within five years.
Amazon is backing this AI push with massive investments. The company is projected to spend over $100 billion in capital expenditures for 2025, with much of it focused on AI infrastructure. Meta committed $60-65 billion for 2025 AI spending, while Google pledged $75 billion for AI-driven data centers. This includes developing proprietary AI chips like Trainium2, expanding AWS capabilities and building data centers globally. Amazon has also invested $8 billion in AI startup Anthropic, securing AWS as Anthropic’s primary cloud provider.
From Alexa+ to AI agents
Jassy’s memo detailed an AI offensive spanning Amazon’s empire. On the consumer front, the company is rolling out Alexa+, a next-generation assistant capable of taking “significant actions” beyond just answering questions. The company first announced Alexa more than a decade ago. Amazon also announced AI shopping tools like Lens (snap a photo to shop) and “Buy for Me” agents that can purchase items from other merchants’ websites. For businesses, Amazon is deploying AI across its AWS cloud platform with custom chips (Trainium2), foundation model services (Bedrock, SageMaker) and its own frontier AI model called Nova.
Behind the scenes, AI is already transforming operations, from robots in fulfillment centers to rebuilt customer service chatbots. But Jassy’s most ambitious vision centers on AI agents: autonomous software systems that can perform actions on behalf of users. Jassy said that the company has “strong conviction that AI agents will change how we all work and live.”
There’s an agent for that
These agents won’t just summarize information or automate basic tasks. According to Jassy, their capabilities will include researching topics in depth, writing code, detecting anomalies, surfacing non-obvious insights, translating languages, and streamlining a wide range of workflows.
Jassy predicts that AI agents will eventually number in the billions, permeating every industry and function. Jassy isn’t the only tech CEO who is enthusiastic about agents. Microsoft CEO Satya Nadella has suggested that traditional business applications may become obsolete in the agentic AI era, as AI agents take over backend business logic and orchestrate tasks across multiple platforms. Google’s Project Astra envisions a “universal AI assistant” using multimodal, context-aware capabilities to see and hear the world in real-time.
At Google, CEO Sundar Pichai has introduced Project Mariner, an early prototype of AI agents capable of interacting with the web to perform tasks on behalf of users. In parallel, Google’s DeepMind division recently unveiled AlphaEvolve, an AI agent that autonomously generates and optimizes algorithms, including surpassing decades-old benchmarks in matrix multiplication and improving Google infrastructure, by combining large language models with evolutionary search methods. Meta’s open-source strategy aims to commoditize AI models, challenging closed-source competitors while building a global developer community around its Llama platform.
