SShortSingh.
TopTechnologyAI & MLStartupsProgrammingScienceBusinessFinanceCrypto & Web3World
Back to feed
Programming

AI Hardware Costs Force Amazon to Shorten Server Lifespans, Dragging Free Cash Flow Negative

0
·69 views

Amazon revised the estimated useful life of some servers downward in 2025, citing rapid advances in AI and machine learning, a move it acknowledged would reduce operating income by roughly $1.3 billion through accelerated depreciation and related charges. The decision stood out because it was made even as AWS reported $42.2 billion in quarterly revenue and an AI business exceeding $25 billion in annualized run-rate. Despite strong earnings, Amazon's trailing-twelve-month free cash flow turned negative at $7.6 billion, driven by a $66.1 billion surge in property and equipment purchases tied primarily to AI infrastructure. In contrast, Microsoft, Alphabet, and Meta had all recently extended their server useful-life estimates, each booking billions in reduced depreciation and higher reported profits. Amazon's reversal highlights a broader tension in the AI buildout: the same investments generating revenue are also consuming capital at a pace that strains conventional financial metrics.

Read the full story at DEV Community

This is an AI-generated summary. ShortSingh links to the original source for the complete article.

Discussion (0)

Log in to join the discussion and vote.

Log in

Related stories

0
ProgrammingDEV Community ·

Citrix Secure Access Client Flaw CVE-2026-118268 Allows Local SYSTEM Escalation; Fixed in 26.11.1.6

CVE-2026-118268 is a local privilege escalation vulnerability in the Citrix Secure Access Client for Windows that lets a standard user gain SYSTEM privileges. The defect is fixed in version 26.11.1.6 and later. Because the flaw affects endpoints rather than internet-facing services, internet search results such as ZoomEye's 0 application matches and 1 title match do not reliably indicate how many installations are exposed. Experts recommend assessing exposure through endpoint management inventory, checking installed versions, service states, and binaries, including on VDI images and devices that have not recently checked in. The remedy is upgrading endpoints, with follow-up steps like removing older binaries and rotating credentials if compromise is suspected.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

Gartner Predicts 90% of B2B Buying Will Be AI-Agent Intermediated by 2028

Gartner forecasts that by 2028, 90% of B2B purchasing will run through AI agents, moving more than $15 trillion of spend through machine-to-machine exchanges. Early indicators already show buyers ahead of sellers: 67% of B2B buyers now prefer a rep-free experience and 45% used AI during a recent purchase. Deloitte reports nearly 40% of buyers use agentic AI for tasks like product evaluation and contract review, while only 24% of suppliers use agents in sales. A 2021 example cited is Maersk, which used AI agents from Pactum to negotiate freight contracts, reportedly securing rates 22% lower in trials. The analysis argues that sellers still designing for human buyers need to adapt their commercial strategies.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

VOLT taxi service pre-launch tests focused on fares, commission splits and role limits

The team behind VOLT, a taxi service built for a small Ukrainian city, shared the checks they ran before launch. An overnight test of all money-related logic exposed that production still had old fares, which would have charged the first passenger 80 hryvnia instead of the agreed 130. Prices shown on the website now come from the same table the ride calculation uses, so the figures match everywhere. The commission split was verified across ten variants, including edge cases with kopecks, so payments add up exactly. The team also tested access boundaries separately, so service owners cannot change platform terms, adjust balances or see others' data.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

Mosaic: A Kotlin Framework for Decoupling BFF API Orchestration Logic

The article introduces Mosaic, a new Kotlin framework built to manage business logic in complex backend-for-frontend orchestration APIs. Developers write small, composable "tiles" that run in parallel, avoid duplicate calls, and declare dependencies on one another. Mosaic assembles a request tree dynamically for each request, aiming to replace sprawling conditional logic with testable, decoupled components. It is positioned as a complement to existing web frameworks such as Ktor, Micronaut, and Spring rather than a replacement. The authors point to examples and a quick start guide for developers looking to adopt it.

0 comments Read more at DEV Community