cd /news/ai-safety/anthropic-employees-say-ai-could-kil… · home topics ai-safety article
[ARTICLE · art-125850] src=247wallst.com ↗ pub= topic=ai-safety verified=true sentiment=↓ negative

Anthropic Employees Say ‘AI Could Kill All Humans!’ Put Odds Above 10% This Decade

Anthropic researcher Evan Hubinger publicly estimated on X that AI could kill all humans with odds of ">10% within the next decade," conceding the lab "do[es] not yet have a plan to solve alignment for superintelligence." The estimate lands as Amazon, which CEO Andy Jassy said has committed Anthropic to up to 5 GW of Trainium capacity anchoring Project Rainier's 500,000+ Trainium2 chips, reported a $53.4 billion one-time pre-tax gain on its Anthropic stake in Q2 FY2026 and AWS revenue growth of 36.7% year-over-year on a $169 billion annualized run rate. Jassy told analysts Amazon is "pursuing our own frontier model" while Anthropic and OpenAI hold multi-year, multi-gigawatt Trainium commitments.

by read3 min views2 publishedSep 10, 2026
Anthropic Employees Say ‘AI Could Kill All Humans!’ Put Odds Above 10% This Decade
Image: 247Wallst (auto-discovered)

An Anthropic researcher just put the odds of AI killing all humans above 10% this decade, and Amazon has staked roughly $200 billion in annual capex on that very lab. What Andy Jassy told investors next will either reassure shareholders…

The company whose chips Amazon has bet roughly $200 billion in 2026 capex to serve just had one of its own staff publicly estimate the odds of human extinction from AI at more than one in ten this decade. On X, Anthropic researcher Evan Hubinger wrote that “we really do earnestly believe AI could kill all humans” and pegged the probability at “>10% within the next decade.” He also conceded Anthropic “do[es] not yet have a plan to solve alignment for superintelligence and [is] not clearly on track to.”

That is the counterparty Amazon (NASDAQ:AMZN | AMZN Price Prediction) CEO Andy Jassy has effectively welded AWS to. And based on his July 30 earnings call, he is doubling down.

Jassy on Anthropic: Multi-Gigawatt, Multi-Year, Fully Booked #

Jassy told analysts that “the two leading AI labs in the world, Anthropic and OpenAI,” have made “multi-year, multi-gigawatt commitments” to Amazon’s Trainium chips. Anthropic alone has committed to secure up to 5 GW of current and future Trainium generations, anchoring Project Rainier, which Amazon describes as the “world’s largest operational AI compute cluster with 500,000+ Trainium2 chips for Anthropic’s Claude training.”

The financial imprint is already enormous. Amazon’s stake in Anthropic generated a $53.4 billion one-time non-operating pre-tax gain in Q2 FY2026, lifting GAAP EPS to $5.75 against a $1.82 consensus. Comparable EPS came in at roughly $1.88 versus a $1.83 estimate. Prior-quarter Anthropic gains ran $16.8 billion in Q1 and $9.5 billion in Q3 2025.

AWS Is Booming Because of the Bet #

Jassy said AWS is “booming right now,” posting revenue growth of 36.7% year-over-year, its fastest growth in 18 quarters, on a $169 billion annualized run rate. Backlog stands at $496 billion, growing triple digits year-over-year. Amazon’s chips business and AI revenue each cleared a $25 billion run rate, and Jassy said “the lion’s share of capacity in 27… is largely reserved,” with “quite a bit of capacity that’s already been reserved for 28.”

Amazon Is Also Building Its Own Frontier Model #

Asked whether Amazon needs to compete with the very lab it funds, Jassy said: “We are pursuing our own frontier model.” His reasoning: “within the next few years, you’re going to have at least a half dozen models that are comparably good to each other. They’ll all be in bedrock, and one of them will be ours.”

Q2 revenue reached $200.61 billion, up 19.6% YoY, with operating income at $27.46 billion. Cash capex hit $53.1 billion in Q2 alone. Shares trade at $252.37, up 9.34% YTD but off 9.25% over the past month. Polymarket traders currently price 90.5% odds that Amazon’s 2026 capex tops $200 billion. The safety debate at Anthropic is now, mathematically, a shareholder issue.

Contact [email protected] for any questions or corrections.

── more in #ai-safety 4 stories · sorted by recency
── more on @anthropic 3 stories trending now
sponsored brought to you by zahid.host 4,200+ EU-deployed projects
reading about agents? ship yours in a single git push.

Run your AI side-project on zahid.host

EU-based hosting, git-push deploys, automatic HTTPS, no cold starts. Free tier with a custom domain — perfect for shipping the agent you just read about.

$git push zahid main
Live at https://your-agent.zahid.host
Get free account → Pricing
from €0/mo · no card required
LIVE [news/anthropic-employees-…] indexed:0 read:3min 2026-09-10 ·