AI capex spending concerns triggered a tech-led selloff with the Nasdaq 100 down 2.9%, semiconductors down 6.7%, and the leveraged semis ETF $SOXL crashing 20.5% — just one week after a record $3.8B inflow into it.
First reported Thu 22 Oct · 21:27 UTCUS hyperscalers led a broad tech selloff with Amazon –4%, Oracle –2.4%, Alphabet –2.3%, while the 10-year Treasury yield dipped below 5% for the first time since late September, with Goldman framing AI weakness as paradoxically bullish for bonds.
First reported Thu 22 Oct · 16:27 UTCMeta reported its first-ever negative free cash flow quarter since its 2012 IPO, driven by an 80% YoY surge in AI infrastructure spending to $34.9B. Reality Labs has now lost a cumulative $92B since 2020, while Meta's credit risk hit a new high.
Updated 2 times since 21 Oct 2026Chinese AI lab DeepSeek has raised $15 billion in a financing round at a $74B pre-money valuation, with CATL, Tencent and others as backers, and is reportedly accelerating its IPO timeline.
Updated 4 times since 20 Oct 2026Anthropic reportedly disclosed an $80 billion net loss for the first half of 2026, with an analyst framing the financial leak as a deliberate pre-raise strategy by the company's CFO and bankers to gauge investor reaction.
First reported Thu 22 Oct · 04:29 UTCAI-related corporate bond issuance hit a record $419B YTD in 2026 (+90% YoY), with speculative data center debt being repackaged and sold to pension funds in ways analysts are comparing directly to the subprime mortgage bubble.
First reported Wed 21 Oct · 15:28 UTCOpenAI is partnering with Google Cloud to distribute frontier models (GPT-6 Astra, Codex) via Google Cloud Marketplace, while simultaneously expanding ChatGPT ads to 89 countries.
Updated 2 times since 21 Oct 2026Anthropic reportedly posted an $80.7B net loss in the first half of 2026 against ~$16.3B in revenue, raising serious questions about sustainability at a $2T IPO valuation.
First reported Wed 21 Oct · 07:25 UTCChamath, Citadel's Ken Griffin, and Charles Edwards all separately sounded alarms or milestones around AI — from a VC recycling critique to AI surpassing human intelligence in math — while the FT reported market jitters.
Updated 5 times since 20 Oct 2026Anthropic faces a two-pronged controversy: criticism that its AI legislation involvement amounts to regulatory capture that suppresses free expression, and a reported $80 billion net loss for the first half of 2026.
First reported Wed 21 Oct · 04:33 UTCCritics argue AI capability research is arriving dead on arrival — papers are still using two-year-old models by the time they pass peer review, making academic publishing structurally unfit for the AI era.
First reported Wed 21 Oct · 03:32 UTCAnthropic posted an $80 billion net loss in H1 2026 per WSJ. Simultaneously, pre-IPO perpetual futures for Anthropic launched on Phoenix with up to 6x leverage, drawing significant crypto community attention.
First reported Wed 21 Oct · 00:33 UTCCorporate AI token spending growth collapsed from 42% to 0.2% in two months per Ramp data, yet SMCI's blowout earnings and SpaceX's massive revenue guidance are repairing the bullish AI infrastructure narrative in the same session.
Updated 2 times since 20 Oct 2026Anthropic's S-1 filing for its IPO revealed an $80 billion net loss for H1 2026 and contained extraordinary risk language warning that sufficiently advanced AI could theoretically transmit information backward in time.
Updated 2 times since 20 Oct 2026Anthropic's draft IPO filing reveals $16.3B revenue in H1 2026 but an $80.7B net loss driven by accounting charges, ~$9B in compute spending, and ~$450B in cloud commitments over 7–10 years — with the S-1 also containing extraordinary language about AI potentially transmitting information backward in time.
First reported Tue 20 Oct · 21:34 UTCA new academic paper predicts a major AI-enabled terrorism event as early as 2027–2028, while Anthropic's internal model was separately tested on creating experimental scorpion antivenom — highlighting the dual-use stakes of frontier AI in biology.
First reported Tue 20 Oct · 18:32 UTCReports that OpenAI is rushing to raise $30B at a $1.4T valuation from SoftBank and MGX sent SoftBank CDS to a 4-year high, as investors simultaneously race to buy CDS protection on AI-linked companies more broadly.
First reported Tue 20 Oct · 17:33 UTCA viral anecdote about a senior FX desk head developing 'terminal AI psychosis' after over-relying on an AI assistant underscores growing concerns about AI misuse in professional trading.
First reported Tue 20 Oct · 06:32 UTCMultiple sources are converging on concerns about overbuilt AI infrastructure — unlit GPU warehouses, surging AI debt default insurance, and a sharp CoreWeave price target cut — while corporate AI deals continue to be announced.
First reported Tue 20 Oct · 04:32 UTCOracle credit default swaps hit all-time wides as data center deals collapse en masse. CoreWeave's own co-founder called life insurers backing the AI buildout the 'dumbest money,' and Anthropic's revenue growth is reportedly slowing dramatically.
First reported Tue 20 Oct · 03:33 UTCA Gartner analyst's claim that AI returns don't yet justify neocloud build-out is being blamed for a tech selloff hitting Seoul and Tokyo. Simultaneously, a major compiled list of collapsed data center deals is circulating widely.
First reported Tue 20 Oct · 02:33 UTCThe Fed warned AI growth could worsen inflation, Morgan Stanley outlined three ways AI's $1.6T capex plan could collapse, and new analysis says the US AI lead over China is a thin ~6 months — but could be doubled with existing federal authority.
Updated 5 times since 18 Oct 2026Multiple threads highlight China's deepening AI compute access via US chips and its deliberate hoarding of industrial tech leads, with calls for the US to cut off remote chip access immediately.
First reported Tue 20 Oct · 00:32 UTCAnthropic is reportedly in talks with NScale for a $45 billion cloud computing deal, while the FT is hosting a reader Q&A on what could be the biggest IPO in stock market history.
First reported Mon 19 Oct · 19:33 UTCA widely-amplified thread from Ansem argues that as AI models commoditize, distribution and customer retention will be the defining competitive advantage — with OpenAI's massive consumer adoption as the prime example.
First reported Mon 19 Oct · 17:34 UTCHedge funds are dumping AI stocks at record pace while retail pours a record $7B into 3x leveraged chip ETFs. NVDA/AVGO credit default swaps hit all-time wides, and DeepSeek-Huawei chip collaboration adds a new competitive threat to Nvidia.
First reported Mon 19 Oct · 16:34 UTCAnthropic is reportedly readying a new model that better reads users' intentions. Separately, the Trump administration's senior Defense official threatened export controls against AI companies that don't cooperate with the government.
First reported Mon 19 Oct · 14:33 UTCAmazon is co-developing a nearly 1 GW nuclear plant in Indiana for data centers, opening its AWS Trainium AI chips to all PyTorch developers, and a new SemiAnalysis report forecasts data center capacity doubling to 51.2 GW in 2027.
First reported Mon 19 Oct · 13:33 UTCAnthropic committed ~5% of shares (worth ~$48bn) to a public benefit fund with a dilution kicker tied to valuation growth, sparking both corporate governance debate and speculative meme-token activity.
First reported Mon 19 Oct · 08:32 UTCZerohedge highlights that Nvidia is projected to capture $690B of total 2027 AI capex, while the other 499 S&P companies combined get just $340B — underscoring extreme concentration in the AI infrastructure buildout.
First reported Sun 18 Oct · 21:33 UTC