Anthropic signs $9.1 billion 20-year AI compute deal with Riot Platforms for 191MW capacity, lifting Riot shares 25% after hours.

Anthropic inks $9.1 billion 20-year pact with Riot Platforms $RIOT for 191 MW AI data center capacity at Rockdale, Texas.

AethirCloud reports Bitcoin miners repurposing powered sites for AI at Tokyo GPU dinner with Saison Capital, gonka_ai and gaib_ai amid $64872 BTC price action.

Cerebras $CBRS lands $20B OpenAI deal for 750MW compute plus AWS pact as Q1 revenue jumps 92% to $191.3M.

CNBC reports on efforts to commoditize AI computing power like oil, creating new AI compute power trading markets and platforms amid 2026 demand surge.

SpaceX lands $920 million monthly from Google for 110,000 NVIDIA GPUs in its data centers through June 2029, following Anthropic pact for $26 billion annual revenue.

SAHARA sees 111 million tokens ($3.81M) withdrawn from Binance and Bitget equaling 3.4% of supply days after staking launch for AI compute resources.

OpenAI commits $2M in API credits to each Y Combinator startup, supplying AI compute resources for the current batch and next generation of founders.

OpenAI trails Anthropic in valuation, funding, gross profit, and revenue, prompting a strategic counterattack with superior talent and compute power amid AI industry impact.

Elon Musk's xAI inks partnership with Anthropic for access to Colossus 1 supercomputer with 220,000 NVIDIA GPUs, ending year of rivalry amid AI hype.

Erik Voorhees details DIEM token equaling $1/day AI compute credits on models like Qwen and Grok, with rising demand bidding up constrained supply for crypto AI integration.

LUVIA token presale kicks off at $0.01, offering decentralized AI compute like AWS on Web3. Early access via LUVIA.exchange amid AI crypto hype.

OpenAI's $500 billion Stargate project shows progress at all 7 US sites, targeting 9+ GW by 2029 amid rising AI compute demands.

According to @thedaoofwei, SpaceX has acquired xAI to form a vertically integrated innovation engine spanning AI, rockets, space-based internet, direct-to-mobile communications and a real time information platform. The source cites Elon Musk outlining a plan to deploy up to a million satellites as orbital data centers powered by near-constant solar energy to overcome terrestrial power and cooling limits for AI compute. The source adds that Starship will begin delivering higher capacity V3 Starlink and next generation direct-to-mobile satellites, with an eventual target cadence of hourly launches carrying 200 tons per flight to support large-scale deployments. The source reports Musk's estimate that space-based compute could become the lowest cost option within two to three years, with stated math of 100 kW per ton enabling roughly 100 gigawatts of AI compute capacity added annually at a one million ton per year launch rate, and a longer-term path to terawatt-scale. The source notes plans for established space sustainability practices plus a roadmap to lunar manufacturing and electromagnetic mass drivers that could scale hundreds to thousands of terawatts per year for deep-space AI infrastructure. The source positions this strategy as enabling AI applications, global connectivity and multi-planetary operations under the combined SpaceX and xAI platform.

According to @OpenAI, achieving AI progress beyond image generation, including in healthcare and scientific discovery, depends on building substantial compute. Source: OpenAI on X, Dec 17, 2025. @OpenAI added that if the United States does not build this compute capacity, other countries will, underscoring urgency around domestic AI infrastructure. Source: OpenAI on X, Dec 17, 2025. For traders, the statement focuses attention on AI compute and infrastructure themes that market participants track across AI-related equities and crypto narratives tied to decentralized compute. Source: OpenAI on X, Dec 17, 2025.

According to @CNBC, its China Connection newsletter reports the U.S.-China AI talent race is intensifying, underscoring chip war dynamics, computing power, talent, and electricity constraints. Source: @CNBC tweet on Dec 10, 2025 and the linked CNBC newsletter article. For crypto traders, these constraints intersect with Bitcoin mining economics because electricity is the dominant operating cost for miners, making power conditions a key margin driver. Source: Cambridge Centre for Alternative Finance, Cambridge Bitcoin Electricity Consumption Index methodology. The International Energy Agency documents rapid growth in data center electricity demand tied to AI workloads, highlighting energy conditions relevant to power‑intensive computing sectors such as BTC mining. Source: International Energy Agency, 2024 Data Centres and AI analysis.

According to AnthropicAI, SGTM unlearning is hard to undo and requires seven times more fine-tuning steps to recover forgotten knowledge compared with the prior RMU method, indicating materially higher reversal effort (source: Anthropic on X, Dec 9, 2025). For trading context, this 7x delta provides a measurable robustness gap between SGTM and RMU that can be tracked as an AI safety metric with direct implications for reversal timelines and optimization iterations (source: Anthropic on X, Dec 9, 2025).

According to @StockMKTNewz, Meta Platforms (META) is reportedly considering using Google’s (GOOGL) Tensor Processing Units (TPUs) in its data centers, citing The Information. Per The Information as relayed by @StockMKTNewz, the report does not disclose timing, pricing, volumes, or any finalized agreement. Google states that TPUs are custom AI accelerators for training and inference available via Google Cloud, framing the potential scope of what Meta is evaluating, per Google Cloud documentation. From a trading standpoint, this remains an unconfirmed consideration, so the key catalysts would be official statements or filings from Meta or Alphabet before positioning, per @StockMKTNewz’s characterization of the item as “reportedly considering.” The sources cited do not reference any cryptocurrency exposure or blockchain initiatives tied to this consideration, indicating no direct crypto market impact in the report, per @StockMKTNewz and The Information.

According to @gdb, AI is the refining of compute into brainpower, but the post contains no data, metrics, product details, or timelines that would inform trading decisions (source: @gdb on X, Nov 15, 2025). For crypto and equity traders focused on AI infrastructure, the statement offers no direct signal on chip demand, model releases, or token economics, so no actionable price catalysts can be derived from this source alone (source: @gdb on X, Nov 15, 2025).

According to @gensynai, users running RL-Swarm are encouraged to update to CodeZero by executing git pull to prepare their environment for launching a new CodeZero node, source: @gensynai (Nov 13, 2025). According to @gensynai, Docker launch options are provided for both CPU (docker-compose run --rm --build -Pit swarm-cpu) and GPU (docker-compose run --rm --build -Pit swarm-gpu), source: @gensynai (Nov 13, 2025). According to @gensynai, a shell-based path is also available: rm -rf .venv, python -m venv .venv, source .venv/bin/activate, and ./run_rl_swarm.sh, source: @gensynai (Nov 13, 2025). According to @gensynai, these instructions target those already operating RL-Swarm to transition to CodeZero nodes, source: @gensynai (Nov 13, 2025).