LIVE · THU, JUL 09, 2026 --:--:-- ET
Issue Nº 79 COST TOTAL $14731.04 ARTICLES TODAY 6 TOKENS TOTAL 9.39B
Podcast
aiexpert
Subscribe · $8/mo
Running the wire
Chips Imec roadmap: 2.5D optical I/O cuts AI inference interconnect power from 1.25kW to <200W Breaking SpaceXAI releases Grok 4.5 at $2/$6 per MTok; claims parity with Opus 4.7, faster Breaking Grok 4.5 launches at $2/$6 per million tokens: SpaceXAI's economics play vs. Opus, not performance crown Funding Rapidus secures ¥267.6B ($1.7B) for 2nm mass production by 2027; Japan largest shareholder Market SK Hynix US IPO (SKHY) launches July 10; $28B raise 7x oversubscribed, AI memory demand drives 200%+ YTD gain Funding QuantumDiamonds closes €91M (€15M equity + €76M grants) for quantum chip-inspection tech Chips Micron breaks ground on $9.3B HBM fab in Japan; adds subsidized capacity for AI Market Luxshare IPO debuts in Hong Kong, raises HK$24.27B; Apple supplier eyes broader electronics and auto market Funding Prime Intellect hits $1B valuation on $130M Series A; $100M ARR; enterprise RL agents move off-frontier Funding ElevenLabs targets $22B secondary valuation; voice AI doubles from $11B in five months on $500M ARR Breaking SpaceXAI launches Grok 4.5 at Opus-class performance; $2/M input tokens vs Claude $5/M Market Blackstone QTS raises $870M via data center asset-backed securities; Nemotron benchmarks open vs closed models Breaking China flags Claude Code monitoring mechanism as security backdoor; Anthropic says it's anti-abuse experiment Funding Bespoke Labs raises $40M for AI agent training environments; frontier labs abandon pure model scaling Chips NVIDIA denies Kyber NVL144 delay reports; roadmap remains intact for H2 2027 launch Funding Taktile raises $110M Series C from Goldman Sachs for AI decisioning in banks and insurers Funding Bespoke Labs raises $40M Series A to build realistic training grounds for long-horizon AI agents Market SpaceX stock slides to $148 post-Nasdaq 100 inclusion; analysts see AI data center, Starlink upside Market Together AI launches Provisioned Throughput for reserved open-model inference at 90% below closed APIs Market Meta breaks ground on $9B Alberta data center; first Canadian AI facility Chips Imec roadmap: 2.5D optical I/O cuts AI inference interconnect power from 1.25kW to <200W Breaking SpaceXAI releases Grok 4.5 at $2/$6 per MTok; claims parity with Opus 4.7, faster Breaking Grok 4.5 launches at $2/$6 per million tokens: SpaceXAI's economics play vs. Opus, not performance crown Funding Rapidus secures ¥267.6B ($1.7B) for 2nm mass production by 2027; Japan largest shareholder Market SK Hynix US IPO (SKHY) launches July 10; $28B raise 7x oversubscribed, AI memory demand drives 200%+ YTD gain Funding QuantumDiamonds closes €91M (€15M equity + €76M grants) for quantum chip-inspection tech Chips Micron breaks ground on $9.3B HBM fab in Japan; adds subsidized capacity for AI Market Luxshare IPO debuts in Hong Kong, raises HK$24.27B; Apple supplier eyes broader electronics and auto market Funding Prime Intellect hits $1B valuation on $130M Series A; $100M ARR; enterprise RL agents move off-frontier Funding ElevenLabs targets $22B secondary valuation; voice AI doubles from $11B in five months on $500M ARR Breaking SpaceXAI launches Grok 4.5 at Opus-class performance; $2/M input tokens vs Claude $5/M Market Blackstone QTS raises $870M via data center asset-backed securities; Nemotron benchmarks open vs closed models Breaking China flags Claude Code monitoring mechanism as security backdoor; Anthropic says it's anti-abuse experiment Funding Bespoke Labs raises $40M for AI agent training environments; frontier labs abandon pure model scaling Chips NVIDIA denies Kyber NVL144 delay reports; roadmap remains intact for H2 2027 launch Funding Taktile raises $110M Series C from Goldman Sachs for AI decisioning in banks and insurers Funding Bespoke Labs raises $40M Series A to build realistic training grounds for long-horizon AI agents Market SpaceX stock slides to $148 post-Nasdaq 100 inclusion; analysts see AI data center, Starlink upside Market Together AI launches Provisioned Throughput for reserved open-model inference at 90% below closed APIs Market Meta breaks ground on $9B Alberta data center; first Canadian AI facility
Breaking

Grok 4.5 launches at $2/$6 per million tokens: SpaceXAI's economics play vs. Opus, not performance crown

SpaceXAI released Grok 4.5 on July 8-9, 2026 after private beta at SpaceX and Tesla. Elon Musk pitched it as 'Opus-class, but faster, more token-efficient and lower cost,' then clarified: 'roughly comparable to Opus 4.7, but much faster.' The model uses 1.5T-parameter V9 foundation trained on Cursor IDE session data. Price: $2/million input tokens, $6/million output (vs. Claude Opus 4.8 at $5/$25; OpenAI GPT-5.6 Sol at $5/$30).

Benchmarks are mixed. On DeepSWE 1.1 (real bug-closing rate), Grok 4.5 scored 53% vs. Opus 4.8's 59%, GPT-5.5's 67%, and Claude Fable 5's 70%. On SWE Bench Pro, Grok 4.5 posted 64.7%, beating GPT-5.5's 58.6% but trailing Opus 4.8's 69.2% and Fable 5's 80.4%. Musk acknowledged Fable is 'definitely better' but claimed most tasks don't require Fable-level capability.

The efficiency delta is where Grok competes: on SWE Bench Pro, Grok 4.5 used 15,954 output tokens per job vs. Opus 4.8's 67,020—a 4.2x gap. At lower cost per token and massively fewer tokens burned per task, Grok targets high-volume developer teams optimizing total spend over benchmark leaderboards. Throughput: 80 tokens/sec (fast-model territory).

For enterprise teams running production multi-agent systems, this is a classic disruption play: Musk bets that 60% of Opus capability at 40% of the cost and 4x fewer tokens per task wins price-sensitive workloads. Musk promised weekly improvements and a C/C++ rewrite expected in ~3 months for 10x+ training speed gains on Colossus's GB300 GPUs.

Sources

Get the signal before the noise.