Sent to 17 readers
Race to AGI Daily Digest - Friday, March 6, 2026
Share:
TLDR
China just made AI and open models a core plank of its economic playbook, locking in long‑term demand for domestic labs and chip supply.
OpenAI at a $25B run‑rate and Anthropic at $19B turn the reliability race into a revenue race, cementing frontier labs as Big Tech peers.
New funding for Unleash and Cent shows ‘govern the AI, not just use it’ becoming its own product category across code and healthcare.
The Full Story
All week we’ve been talking about an AI market that can’t decide between fear and FOMO. Monday and Tuesday swung the AI index hard, Wednesday punished Intel, and yesterday OpenAI’s GPT‑5.3 Instant turned “fewer hallucinations” into a product feature. Today the same story shows up at national‑policy scale and in someone’s Series B deck.
On the policy side, China just upgraded AI from buzzword to backbone. Its latest planning documents put an “intelligent economy” and AI+ at the center of 2026 goals, with a five‑year plan that treats AI and open models as core economic infrastructure rather than side projects. That’s a huge signal for players like Alibaba ->, Huawei, iFlytek, SenseTime, and newer labs like DeepSeek ->. You can see the plan’s contours here: China five-year plan puts AI and open models at economic core ->.
Building on yesterday’s reliability theme, we’re also seeing a market form around “AI that doesn’t wreck your codebase.” Unleash just raised $35M to govern AI‑generated code, effectively sitting between ChatGPT‑style tools and production repos. That’s reliability arms race, but for software supply chains. More quietly, Cent is funding AI health scans in India, pushing the same idea into medicine. Both are covered in: Unleash secures $35M Series B to govern AI‑generated code ->.
Then there’s the money scoreboard. OpenAI is reportedly at a $25B annualized revenue run‑rate, with Anthropic -> at $19B. That’s not “promising startup” territory; that’s Big Tech revenue, built almost entirely on selling access to frontier models. Check the numbers: OpenAI hits $25B and Anthropic $19B in annualized AI revenue ->. No surprise Hacker News is already buzzing about “GPT‑5.4” and a rumored GitHub rival.
Markets rounded out the picture. Snowflake, Broadcom, Salesforce, and IBM all climbed, while ARM slipped. So the week that started with anxiety about chip exports now looks more like a positioning game: who owns the reliable models, who controls the data, and which countries get to set the rules. If you want to see the geopolitical layer under all this, keep watching our running story on export controls: AI Chip Exports Spark National Security Reckoning ->. The question hanging over Friday: in the race to AGI, does reliability become the real moat, or do governments and customers force these labs to share the keys?
Get This Delivered Daily
Join thousands of AI professionals who start their day with Race to AGI.