AI Industry Development & Policy Dynamics

Jason Calacanis is making a much bigger bet on decentralized AI than a casual market call. In an episode of This Week In Startups, the veteran angel investor appears to frame Bittensor’s TAO token as the kind of asymmetric opportunity that venture investors spend years hunting for, and one that, in

According to CoinDesk 20 data, Bittensor (TAO) jumped 4.1% over the weekend, with Chainlink (LINK) also rising 2.7% since Friday.

According to BlockBeats, on May 4, OpenAI raised over $4 billion to establish a new joint venture focused on helping enterprises adopt its artificial intelligence software. The venture, called The Deployment Company, is backed by 19 investors including TPG Inc., Brookfield Asset Management,

Karpathy pointed out that AI cognition gaps come from two diagnoses: 1) having only used free/older models, which cannot reflect the capability of the most advanced agentic models; 2) ability improvements are only significant in highly technical domains and they show asymmetry. This causes paid users and general readers to see different phenomena, leading to mutual misunderstandings in parallel worlds. The takeaway for readers in Taiwan is that only by actually using the most advanced models to complete production tasks and verify the basis for judgments can you avoid being swayed by media myths.

According to Globenewswire, Applied Digital secured $300 million in senior secured bridge financing led by Goldman Sachs on May 4 to advance its AI data center development. The financing is collateralized by project assets, can be repaid early without penalties, and the company plans to pursue

OpenAI founding team member and former Tesla AI chief Andrej Karpathy posted on X a “LLM Knowledge Bases” workflow, explaining how he recently shifted his large token usage from “controlling code” to “controlling knowledge”—using an LLM to turn scattered papers, articles, folders, and images into a self-maintaining personal wiki. The whole process has already accumulated ~100 articles, ~400k words in his own research projects, and the entire pipeline is written and updated by the LLM. This article summarizes Karpathy’s complete setup and provides developers who want to replicate it with an actionable checklist. Core idea: raw data → LLM compilation → wiki → Q&A Karpathy’s design philosophy can be condensed into a

This article summarizes Andrej Karpathy’s four-step method for “making an LLM argue against its own claims”: first, use the LLM to refine an initial draft; then, start a new conversation and ask it to propose opposing arguments and provide evidence; next, turn it into a complete rebuttal article; finally, compare both sides and the verifiable evidence, with humans making the judgment. Remind that the LLM may exhibit sycophancy—don’t treat agreement as truth; use it instead as an argument-generating machine rather than a final arbiter. You can also have two LLMs debate each other to reduce bias. This method is especially valuable for content creation in 2026, improving cognitive depth and reducing homogenization.

According to ChainCatcher, Nasdaq-listed bitcoin treasury company K Wave Media announced on May 4 a strategic pivot to AI infrastructure, securing up to $485 million in capital support for data center investments, GPU rental services, and AI infrastructure acquisitions and partnerships. The company

A well-known Silicon Valley podcast, 《All-In Podcast》, discussed the AI industry and offered a rather sharp observation: American society’s attitude toward AI is shifting negative, and the most concrete outlet for this sentiment is the data centers that AI companies are racing to build across the U.S. The source of this backlash may include AI doomsday fears, fear of unemployment, or a deeper sense of dissatisfaction: it seems that yet another wave of technological innovation will only make a small number of people get rich, while most people’s lives show no clear improvement. Local U.S. governments have already overturned data center construction cases On the show, Chamath Palihapitiya said that the current problems facing the AI industry are not just competition among models, capital expenditures, or a shortage of compute power, but that “the American people, overall, are becoming increasingly repulsed by the whole issue of AI.” He noted that the source of this repulsion may include AI companies

Antimatter, a France-based cloud infrastructure company for AI workloads, launched on May 4 by combining three existing companies: Datafactory, Policloud, and Hivenet. The company is raising 300 million euros (US$351 million) to deploy 100 micro data center units in 2026 for AI inference

According to BlockBeats, on May 4, AI community user X Freeze questioned whether mainstream AI models including ChatGPT, Claude, and Gemini show systematic bias by being less aligned with conservative positions on issues such as gender, immigration, and crime. The user suggested that as AI

Cerebras is set to list on Nasdaq, with an estimated valuation of about $40 billion and a fundraising cap of $4 billion, issuing shares at $115–$125. Powered by its wafer-scale processor WSE, it positions itself as an AI infrastructure platform, with clients including OpenAI and AWS, and TSMC as a key contract manufacturer. Revenue in 2025 is expected to be $510 million but with losses, relying on a small number of major customers such as G42 and MBZUAI; OpenAI plans to deploy 750MW of low-latency compute, and AWS Bedrock also has a partnership.

According to Reuters, AI chipmaker Cerebras Systems started its IPO roadshow on May 4 for a Nasdaq listing under the ticker CBRS. The company plans to price shares at US$115 to US$125 and could raise up to US$4 billion at a valuation of approximately US$40 billion. This marks Cerebras' second

The Ministry of Education is promoting “AI in Every Library” (“AI in Every Library”). Starting in the fourth quarter this year, national libraries such as the National Library of Taiwan will set up five AI computers per library. The public can use tools such as ChatGPT, Claude, and Gemini for free with a library card, narrowing the paid AI gap, and expanding to libraries at 47 national universities. Funding will come from school self-raised sources or approved grants; meanwhile, challenges such as usage time slots, account management, privacy, and licensing must be overcome.

Morgan Stanley once again raised its forecast for capital expenditure by the top five U.S. cloud giants: about $805B in 2026 and about $1.116 trillion in 2027, and in 2026 spending on non-tech sectors will be roughly comparable to 2025. David Sacks said AI capex contributes about 2.5% to GDP and could exceed 3% next year, viewed as a new engine for the U.S. economy.

The Wall Street Journal said that, to prepare for IPOs from SpaceX, OpenAI, Anthropic, and others, the S&P and Nasdaq have loosened the criteria for component stock selection. The S&P has cut the new-share observation period to six months, or waived the profit threshold; the Nasdaq 100 has introduced fast-track inclusion, allowing ultra-large IPOs to be added to the index on the 15th day after listing. Selection criteria have been changed to rely on total market capitalization, remove the minimum free-float requirement, and adopt dynamic weighting. Experts questioned whether it could amplify the risk of popular stocks and distort price discovery when it may be insufficient.

According to Bloomberg, on May 4, Australia's stock exchange operator ASX warned companies not to overstate the impact of artificial intelligence on their businesses. ASX Chief Compliance Officer Lucinda McCann said the exchange was watching for "ramping" or claims intended to lift share prices

Megazone Cloud, a South Korea-based cloud services firm, was selected by Amazon Web Services on May 4 as the operations partner for AWS Agentic AI GameDay in South Korea, according to the company announcement. The program is a hands-on training event led by AWS's Generative AI Innovation Center,

According to Solana co-founder Anatoly Yakovenko, speaking at the 2026 Solana Breakpoint conference in Amsterdam, artificial intelligence poses a greater existential threat to blockchain security than quantum computers. Yakovenko warned that AI models could exploit subtle mathematical patterns in po

AI virtual models have moved from discussion experiments to fully fledged business entities earning tens of thousands of dollars per month, with the entire technical stack rapidly maturing within 18 months. This article compiles three typical cases: Aitana López under The Clueless (Barcelona, earning €10k per month), Emily Pellegrini created by an anonymous creator (about $10,000 per week), and the recent “Maya” case circulating on X—an alleged AI virtual persona built by a Texas student using 4 markdown files, generating $43k in revenue on her OnlyFans account in the first month. Aitana López: The Clueless’ in-house virtual model, earning up to €10,000 per month Aitana López is from