Photo by Steve Johnson on Unsplash

Catch up on the top artificial intelligence news and commentary by Wall Street analysts on publicly traded companies in the technology space with this short recap.

X-Like Social Network 

Microsoft-backed OpenAI is developing its own X-like social media platform, The Verge's Kylie Robison and Alex Heath reported, citing multiple sources familiar with the matter. While the project remains in preliminary stages, an internal prototype exists that is focused on ChatGPT's image generation and includes a social feed, the report stated, noting that sources said CEO Sam Altman has been privately asking outsiders for feedback on the matter.

AI Coding Tool Maker 

OpenAI is also in talks to acquire Windsurf, an artificial intelligence-assisted coding tool maker formerly known as Codeium, for about $3 billion, according to Bloomberg's Rachel Metz, Kate Clark, and Shirin Ghaffary, who cited a person familiar with the matter.

Such a deal would be OpenAI's largest acquisition to date and could help the company take on rising competition in the market for AI-driven coding assistants, though terms haven't been finalized and talks could still change or fall apart, the source said.

DeepSeek

After the Trump administration moved to restrict Nvidia's (NVDA) sale of AI chips to China, it has also begun weighing penalties that would block DeepSeek, China's artificial intelligence star, from buying U.S. technology and is debating barring Americans' access to its services, three people with knowledge of the matter told The New York Times' Tripp Mickle, Ana Swanson, Meaghan Tobin, and Cade Metz.

The House Select Committee on the Chinese Communist Party opened an investigation into Nvidia's sale of chips across Asia to assess whether the U.S. chipmaker knowingly provided DeepSeek with critical technology to develop A.I., potentially in violation of U.S. rules, added the report.

H20 Restrictions

In a regulatory filing, Nvidia stated the following: 

"On April 9, 2025, the U.S. government, or USG, informed NVIDIA Corporation, or the company, that the USG requires a license for export to China (including Hong Kong and Macau) and D:5 countries, or to companies headquartered or with an ultimate parent therein, of the Company's H20 integrated circuits and any other circuits achieving the H20's memory bandwidth, interconnect bandwidth, or combination thereof.

"The USG indicated that the license requirement addresses the risk that the covered products may be used in, or diverted to, a supercomputer in China. On April 14, 2025, the USG informed the company that the license requirement will be in effect for the indefinite future. The company's first quarter of fiscal year 2026 ends on April 27, 2025. First quarter results are expected to include up to approximately $5.5 billion of charges associated with H20 products for inventory, purchase commitments, and related reserves."

Wells Fargo noted that following Nvidia's 8K announcement earlier in the week, AMD (AMD) filed an 8K as well, announcing that it too expects new U.S. Government license controls on GPU sales into China to impact the company's MI308X.

AMD said that while it expects to apply for licenses, there are no assurances that such licenses will be granted. Therefore, the company announced that it expects to take up to an $800 million inventory, purchase commitment, and related reserve charge, likely in Q2, Wells Fargo stated.

Bitnet That Can Run on CPUs

Microsoft (MSFT) researchers claimed they have developed the largest-scale 1-bit AI model, also known as a "bitnet," to date. Called BitNet b1.58 2B4T, it's openly available under an MIT license and can run on CPUs, including Apple's M2, TechCrunch's Kyle Wiggers reported.

The Microsoft researchers said that BitNet b1.58 2B4T is the first bitnet with 2 billion parameters, "parameters" being largely synonymous with "weights." Trained on a data set of 4 trillion tokens, BitNet b1.58 2B4T outperforms traditional models of similar sizes, the researchers claimed.

According to the researchers' testing, the model surpasses Meta's Llama 3.2 1B, Google's Gemma 3 1B, and Alibaba's Qwen 2.5 1.5B on benchmarks including GSM8K and PIQA, the author said.

More By This Author: