TheChek News Flow

President offers reassurance after crackdowns even as he stresses need to serve party’s strategic aims

In a meeting with leading entrepreneurs this week, China’s President Xi Jinping celebrated a rising tech industry star while rehabilitating a once-venerated tycoon as he sought to restore investor confidence in the country’s struggling economy.

State media on Monday showed the Chinese president warmly greeting Liang Wenfeng, the youthful chief of artificial intelligence start-up DeepSeek, and shaking hands with Jack Ma, the Alibaba founder who fell from favour after criticising regulators in 2020.

In one step, Xi sought to reassure established figures whose tech groups powered China’s recent economic rise — some of whom have been personally targeted by government crackdowns — even as he signalled a new era for the private sector following Deepseek’s dramatic emergence this year as a rival to western AI companies.

The meeting showed how quickly “political winds” can change for Chinese business, said Han Shen Lin, China country director for US consultancy The Asia Group. 

“It was only a few years ago that the same sector was under a regulatory chokehold,” said Lin. “Now suddenly they’re the favoured son again.”

The meeting on Monday was Xi’s first high-profile encounter with private entrepreneurs in several years. He took pains to emphasise the entrepreneurs’ importance to China’s economic strength, referring to the “two unshakeable principles” — meaning that both the public and private sector should be supported.

But he also reiterated the ruling Chinese Communist party’s control over business, stressing that companies should be “ambitious in serving the country”.

“It’s a recognition that the dynamism of the private sector matters, it adds strength to the Chinese economy and to the Chinese state’s capacity,” said Manoj Kewalramani, author of a newsletter that provides daily interpretations of the CCP’s flagship People’s Daily newspaper.

Kewalramani said Xi was conveying to business leaders that “you are valuable players — but . . . in the grand scheme of things, you are not there just to serve your own needs”. 

Most of the attendees were involved in advanced industries, such as electric vehicles and batteries, robots and other electronic hardware sectors, underlining Xi’s strategy of elevating China’s manufacturing value chain as it competes with the US for tech supremacy.

Aside from Ma and Liang, the attendees included Robin Zeng, chair of leading battery maker CATL, Unitree robots’ founder Wang Xingxing and Wang Chuanfu, chair of electric vehicle maker BYD. 

“Xi wants to have a conversation with the CEOs who are going places, pushing their industries forward,” said Rupert Hoogewerf, whose research company Hurun Report has chronicled the rise of China’s top entrepreneurs.

“This is the new guard that Beijing wants to encourage,” he added.

Chen Long, founder of research group Plenum, said the timing indicated that Xi’s administration was looking to harness the positive momentum in business and investor confidence following the launch of DeepSeek’s groundbreaking AI model, which “showed China could innovate”.

“Now, this meeting removes policy risk for private business which some investors were still worried about,” Chen said.

Encouraging private sector investment is critical for Xi to revitalise the economy, which has been beset by slowing growth in the wake of a years-long property sector bubble.

Government data shows that public sector spending in recent years has driven fixed asset investment growth, which measures spending on infrastructure, property and equipment. Private investment, by contrast, has pulled back, contracting 0.1 per cent in 2024 from the previous year, following a 0.4 per cent decline in 2023, reflecting caution among business.

But Chen said the economy was gradually pivoting towards Xi’s priorities of advanced tech and high-end manufacturing. “It’s a structural transition,” he said.

Lin of The Asia Group stressed that Xi’s message was crucially to “incentivise innovation in service of the state, not for its own sake”, with the private sector ultimately serving the party’s strategic goals.

The Chinese leader urged the business leaders present at Monday’s meeting to “actively fulfil social responsibilities” and “promote common prosperity”.

The latter term — which experts say decries excessive wealth — was invoked to justify Xi’s crackdown on the tech sector, which began with the scrapping of the IPO of Ma’s Ant Group in late 2020 and wiped billions of dollars from the market values of leading Chinese companies.

The Chinese leader promised a level playing field for private businesses this week, and the resolution of persistent challenges such as high financing costs and late payment by state bodies as well as an end to arbitrary fees, fines and inspections.

But some analysts questioned whether those pledges would stimulate a broader recovery in private sector confidence in China, beyond cheering big tech companies and their investors. 

Xi and other party leaders have regularly sought to boost business sentiment in the past year with promises to curb mistreatment of business by cash-strapped local governments looking to raise money. But the problems, which can include detentions of executives, have persisted. 

A Beijing-based entrepreneur said he hoped Xi’s messaging would resolve issues such as delayed payments from state-owned groups. 

He said a state-owned automaker had owed him almost Rmb10mn ($1.4mn) for nearly a year. “It’s very difficult to pay staff without this money,” he said, asking not to be named to avoid offending his state-owned partner.  

OpenAI considers special voting rights to fend off hostile bidders such as Musk

Start-up in talks over ways to allow its non-profit board to maintain control after conversion to a for-profit business

OpenAI is considering granting special voting rights to its non-profit board in order to preserve the power of its directors, as the $157bn start-up fends off an unsolicited takeover bid from Elon Musk.

Chief executive Sam Altman and other board members are weighing a range of new governance mechanisms after OpenAI converts into a more conventional for-profit company, according to people with direct knowledge of the discussions.

Giving the non-profit’s board outsized voting power would ensure it retained control of the restructured company and was able to over-rule other investors including existing backers such as Microsoft and SoftBank.

While no firm decisions have been made, special voting rights would also ensure OpenAI can fight off hostile bids from outsiders such as Musk. The billionaire made a surprise $97.4bn cash bid for the assets held by the non-profit, including its controlling stake in the start-up’s for-profit subsidiary.

Special voting rights could keep power in the hands of its non-profit arm in future and so address the Tesla chief’s criticisms that Altman and OpenAI have moved away from their original mission of creating powerful AI for the benefit of humanity.

The ChatGPT-maker is in the midst of a complex conversion from a non-profit research organisation to a more conventional for-profit business known as a public benefit corporation (PBC).

Its executives argue the shift would allow it to receive billions of dollars more investment as it races rivals such as Google to build the technology, while potentially unlocking greater returns for investors. OpenAI is also in talks to raise $40bn from an investment group led by SoftBank at a valuation of $260bn.

The San Francisco-based group’s plans have come under fire from Musk, who co-founded OpenAI and donated tens of millions of dollars before leaving its board in 2018 in a fallout with Altman. He said at an event in Dubai last week that “OpenAI is trying to completely delete the non-profit, [and] that seems too far.”

The comments come as the board seeks to find a fair price for the non-profit’s assets and decide what role its charitable arm should play in the PBC.

The FT has previously reported that OpenAI has discussed trading the non-profit’s assets for $30bn, paid largely in equity in the PBC. Musk’s $97.4bn offer implies that OpenAI is vastly underpricing the deal.

To complete the conversion, the attorney-general of Delaware — where OpenAI is incorporated — must decide whether the transaction represents fair value and is for the public benefit. People close to OpenAI said the board is under no obligation to sell, or to maximise the valuation of the non-profit’s assets in a competitive auction.

On Friday, the board unanimously rejected the offer from Musk, calling it his “latest attempt to disrupt his competition.”

“OpenAI is not for sale,” Bret Taylor, chair of OpenAI, added. “Any potential reorganisation of OpenAI will strengthen our non-profit and its mission.”

“Of course they are putting the charity’s assets up for sale. That is what their ‘reorganisation’ is all about,” Musk’s attorney, Marc Toberoff, responded in a statement to the FT.

“They are just selling it to themselves at a fraction of what Musk has offered, enriching board members in a classic self-dealing transaction. Will someone please explain how that benefits ‘all of humanity’?”

Carl Tobias, professor at the University of Richmond School of Law, said Musk’s bid was unlikely to affect OpenAI’s conversion to a PBC.

“I think Musk is simply trying to torture Open AI generally and Sam Altman specifically,” he said.

However, the takeover offer has focused minds at the start-up on how to retain control once a conversion is completed.

Silicon Valley groups have often adopted corporate structures that include special voting rights, often to entrench the power of founders. Meta’s Mark Zuckerberg has retained control over his company, despite owning less than 15 per cent of the social media group, thanks to his ownership of shares with outsized voting power.

Another option that could be considered by OpenAI’s board is introducing a shareholder rights plan, or poison pill, which would allow shareholders to buy up additional shares at a discount in order to fend off hostile takeovers. Poison pills were pioneered by lawyer Martin Lipton, co-founder of the firm now advising OpenAI’s board, in the 1980s.

In 2022, Twitter launched a poison pill defence in a failed attempt to prevent Musk’s $44bn buyout of the company.

As part of the conversion under discussion, the current board is expected to oversee the PBC. The non-profit will recruit new directors.

OpenAI is being advised by Goldman Sachs and M. Klein & Company, who are tasked with ensuring that the conversion into a PBC, with a separate non-profit arm, is fair and responsible, according to multiple people with knowledge of the matter.

“A well-thinking board would actually think about [the $97.4bn bid],” said one person with knowledge of OpenAI’s deliberations. But, they added, “the board needs to consider the mission and nobody else — not Sam, not the brand OpenAI, and certainly not Mr Musk.”

Gapping down
In reaction to earnings/guidance:

Other news:

Gapping up
In reaction to earnings/guidance:

Other news:

AI models are being cranked out at a dizzying pace, by everyone from Big Tech companies like Google to startups like OpenAI and Anthropic. Keeping track of the latest ones can be overwhelming.

Adding to the confusion is that AI models are often promoted based on industry benchmarks. But these technical metrics often reveal little about how real people and companies actually use them.

To cut through the noise, TechCrunch has compiled an overview of the most advanced AI models released since 2024, with details on how to use them and what they’re best for. We’ll keep this list updated with the latest launches, too.

There are literally over a million AI models out there: HuggingFace, for example, hosts over 1.4 million. So this list might miss some models that perform better, in one way or another.

AI models released in 2025

OpenAI o3-mini

This is OpenAI’s latest reasoning model and is optimized for STEM-related tasks like coding, math, and science. It’s not OpenAI’s most powerful model but because it’s smaller, the company says it’s significantly lower-cost. It is available for free but requires a subscription for heavy users.

OpenAI Deep Research

OpenAI’s Deep Research is designed for doing in-depth research on a topic with clear citations. This service is only available with ChatGPT’s $200 per month Pro subscription. OpenAI recommends it for everything from science to shopping research, but beware that hallucinations remain a problem for AI.

Mistral Le Chat

Mistral has launched app versions of Le Chat, a multimodal AI personal assistant. Mistral claims Le Chat responds faster than any other chatbot. It also has a paid version with up-to-date journalism from the AFP. Tests from Le Monde found Le Chat’s performance impressive, although it made more errors than ChatGPT.

OpenAI Operator

OpenAI’s Operator is meant to be a personal intern that can do things independently, like help you buy groceries. It requires a $200 a month ChatGPT pro subscription. AI agents hold a lot of promise, but they’re still experimental: a Washington Post reviewer says Operator decided on its own to order a dozen eggs for $31, paid with the reviewer’s credit card.

Google Gemini 2.0 Pro Experimental

Google Gemini’s much-awaited flagship model says it excels at coding and understanding general knowledge. It also has a super-long context window of 2 million tokens, helping users who need to quickly process massive chunks of text. The service requires (at minimum) a Google One AI Premium subscription of $19.99 a month.

AI models released in 2024

DeepSeek R1

This Chinese AI model took Silicon Valley by storm. DeepSeek’s R1 performs well on coding and math, while its open source nature means anyone can run it locally. Plus, it’s free. However, R1 integrates Chinese government censorship and faces rising bans for potentially sending user data back to China.

Gemini Deep Research

Deep Research summarizes Google’s search results in a simple and well-cited document. The service is helpful for students and anyone else who needs a quick research summary. However, its quality isn’t nearly as good as an actual peer-reviewed paper. Deep Research requires a $19.99 Google One AI Premium subscription.

Meta Llama 3.3 7B

This is the newest and most advanced version of Meta’s open source Llama AI models. Meta has touted this version as its cheapest and most efficient yet, especially for math, general knowledge, and instruction following. It is free and open source.

OpenAI Sora

Sora is a model that creates realistic videos based on text. While it can generate entire scenes rather than just clips, OpenAI admits that it often generates “unrealistic physics.” It’s currently only available on paid versions of ChatGPT, starting with Plus which is $20 a month.

Alibaba Qwen QwQ-32B-Preview

This model is one of the few to rival OpenAI’s o1 on certain industry benchmarks, excelling in math and coding. Ironically for a ‘reasoning model,’ it has “room for improvement in common sense reasoning,” Alibaba says. It also incorporates Chinese government censorship, TechCrunch testing shows. It’s free and open source.

Anthropic’s Computer Use

Claude’s Computer Use is meant to take control of your computer to complete tasks like coding or booking a plane ticket, making it a predecessor of OpenAI’s Operator. Computer use, however, remains in beta. Pricing is via API: $0.80 per million tokens of input, and $4 per million tokens of output.

x.AI’s Grok 2

x.AI, the Elon Musk-owned AI company, has launched an enhanced version of its flagship Grok 2 chatbot it claims is “three times faster.” Free users are limited to 10 questions every two hours on Grok, while subscribers to X’s Premium and Premium+ plans enjoy higher usage limits. x.AI also launched an image generator, Aurora, that produces highly photorealistic images, including some graphic or violent content.

OpenAI o1

OpenAI’s o1 family is meant to produce better answers by “thinking” through responses through a hidden reasoning feature. The model excels at coding, math, and safety, OpenAI claims, but has issues deceiving humans, too. O1 requires subscribing to ChatGPT Plus, which is $20 a month.

Anthropic’s Claude Sonnet 3.5

Claude Sonnet 3.5 is a model Anthropic claims as best-in-class. It’s become known for its coding capabilities and is considered a tech insider’s chatbot of choice. The model can be accessed for free on Claude although heavy users will need a $20 monthly Pro subscription. While it can understand images, it can’t generate them.

OpenAI GPT 4o-mini

OpenAI has touted GPT 4o-mini as its most affordable and fastest model yet thanks to its small size. It’s meant to enable a broad range of tasks like powering customer service chatbots. The model is available on ChatGPT’s free tier. It’s better suited for high-volume simple tasks compared to more complex ones.

Cohere Command R+

Cohere’s Command R+ model excels at complex Retrieval-Augmented Generation (or RAG) applications for enterprises. That means it can find and cite specific pieces of information really well. (The inventor of RAG actually works at Cohere.) Still, RAG doesn’t fully solve AI’s hallucination problem.