Skip to main content

icnlive

Menu
Menu
icn.lıve
WATCH LIVE. THINK BUSINESS.
Menu
X-twitter Linkedin-in Instagram Youtube

© 2026 ICN.LIVE
  • GPT-6 Astra

OpenAI GPT-6 Astra Launch Puts Frontier AI Cost and Cyber Risk in Focus

GPT-6 Astra is OpenAI’s new frontier model, and its arrival brings two questions to the center of the AI market: what these

Trending AI

Explore Section:

TECH
AI
Capital
Politics
Banking
Health
OpenAI's executive departures continue
  • OpenAI

OpenAI’s executive departures continue, and the newest one hits a corner of the company that decides how fast its AI can grow. The Wall Street Journal reported that Chris Malone, the OpenAI head of data centers, left last week. He joined in March 2025 after more than a decade at Google and nearly five years at Meta, so his run at the lab stayed short.

Picture what that seat holds. Every AI model needs computing power, and that power needs buildings packed with servers. The person steering that work holds one of the most watched jobs in the field. When they leave, people pay attention.

A short-lived reorganization?

Before Malone left, OpenAI changed the shape of his role. He stopped reporting to president Greg Brockman and began reporting to vice president Sachin Katti, who took over the group. In a statement, OpenAI said it had “recently reorganized” its infrastructure organization to match the scale and pace of its work. The company said it keeps a deep data center team with clear leadership.

Several leaders now share the load. Uday Ruddarraju runs the data center team. Brent Mayo handles build and delivery. Spas Lazarov, a veteran of the data center and energy sectors, leads all data center engineering.

Why OpenAI’s executive departures continue to draw scrutiny

Malone is not walking out alone. His exit follows a long run of OpenAI executive departures, and Business Insider counted 13 in 2026, with several in the past month. These are not junior staff. They sit near the top. OpenAI’s executive departures continue at a pace few large companies would shrug off.

Two weeks ago, OpenAI replaced chief revenue officer Denise Dresser after roughly eight months. Days before that, longtime chief operating officer Brad Lightcap said he would leave to “start something new,” though he shared no plans. About a month earlier, Fidji Simo stepped down as the company’s second-in-command to recover from a “chronic illness.” She stays on as an adviser.

The safety and ethics side has thinned too. OpenAI lost its head of ethics, Chloé Bakalar, in July. Reports also said the company disbanded its preparedness team, the unit that studied whether its models could cause catastrophic harm. Some leaders left because their projects closed. Bill Peebles, who ran the shuttered image tool Sora, was one. Kate Rouch, the former chief marketing officer, left in April for health reasons.

What the OpenAI executive exodus means before the IPO

The company plays down the OpenAI executive exodus. Brockman has said the bright spotlight on the company means every exit gets picked apart in a way rivals avoid. There is truth in that. Even so, the pattern raises real questions.

Reporting from CNBC and Bloomberg points to a shift in strategy. The company has leaned less on the Stargate project and moved toward leasing data center space from cloud providers. That change helps explain the reshuffle around Malone’s team.

Timing sharpens the worry. OpenAI’s public listing, once expected this year, is now reported to be pushed to 2027. A looming OpenAI IPO in 2027 brings a hard look at the numbers. Reporting has questioned whether the company is overvalued and whether its profit can match the huge sums flowing in. As OpenAI’s executive departures continue, investors will weigh what the churn says about the company, while the people still inside keep the work moving.

  • 3 min read
Alibaba eyes AI infrastructure spending
  • Data Centers

Alibaba AI infrastructure spending is climbing again, and the company wants shareholders to fund it. The Chinese ecommerce and cloud group is selling HK$80 billion of new shares, roughly $10.2 billion, with every dollar of net proceeds going into its full-stack AI capabilities. That covers chips, data centres, and the models running on top of them. The Alibaba share placement is the largest primary follow-on offering ever from a Hong Kong-listed company. Globally it ranks third this year, behind Alphabet and Intel.

Pricing tells you how the market took it. Alibaba set 710 million new shares at HK$112.70 each, against a Friday close of HK$123. Hong Kong-listed shares dropped as much as 10 percent on Monday. Buyers at the discount get exposure to the buildout. Existing holders get dilution and a longer wait for returns. US investors were excluded from the deal.

The numbers behind the raise

Alibaba AI capex hit 67.7 billion yuan in the June quarter, up 75 percent from a year earlier. Net profit fell by the same proportion over that period, to roughly $1.5 billion, and free cash outflow reached $6.6 billion. Alibaba AI infrastructure spending sits inside a three-year plan worth at least 380 billion yuan, and the company says it has already spent close to half. CEO Eddie Wu told analysts the compute capacity has to exist before the growth can be captured.

Revenue is arriving behind the bill. Alibaba Cloud revenue from AI and compute services rose 45 percent to 48.44 billion yuan in the quarter, the fastest pace in 22 quarters. Payback on AI-related investment is now expected in about 2.5 years, down from three.

What Alibaba AI infrastructure spending means for you

Hold the stock, and you absorb the dilution today for capacity that pays later, if the demand holds. Build with AI in Asia, and the calculation flips, because more compute usually means cheaper inference and stronger models. The Qwen AI model family sits at the centre of that trade. Alibaba released Qwen 3.8-Max weeks ago, and early benchmarking points to strength in agentic coding, where bots write and repair code from high-level instructions.

China AI investment runs hot

This raise lands in a market already paying up. Chipmaker CXMT pulled in $8.6 billion at listing, and its shares rose 466 percent on debut. Humanoid robotics group Unitree raised $900 million last week, with shares climbing more than 600 percent on day one after retail demand topped 5,500 times the available allotment. Moonshot’s Kimi K3 launch last month added to the mood. China AI investment at these valuations carries real risk if earnings arrive slowly.

Washington is still a problem

Regulation shapes the rest of the story. The Pentagon in June returned Alibaba to a blacklist of Chinese companies treated as a national security risk, alongside Baidu and BYD, citing alleged links to the People’s Liberation Army. Alibaba has asked a US court to overturn the order. The company denies any PLA ties and rejects the claim it takes part in military fusion, where civilian industry works with the state defence sector. Xi Jinping and Donald Trump meet in the US next week, their second summit this year, with export controls and technology restrictions on the agenda. What comes out of that room decides how far Alibaba AI infrastructure spending can travel outside China.

  • 3 min read
Gulf Investment Priorities
  • Business Intelligence

Gulf investment priorities are under fresh scrutiny, and the questions investors ask have changed shape. Bilal Sabouni runs the Middle East, Africa, Turkiye and Central Asia business at Guidepoint, the global expert network. He told ICN Business that client demand now circles one theme.

“A lot of the questions we are hearing come back to one issue: how investment priorities in the Gulf may change in response to the current geopolitical environment,” Sabouni said.

The sectors under review are the ones the region already built. Oil, liquefied natural gas, chemicals and aluminium. Aviation, logistics, construction and defence. Capital went into all of them. Now each one gets a second look. “These investments are being examined more closely than in the past,” Sabouni said. Clients want to know “which areas remain resilient, where risks are increasing, and where new opportunities may emerge.”

That list tracks GCC economic diversification plans almost line for line. Sovereign wealth funds and state-linked investors sit behind much of the money in question.

Gulf investment priorities reach past the region

Two markets keep coming up in client work. India buys heavily from the Middle East. China is tied in through its exports and trade flows.

“A change in investment or production here can have consequences across several markets,” Sabouni said. Speed is the part he thinks people miss. Geopolitical risk is not nudging plans slowly. It is landing inside live decisions. “We are seeing more attention given to logistics networks and alternative routes that reduce dependence on vulnerable trade corridors,” he said.

Read that as supply chain resilience being priced in real time. The route now matters as much as the asset. Sabouni puts the shift plainly. The question is no longer where capital lands. It is how those investments “could reshape trade flows, strengthen regional resilience, and create lasting value.”

Announcements tell you what a government or company plans to do. Operators tell you what is moving.

Why a phone call still costs more than software

Guidepoint sells access to experience. Its network runs to more than 2 million vetted experts across 300-plus industries and 150 countries. The Guidepoint Library holds more than 120,000 expert interviews. Its AskGP tool returns source-cited answers in seconds.

Gulf investment priorities now move faster than published research can track. So why pay a premium for a human in 2026? “Information has become cheaper and more abundant, but abundance does not automatically create understanding,” Sabouni said.

Search engines retrieve what has been published. AI tools summarise what is public. Neither one explains why a rollout failed, how buyers decide, or which local dynamic flips the outcome. “AI can produce a fast answer, but speed alone does not make an answer reliable, current, or decision-ready,” Sabouni said. Clients want to push back, test contradictions, and hear the minority view. You cannot do that with a summary.

What keeps an expert network on the right side of the line

The model rests on a boundary. Clients get experience and informed opinion. They never get confidential, proprietary, or material non-public information. Sabouni says compliance is built in rather than bolted on. Advisors go through vetting and a third-party background check. They take compliance training when they join and every 12 months after that. Before each project, they reconfirm what they will not share.

Clients layer on their own controls too. Extra screening questions, required affirmations, pre-approval of Advisors, chaperoned calls. Guidepoint360 keeps an audit trail and lets compliance teams pull consultation records in real time. “Speed is important in research, but it can never come at the expense of integrity,” Sabouni said.

Gulf investment priorities will keep moving with the routes. My read on his answers is simple. The Gulf story is no longer about how much capital exists. It is about who can tell you what is happening on the ground this week, and prove where the answer came from.

  • 4 min read
Load More Economy News