Skip to main content

icnlive

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

© 2026 ICN.LIVE
  • OpenAI

Sam Altman calls an AI warning and holds the labs to it

Sam Altman calls an AI warning at a moment when the people who build the technology, rather than those who regulate it,

Trending AI

Explore Section:

TECH
AI
Capital
Politics
Banking
Health
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
SDAIA Academy Summer Bootcamps
  • Data Centers

SDAIA Academy summer bootcamps return this August with 18 specialized training programs in data and artificial intelligence. The Saudi Data and Artificial Intelligence Authority announced the lineup on July 20 under what it calls the future summer initiative. Two groups are in scope. Specialists already working in the field and people trying to break into it.

Think of it as a compressed apprenticeship. Rather than a semester of theory, participants spend the days building.

What the 18 tracks cover

The curriculum leans technical. Programs run across generative programming, AI agent systems engineering, AI agent development, and modern data engineering for AI applications. Computer vision for developers is on the list. So are prompt engineering, responsible AI use, and applying artificial intelligence to raise productivity and tighten business processes.

Those titles map closely to what companies are hiring for right now. An AI agent that books, checks, and escalates on its own needs someone who can wire it together and keep it from going off the rails. That is a job, not a concept.

Data governance gets equal billing.  A second cluster of programs handles data management, governance, quality, and integration. It sounds less glamorous than agent engineering. It usually decides whether the agent works at all. A model reads whatever you feed it. Feed it duplicated records and stale fields, and the output looks confident and wrong. SDAIA has spent years building the rules around national data use, so training people to apply those rules is part of the same project.

How the SDAIA Academy summer bootcamps run

The format is intensive. Short, dense, hands-on. The academy focuses on practical skills, smart solution development, application design, and best practices in responsible AI deployment and data management. Participants are meant to walk out with work they built rather than notes they will lose.

That design says something about the hiring market. Shipping an AI agent or repairing a broken pipeline is demonstrable. Reading about either is not.

Registration and the AthkaX platform

Registration for the SDAIA Academy summer bootcamps runs through the AthkaX platform at athkax.sdaia.gov.sa. AthkaX is SDAIA’s national platform for capacity building in data and AI, launched during the International Conference on Data and AI Capacity Building. Program details and schedules are on the same platform.

Where the bootcamps fit in Saudi Vision 2030

August is the second wave of the season. In early July, SDAIA Academy launched eight training programs and bootcamps under a July summer initiative, mixing professional tracks with beginner courses on prompt engineering and AI at work. Eighteen bootcamps are more than double that count, in a single month.

The SDAIA training bootcamps sit within the Year of Artificial Intelligence and the workforce ambitions of Saudi Vision 2030, which calls for a globally competitive digital workforce. AthkaX connects to the Human Capability Development Program, one of the Vision 2030 programs.

AI training programs in Saudi Arabia have turned into a steady pipeline rather than scattered one-off events. Over the past year, SDAIA Academy has run a quantum computing camp and a professional program in large language models built toward NVIDIA certification. The SDAIA Academy summer bootcamps extend that cadence into the summer months, when students and job seekers have time to commit.

If a move into data or AI work is on your list, August is the window.

Saudi Arabia’s AI market, 2025–2032

saudi-ai-market-chart

 

Three houses looked at the same market and came back with numbers that barely overlap. Ken Research starts Saudi AI at $4.3 billion in 2025 and takes it to $19.36 billion by 2031. A second forecast begins higher, at $5.2 billion, but grows more slowly and lands at $14.3 billion. A third starts at just $2.14 billion and nearly matches the top line by 2032.

The base year is where they disagree most: the highest estimate is more than double the lowest. That gap is a definitional problem, not a forecasting one. What all three agree on is direction, and Saudi enterprise spending backs it up.

  • 4 min read
Load More Economy News