
Salma Makboul

AI's picks and shovels: the investment case moves down the value chain
After the first wave of enthusiasm for model builders, investors are increasingly trying to isolate the segments of the AI supply chain where scarcity, pricing power and free cash flow may ultimately emerge. After looking at AI through the lens of infrastructure costs, then through the valuations implied by the market's enthusiasm for model builders, the next question becomes more practical for investors: where is economic value actually being captured? To make that question more concrete, we have built a set of representative market baskets designed to benchmark specific segments of the AI value chain. The objective is to break a broad (and sometimes rather imprecise) investment theme into smaller building blocks that can be compared across performance, valuation, margins and financial quality.
27 June 2026 at 12:41 am AEST
What does a $1 trillion valuation actually mean for OpenAI and Anthropic?
Is AI really a productivity revolution? Can companies building it turn its vast (mostly free) usage into vast free cash flow? Artificial intelligence is being sold as a productivity engine for the global economy. It is expected to help companies accelerate research, support customers, assist doctors and lawyers, and reduce the cost of repetitive knowledge work. In the optimistic version of the story, AI allows companies to produce more with fewer resources, lifting margins while reducing inflationary pressure. On the demand side, if AI "threatens" qualified employment, what happens to purchasing power and consumer demand? But before asking whether AI can replace workers, investors need to know whether AI is actually cheaper than the work it claims to replace.
26 June 2026 at 12:54 am AEST

SpaceX, Reflection AI and the Compute Boom: When AI Starts Looking Like Heavy Industry
The latest signal from the artificial-intelligence boom is not a new chatbot, a consumer application or a software feature, but a multibillion-dollar compute agreement that underlines how physical the industry has become. SpaceX's reported $6.3 billion deal with Reflection AI, including roughly $150 million a month for access to compute capacity, has become a useful case study in the economics now sitting behind the sector's growth story. Reflection AI, backed by Nvidia, is trying to secure the processing power needed to compete with leading frontier laboratories, while SpaceX, still better known for rockets, satellites and Starlink, is increasingly being discussed as a supplier of large-scale AI infrastructure. The arrangement has therefore been presented as part of the open-source AI race, but its more important message may be financial rather than strategic. The market reaction was notably restrained. SpaceX shares fell on Monday, according to the Wall Street Journal, suggesting that investors are no longer prepared to treat every AI-related announcement as an automatic positive. The question now being asked is less whether demand for AI exists, and more whether the cost of serving that demand can support the valuations already attached to the sector.
24 June 2026 at 06:21 pm AEST

Hormuz: The Hidden Cost of Dependency
The conflict involving Iran has served one primary purpose: reminding us that globalization still relies on very real and highly vulnerable physical bottlenecks. The Strait of Hormuz measures barely 29 nautical miles at its narrowest point, with two navigable channels of approximately 2 miles each. Some 20m barrels pass through it daily, representing nearly 20% of global liquid fuel consumption and a quarter of the world's maritime oil trade. For liquefied natural gas, the scale is comparable.
19 June 2026 at 01:49 am AEST
Discover our authors
Chief Editor
Journalist
Analyst
Analyst
Analyst
- Stock Market
- Our authors
- Salma Makboul























