Nov 1 – Nov 7, 2026
Not investment advice — these are resources for learning and forming your own view.
Deep dive — one theory Organization
Why Nvidia's robot hand doesn't end with 'buying it in': the price of costs a contract cannot capture
- Origin
- It began with Ronald Coase's 1937 paper ("The Nature of the Firm") and was built into a system by Oliver Williamson in the 1970s and 1980s.
- In short
- Asked why firms exist at all if markets are so efficient, Coase answered that using the market also has a cost: the cost of searching, bargaining, writing contracts and monitoring whether they are honoured. When that cost exceeds the cost of running an organisation directly, firms bring the activity in-house instead of using the market. Williamson added conditions: the larger the investments useful only within the relationship, and the less the future can be fully written into a contract, the stronger the pull toward bringing it in-house.
- Limits
- Transaction costs are hard to measure in advance, so they are easy to fit to outcomes after the fact. The theory is also criticised for underrating the efficiency of cooperative networks that run on trust and long relationships.
Transaction cost economics is worth rereading now because companies can no longer decide whether to buy cheaply or make in-house from a price list alone. Coase and Williamson described three conditions: large investments that are useful only within a particular relationship, a future that cannot be fully written into a contract, and the risk that the other party will later change its terms. By coincidence, all three showed up in this week's news at once. In the same week, the US warned China to honour its rare-earth supply commitments under the Busan agreement immediately. That episode showed that component supply is not a purely commercial contract; it can become a hostage in negotiations between states. If a supplier that is 10% cheaper can one day stop shipping altogether, that 10% is less a discount than an insurance premium left unpaid.
News that Nvidia's humanoid robot system uses, as a core component, a mechanical hand made by a China-linked start-up fits this framework exactly. A robot hand is not a general-purpose part you can plug in anywhere. It is designed together with a specific actuation method and control software, making it a classic relationship-specific asset in which both sides pile up investments useful only to each other. That design experience cannot move to another supplier overnight. And because no contract can specify in advance when export controls or retaliation will hit which items, Nvidia holds fewer cards than it might seem the moment the supplier reprices or resets delivery dates. The news is usually consumed as a matter of pride, that 'American technology still depends on Chinese manufacturing'. Through this lens it is not about pride but about bargaining power, and the price of that bargaining power is ultimately paid in organisational design costs such as insourcing, equity stakes or dual sourcing.
The generalisation for alumni is that the number to compare when choosing a supplier is not the unit price but 'the unit price plus the value of risks a contract cannot block'. If investments useful only within the relationship are large and future situations cannot all be tied down in clauses, outsourcing is more expensive than it looks. Conversely, bringing a standardised part with many suppliers in-house only makes the organisation heavier. The theory has a weakness, though: transaction costs are easy to fit after the fact. In practice it is safer to pin down a number in advance, along the lines of 'if this supplier cuts us off tomorrow, what percentage of our revenue stops, and for how many months?' In next quarter's supply-chain review, try writing the number of months needed to secure alternative sourcing next to each item's unit price.
Source story: CNBC's The China Connection newsletter: Nvidia's humanoid push relies on China-made robot hands
Cases by layer 6
US economy outstrips forecasts to add 165,000 jobs in October
The US economy added 165,000 jobs in October, beating market expectations. The solid payroll data pushed expectations back toward the view that the Fed will find it hard to cut rates quickly. That put upward pressure on the dollar and Treasury yields and downward pressure on growth-stock valuations.
For this lens to work, slack in the labour market must be observable. The market immediately read the stronger-than-expected job gains as a sign that this slack is shrinking, and that reading showed up directly in rates and the dollar. Inflation expectations not being wholly reset in a short period also fits the conditions. Rate expectations moved up even though oil plunged the same week, showing that people are locating the source of inflation in the labour market.
This news is usually read as simple good news: 'the economy is strong, so corporate earnings will be strong too'. Bring out this curve, however, and you see the path by which strong hiring translates into wages, then prices, then interest rates. That explains why growth stocks get hit on good data. It also prompts a check on the conditions under which the curve breaks down: when inflation comes not from demand but from oil or component shortages.
Through this lens, this week was a rare combination of hot demand and cooling supply. With oil falling on peace hopes, supply-driven price pressure eased, leaving room for inflation not to follow strong hiring right away. On that reading, this week's jump in rate expectations was somewhat overdone. Conversely, if more items see supply-side price rises, as with the memory shortage, the same payroll number takes on an entirely different meaning.
Takeaway — Do not prejudge the direction of rates from a single payroll number; first separate whether inflation is coming from the labour market or from supply chains. Get this distinction wrong and your portfolio's entire duration choice will be wrong.
Financial Times
US warns China it must ‘immediately’ meet terms of Busan trade deal
The US warned China to deliver immediately on the expanded supply of rare earths and key items it promised in the Busan trade agreement. Although an agreement exists, delays in implementation have increased supply uncertainty for critical resources. Volatility in related commodity prices and concern about rising global supply-chain costs rose together.
There are effectively two players, each player's choices feed straight into the other's gains and losses, and both know the other's cards, tariffs and export controls, so the payoff structure is reasonably legible. The very fact that implementation is lagging despite a signed agreement, and that a warning followed, shows this is not a one-off deal but a repeated game in which each side plays while watching the other's next move.
This news is usually read as a question of good faith: 'China does not keep its promises'. Through the lens of game theory, the issue is not good faith but structure. Both sides would be better off if both kept their promises, but as long as each suspects the other may defect first, it is rational for each to keep hold of its supply card. The incentive to break the agreement was growing from the moment it was signed.
In this framework, Friedland's push, reported the same week, for a US Defense Department-backed rare-earth mine in Australia is not simple investment news but a move that rewrites the game's payoffs. The moment an alternative supplier actually comes on stream, the value of China's delay card falls, so the outcome of negotiations is decided not by what is said at the table but by which alternatives stand away from it. Because the lag before that mine produces anything is long, however, the equilibrium is likely to stay, for now, with both sides holding their cards and waiting each other out.
Takeaway — What makes a counterparty's promise credible is not persuasion but the outside alternatives you hold. Whether in procurement or with customers, before negotiating, count your alternatives and how long it would take for them to actually come on stream.
Financial Times
Qualcomm stock drops 7% on weak earnings forecast as memory shortage weighs on smartphone sales
Qualcomm's third-quarter results beat market expectations, but it issued a weak outlook, saying memory shortages would hurt smartphone sales, and its shares fell 7%. In the same week Apple also warned of rising costs from a shortage of memory chips.
The dependency structure, in which one stage's output is the next stage's input, is clear. If memory runs short, finished smartphones cannot be built no matter how well chipsets sell, so total shipments are pinned to a single point. Expanding memory capacity takes years of fab construction and process conversion, making it hard to substitute or increase in the short term, which also fits the conditions exactly.
This news is usually read as a demand story: 'the chip cycle has turned and smartphone demand has slowed'. Through the lens of constraints, though, what has turned is not demand but the channel. Set it beside AMD's data-centre revenue nearly doubling the same week, and the picture is that AI servers are absorbing limited memory capacity first and mobile is being pushed to the back of the queue.
Through this lens, no amount of improvement in mobile chip design adds a single unit to total shipments, and the profit pool shifts toward memory, which holds the bottleneck. This week's decline is therefore less a signal that profits across the tech sector are shrinking than a signal that profits are being redistributed within the sector. The point at which memory prices turn down will be the point at which the outlook for the mobile chain recovers. Remember, too, that this prescription wobbles if the bottleneck moves.
Takeaway — When reading supply-chain news, ask 'who holds the bottleneck right now?' before 'has demand fallen?'. In the same cycle, the margins of whoever holds the bottleneck and of whoever is blocked move in opposite directions.
CNBC Markets
AI-Generated Movie Heading Theaters
A 30-person team used AI to make the animated film 'Gods Don't Give Gifts', which is heading for a theatrical release. AI has moved beyond experiments into real commercial production, and this is putting fundamental downward pressure on incumbent studios' staffing structures and costing methods.
What matters is that the new entrant is still inferior on the established performance axis. Its polish falls short of a big studio's handcrafted quality, but on other axes, crew size, production time and cost, it is incomparably better. Incumbents' organisations and contracts are built to extract margins from big-budget productions, so they have almost no incentive to enter the low-budget corner. Meanwhile, generative models are improving in quality far faster than the quality audiences demand is rising.
This news is usually read as a labour issue: 'AI is taking creators' jobs'. This lens turns the question toward the incumbents. It reveals the paradox that big studios are not slow because they are foolish; they left the low-budget space empty because they stayed loyal to their profitable blockbuster customers, and that very rationality is what makes them slow.
Read through this framework, the first thing to be eroded is not blockbusters but mid-budget content. Once the era in which production cost was the core of competition ends, the axis of competition shifts toward intellectual property, distribution windows and audience data. It is all too easy, however, to label something 'disruptive innovation' after the fact, so whether this path has really opened must be confirmed by whether the next few films make box-office money.
Takeaway — Check every quarter who is now entering the 'unprofitable low-end corner' of your own industry. The decision to abandon that corner always looks rational at the time.
Bloomberg Markets
Apple’s Growth Machine Is Slowing. High Prices Might Not Save It.
Apple's shares slipped after new CEO John Ternus warned that margins could weaken in coming quarters. Analysts added that with memory-chip shortages raising costs, a high-price strategy alone may not be enough to stop slowing growth.
The lens's premise holds because the live question is whether costs will actually pass through to consumer prices. There is a clear input shock in rising memory prices, and the company must choose whether to add it to prices or absorb it in margins. Longer replacement cycles mean more consumers who can simply not buy now, and a wide range of mid- and low-priced alternatives exists, so it is possible to observe how much volume falls when prices rise.
This news is usually read as an accounting identity: 'costs went up, so margins shrink'. Through the lens of elasticity, a margin squeeze is not fate but the result of a choice. The real question is who bears the higher cost, and the answer depends not on brand strength but on how much room customers have to delay their purchases.
In this framework, Apple has two options: raise prices and lose volume, or hold prices and cut margins. In a mature market with longer replacement cycles, it is more natural to read the room for pass-through as narrower than in the past. Services revenue from the installed base acts as a buffer, though, so the damage to overall profitability will likely be smaller than the damage to device margins. Note also that elasticity figures measured in the past cannot simply be reused for a structural cost change in which prices jump across the board.
Takeaway — When assessing a company hit by a cost shock, look not at the brand's reputation but at what percentage of volume it lost after its last price increase. Pricing power is an asset confirmed by data, not by feel.
WSJ Markets
His Job Is to Turn Walmart’s Customer Data Into Profit
Walmart put an organisation and strategy for turning customer data into revenue front and centre. The plan is to keep its low-price principle while using data to build a new path to growth, and the market took it as a sign of better long-term profitability for the retail sector.
The unit of analysis is the individual firm, not the industry, and the key point is that Walmart's resources differ in kind from its competitors'. The real purchase records that build up every day across its national store network and payments are not something you can buy on the market. They are the product of scale and long years of operation together, so a latecomer investing the same amount could not replicate them within a few years.
This news is usually read as a diversification story: 'retailers are getting into advertising too'. The resource lens explains why the business pays only when this particular company runs it. Advertisers pay not for Walmart's sales force but for the scarcity of data that confirms whether ads led to actual purchases.
In this framework, the data business reads as a complement to the low-price strategy, not a substitute. If margins are earned from data and advertising, prices can be kept even lower; lower prices bring more traffic, which builds more data, and a virtuous circle turns. The lens's warning that yesterday's strength becomes tomorrow's shackle still applies, however: a data structure tied to a physical store network risks slowing the transition when consumer channels shift again.
Takeaway — List the resources your company holds that competitors cannot buy with money. If that list is empty, your current profitability is likely resting on luck rather than structure.
WSJ Markets
Other signals this week 18
- A Berkshire Hathaway Electric Utility Wants a Historic Rate Hike in California WSJ Markets
- Mississippi Senate election: Democrat Scott Colom narrowly leads Republican Cindy Hyde-Smith CNBC Markets
- Democrats flip dozens of GOP seats. Republicans haven't flipped a single one CNBC Markets
- First polls close in high-stakes vote for control of US Congress Financial Times
- Supreme Court rejects DOJ's bid to block California from counting disputed ballots CNBC Markets
- Judge dismisses Alaska Senate election lawsuit; DOJ vows appeal CNBC Markets
- Tusk vows to press on with overhaul of Poland’s judiciary Financial Times
- UOB Profits Miss Estimates as Lower Trading Income Dampens Gains Bloomberg Markets
- Turkey Appoints New Top Market Regulator in Fund Crisis Response Bloomberg Markets
- Colombia at Risk of Default Without a Fix, Restrepo Tells Semana Bloomberg Markets
- Billionaire Friedland to Build Major US-Backed Rare Earths Mine Bloomberg Markets
- Latest Oil Market News and Analysis for Nov. 6 Bloomberg Markets
- A 'demand decade' ahead for energy? BP, Baker Hughes CEOs weigh the investment outlook CNBC Markets
- Latest Oil Market News and Analysis for Nov. 4 Bloomberg Markets
- Latest Oil Market News and Analysis for Nov. 3 Bloomberg Markets
Based on 171 items over 7 days