Technology. Money. People.
28 Sep – 4 Oct 2026
This week’s lead story
Supplier.Financier.Rival.
Google competes with Claude while selling compute to Anthropic. Broadcom supplies the chips and offers financing to help pay for them.
Rivalry and shared interests can coexist. Separate the roles to see what changes.
Supplier. Google sells compute to Anthropic. Using Claude can also bring revenue to Google.
Shareholder. Google owns a stake in Anthropic. Growth at a rival can also increase the value of that holding.
Rival. Google’s Gemini and Anthropic’s Claude compete in the model market. A shared financial interest does not remove that rivalry.
01 / ANTHROPIC
Buying compute years before it is used.
About 80% of Anthropic’s compute commitments do not flex with usage. The company is reserving capacity for years. Its customers can still switch services or cut their spending while its own bills keep coming.
According to a draft IPO prospectus seen by Reuters, Anthropic’s revenue grew roughly twelvefold in 2025 to $4.6bn. Compute costs tripled to $7.33bn, and its operating loss exceeded $8bn. Reuters [1]
The net loss was $42bn, but roughly $34bn came from a non-cash financing item. It does not represent an equivalent cash outflow. Reuters [1]
Anthropic’s compute commitments
Terms determine flexibility.
The customer can leave. The capacity bill remains.
Google and Amazon are also Anthropic shareholders and distributors of Claude. Combined purchase commitments to them total about €197bn. Google is paid even if usage falls below the agreed minimum; Amazon has a similar provision. Reuters [2]
Broadcom ↔ Anthropic
Broadcom finances its own customer.
A financing commitment of up to about €37.4bn ($42bn) covers a third of a five-year TPU lease worth about €111.5bn ($125.2bn). TPUs are AI chips developed by Google with Broadcom. The debt can convert into equity.
A non-cancellable contract creates an obligation to pay; it does not make the customer solvent. According to the filing, a default could also accelerate lease payments and restrict access to financing. Reuters [3]
Company relationships
Partners can be rivals, too.
Growth can benefit a rival, too.
Google sells Anthropic compute, holds an equity stake and distributes Claude. Its Gemini models compete with Claude. Who benefits also depends on where customers run their work.
Reuters [2]Broadcom
The tenant also receives financing.
Broadcom’s financing commitment would help Anthropic pay for leased TPUs. The hardware lease and financing commitment are separate agreements. A commitment is not money already borrowed.
Reuters [3]Amazon
A minimum purchase still applies in quiet periods.
Amazon supplies compute, owns a stake and distributes Claude. Its minimum-purchase provision resembles Google’s: agreed spending cannot freely shrink with usage.
Reuters [2]xAI
Not every contract locks the buyer in.
The xAI capacity agreement can largely be cancelled with 90 days’ notice. That differs from Google’s and Amazon’s binding minimum purchases.
Reuters [2]The terms of competition
When a rival’s success serves your interests.
Google has a reason to compete with Claude in the model market. As a compute supplier and shareholder, it may also have a reason to welcome Anthropic’s growth. Broadcom would finance the customer whose lease payments provide its revenue. Rivalry and a shared financial interest can coexist.
An ownership chart alone cannot tell us how competitive a market is. What matters is how many independent routes a customer has to compute, distribution and their own data. Closely connected companies can still compete on price and products.
If partnerships speed up construction and lower costs, customers may benefit. If switching suppliers also requires replacing tools, contracts and working practices, more of that benefit may stay within the supply chain. Evidence of new entrants reaching customers—and customers actually moving their workloads—would tell us more.
Follow the implications
Change the assumption. What follows?
Three questions raised by this week’s news. These paths explore possible consequences, not forecasts.
Who does the supply chain work for?
Google is simultaneously a compute supplier, shareholder, distributor and model competitor. Source [2]Source [3]
Customers can switch
Partnerships accelerate capacity building.
A rival offer is a practical option for the customer.
Cost savings could reach customers through lower prices or better service.
What would we look for? Evidence of portable workflows, viable alternatives and improving customer terms would support this path.
Switching becomes difficult
Compute, tools and distribution become tied to one environment.
Changing the model is not enough to move the customer’s work.
Efficiency gains could strengthen the supplier’s bargaining position.
What would we look for? Rising migration costs or barriers to entry would support this path. Equity ties alone do not establish it.
Where does the time saved go?
Barclays says Claude sorts and routes messages. Reported usage does not yet measure the outcome of the work. Source [15]
Room for judgement
Less time goes into routine sorting.
People have time to understand unusual cases.
Customers could reach the right resolution in fewer rounds.
What would we look for? Look at resolution time, reopened cases and staff experience. Throughput alone is insufficient.
A harder queue
Easy cases are handled automatically.
Difficult situations accumulate in the human queue.
Work could become more demanding even as the number of cases falls.
What would we look for? This is not a measured Barclays outcome. Assess it through task difficulty, time for judgement and available support.
With more options, who chooses the objective?
GPT-Synopsys aims to connect a model with design tools in an iterative workflow. An engineer evaluates the result. Source [10]Source [11]
Judgement gains scope
The model tests more proposals.
The engineer weighs energy, speed and size.
Expertise could become more valuable in choosing between options than in generating them.
What would we look for? Better outcomes across the full design process, including checks and corrections, would support this path.
Defaults become objectives
The tool offers a ready-made optimisation approach.
Accepting its defaults saves time.
Decision-making power could quietly shift towards the toolmaker.
What would we look for? Ask whether users can change objectives and understand the trade-offs. Faster computation does not settle this on its own.
What stays with us?
Who can still choose differently?
The same ties can accelerate investment and make changing direction harder. This week’s news does not yet settle which effect will prevail. Customer terms, new competitors’ access to the market and who retains the gains from more efficient work would help answer that.
We can want technology to advance and still examine how it distributes power. If a system becomes more efficient, should success also mean that people and businesses find it easier to decide how they use it?
What should we be able to change—the model, the supplier, or the way we work?