Weekly Signal · November 6, 2026

November 6, 2026: AI stocks, election shock and what to do next

Week ending November 6, 2026 · Rob Tyrie · Ironstone Advisory

Markets through the November 5 U.S. close. Election calls and sources checked November 6, before the U.S. market opened.

Lead photograph: the U.S. Capitol at dusk, September 3, 2013. Martin Falbisoner, via Wikimedia Commons, CC BY-SA 3.0. Earlier context photograph, not this week's election. Resized, not cropped. No endorsement is implied.

Quite a week. AI companies reported serious growth, voters changed the political arithmetic, and then the market took a very large breath in. The money, the machines and the permission to build them are connected. So are the decisions facing a Canadian founder, an insurer buying software and a services firm trying to make agentic systems useful. Five readings. Then some work to do.

1. The reaction is real. The explanation needs care.

On November 5, the Nasdaq Composite fell 5.5%, the S&P 500 fell 4.4%, Nvidia lost 9.3% and AMD lost 17.3%, according to Cboe's closing commentary. Cboe linked investor anxiety to possible constraints on data-centre construction following the elections. That is a market interpretation, not proof of a single cause or evidence that a nationwide moratorium has become law.

For a business, a sell-off can reach the operating plan through financing, customers' budgets and suppliers' commitments. I would check the assumptions behind the next twelve months: cash runway, funding dependencies, cloud concentration and the sales cycle. Keep the evidence separate from the reaction. A red screen is remarkably good at making everybody sound certain.

Read Cboe's November 5 market commentary

2. AMD's growth still needs somewhere to run

AMD reported third-quarter revenue of US$13.4 billion on November 3, up 45% from a year earlier. Its data-centre segment brought in US$8.5 billion, up 97%. Those are reported operating results, alongside a market that subsequently repriced the shares sharply.

Chip demand meets electricity, construction, customer deployment and the price somebody will pay for the finished service. That is business architecture. For an agentic-systems project, test the economics against the capacity you can actually contract for. Include inference, integration, review and recovery. Watch delivery commitments and utilisation as well as the next accelerator announcement. Buying more computational juice is easy to explain. Getting a profitable result takes more work.

Read AMD's November 3 results

3. Palantir puts the customer into the AI argument

Palantir's November 2 business update reported US$984 million in U.S. commercial revenue for the quarter, up 148% year over year. The company said that business exceeded its U.S. government revenue for the first time. These are company-reported results, not an independent measurement of every customer's return.

Still, this is worth studying if you build enterprise software or run professional services. Customers are paying for something that crosses data, software and how people work. For our own projects, I want a named customer problem, a working baseline and evidence of the result after deployment. Watch renewals, expansion and the cost of delivery. A demonstration gets attention. A customer who comes back gives you a business.

Read Palantir's third-quarter business update

4. The election changes the questions businesses will face

The Associated Press called Democratic control of the U.S. Senate on November 3 and the House early on November 4. These are AP's election calls, not a claim that certification is complete or that new policies have already taken effect.

My reading: prepare for changed oversight priorities and contested implementation, without writing campaign promises into the forecast as settled law. Insurance, financial services and cross-border software all need clear responsibility for data, decisions and contracts. Map the federal, state and local dependencies in the projects you sell. Keep a record of what has actually changed, who owns the response and when it matters. There better be a log. Political uncertainty is a poor reason to abandon operational discipline.

Read AP's account of its calls for control of Congress

5. Canada's budget brings the execution question home

Canada's November 5 budget release proposes investment and productivity measures, support for commercialisation, trade diversification and an electricity strategy. Its stated ambition includes doubling grid capacity by 2050. Proposals and ambitions still have to become usable programmes, rules and projects.

Canadian software businesses should examine the published eligibility and implementation details as they emerge, rather than putting a budget headline into booked revenue. This touches the whole company: financing, architecture, partnerships and go-to-market. Build the commercial case for a customer who will pay. Then determine whether a programme improves its economics. And keep a Canadian and a U.S. scenario in the plan. An election result does not, by itself, remove a trade barrier.

Read Canada's November 5 budget announcement

What I would do with this

Put three things on one page:

• What the business depends on: capital, customers, compute, power and permission.

• Which assumptions changed this week, with the source and the owner of the response.

• One action to take now, one experiment to run and one decision to defer until there is evidence.

Then use the Creation Model: Ideate, Think, Make, Do, Judge, Decide, Execute, Archive. Judge what happened. Keep the record. Return to it when the next confident forecast arrives. This week's overlap is the useful signal: technical capability, political permission and commercial demand have to meet in an operation that works. Execution is everything. Profitable revenue is a fairly direct review of the argument.

Explore Rob's Substack for deeper writing and subscription options.

Browse the Weekly Signal Archive.

AI-assisted research and drafting for Rob Tyrie's Weekly Signal. Business analysis, not a recommendation to buy or sell securities.