What the VCs are talking about this week: 3–9 Nov 2026
Two companies Stephen Millard works with are brilliant at building pipeline. What they’re struggling with is getting customers live. It stuck with me most from this week’s VC newsletters. Building software is getting cheap, and nearly everything around it is getting harder.

1️⃣ Notion Capital, Stephen Millard: Winning the deal isn’t the finish line
“A company that keeps winning customers it cannot activate isn’t building a flywheel.”
Activating means getting a customer live and seeing real value. He says founders hire sales teams and senior leaders before they’ve learned how their customers buy.
2️⃣ Point Nine, Christoph Janz: AI rebuilt his 1990s game in 34 hours
“almost any existing piece of software can be reverse-engineered and replicated by AI in a matter of days.”
He and a friend spent about two years writing the game. Claude rebuilt it from the old program file alone, with no source code.
3️⃣ Theory Ventures, Tomasz Tunguz: Running AI is getting more expensive
“When inference becomes more expensive, off-the-shelf software (with AI) can be less expensive than bespoke built & maintained internal software.”
Inference is the cost of running AI every time you use it. He says the hourly cost of renting GPUs (the chips AI runs on) has doubled in six months, so building your own tools isn’t always cheaper.
4️⃣ 20VC, Harry Stebbings with Erik Bernhardsson: The 90% software margin is gone
“Almost no company today has the 90%-plus software margins common in Silicon Valley a decade ago.”
The Modal CEO says every layer, from apps to models to cloud to NVIDIA, takes its own 20% to 30% cut.
5️⃣ Boldstart, Ed Sim: Startup funding is slowing down
“What was getting funded with $100M a month ago is now taking more time and a lot more questions.”
He’s seeing rounds take longer. His advice to founders is to have more than one plan, and to show what each smaller amount of money would buy.
6️⃣ a16z, Moses Sternstein: Investors reward Apple for skipping AI spend
“The market is obviously rewarding Apple for not investing so much (or really anything) in AI.”
Apple now spends less on things like data centres than it did before the AI boom, while the big cloud firms spend roughly 6 to 9 times more than in 2023.
So AI makes building cheaper than ever, but the money, the margins and the running costs are all getting tighter.
For those of us in GTM, that puts more weight on what happens after the sale. If a product is easier to copy, getting customers live and seeing value is what keeps them.



