💥When to Raise, Biggest Opportunity in Software, Building a VC AI Brain & More
1. Key Themes
Theme: AI inference is becoming the largest software market, and it reshapes software economics
AI inference to overtake databases
Tunguz projects AI inference spending will roughly double the database market within a few years.
"AI inference is projected to reach about $350B by 2027, passing the $190B database market by nearly 2x. Companies paid roughly $25B to run AI models in 2025."
Every application becomes an AI reseller, with usage-based bills
"Tunguz expects every software application to become a reseller of AI, so customer bills follow usage and could far exceed the seat license or base subscription."
Gross margins compress
"Traditional software keeps about 72 cents of every revenue dollar after direct costs. AI usage could take more than half of revenue, so margins fall unless companies cut AI use per task or prices drop."
"The budget line starts to behave like cloud spend, which means tracking cost per unit of use rather than per user."
Theme: When traditional metrics stop predicting outcomes, fundraising timing and founder quality dominate
Hypergrowth has scrambled the benchmarks
"AI accelerated innovation so much that some companies now reach $10M in 18 months and $1B in under two years. When that keeps happening, investors lose the ability to tell normal from exceptional."
Rounds rest on team, not traction
"Rounds worth hundreds of millions have closed at very small revenue, while companies with far more revenue have failed to raise. With copied technology no longer differentiating, nearly every round now rests on the founder and team."
The "Decisive Moment" framework
"A decisive moment is when three things line up: the right business (mostly in the founder's control), the right reason to raise (partly controlled), and an investor with a prepared mind (almost entirely outside the founder's control)."
Theme: AI exposure splits work into routine tasks that get absorbed and judgment that stays human
Routine work is most exposed
"BCG's chart estimates the share of work AI could take on at 45% to 55% in onboarding and KYC (know your customer checks) and 40% to 55% in account servicing, against 20% to 25% in client acquisition and in mid- and back-office work."
Relationship-based advice stays human
"Relationship-based referrals, life-stage coaching and goals, multigenerational strategy, investment decisions, and the trusted advisor relationship are marked human-led, with AI handling only 10% to 20%, so the more tailored the service, the less exposed it is."
Redesign beats bolt-on AI
"Adding AI to existing processes delivers small gains. The post argues firms that redesign their work around AI should cut costs, widen margins, and charge more, while firms with scattered, disconnected data rarely get past early trials."
Theme: VC itself is digitizing, and the real edge comes from custom AI workflows
Efficiency vs. effectiveness
"The move from old-school to productivity VC is a push for efficiency through off-the-shelf tools. The next step, into Workflow and Fullstack builders, is a push for effectiveness through custom AI workflows and in-house data."
Most of the market is still early
"The report places most of the global VC market in the old-school stage, run on spreadsheets and gut feel, while fully autonomous quant VC remains rare, emerging, and contested."
Small teams can reach custom-edge stages
"The median Workflow firm has 7 employees, no engineers, 5 investors, and $151M AUM, so its investors and operators take on the engineering role and build the workflows themselves."
2. Contrarian Perspectives
Emerging managers, not megafunds, are where outlier discovery happens Conventional wisdom says the largest platforms see the best deals. Gray's Dealroom analysis argues the opposite.
"Analyzing the data, Gray finds that the average emerging manager backs more than twice as many outlier profiles per investment as a megafund."
Supporting evidence is that megafunds are crowding into the same category:
"AI makes up 82.7% to 89.5% of five megafunds' technology-labeled Seed rounds in 2023 to 2025, while 11 emerging managers range from 15.4% to 100%. Gray sees megafunds crowding into the most legible category, which leaves discovery of companies outside it to emerging managers."
A diversification argument also supports a basket of small funds:
"Combining five emerging manager portfolios widens the range of investments significantly, while the same does not hold for the more homogeneous megafunds, so a basket of small funds buys broader exposure."
Strong metrics no longer guarantee a raise, and weak ones no longer block one This cuts against the standard "hit the benchmarks and the round will follow" playbook.
"Rounds worth hundreds of millions have closed at very small revenue, while companies with far more revenue have failed to raise."
The implication is that founders should spend less effort optimizing metrics and more on pre-positioning investors:
"Harris's practical lever is finding prospective investors before fundraising and shaping their thinking in low-pressure conversations months ahead of any pitch."
Building a useful AI agent is the easy part; constraining it is the real work Most AI-agent enthusiasm centers on capability. Schmincke reports the opposite cost structure.
"A useful agent took an afternoon. The rest of the quarter went on deciding what it may touch and proving that boundary holds, and only three of eight agents can write anything."
3. Companies Identified
- Description: Venture capital firm (Simon Schmincke's employer)
- Why mentioned: Case study of a VC building a personal AI operating system
- Quote: "Simon Schmincke at Creandum describes how his personal AI system moved from answering questions to acting every ten minutes, with every outgoing message needing a matching approval."
- Description: Firm where Dan Gray conducted the emerging-manager outlier analysis
- Why mentioned: Source of data-backed research on emerging managers vs. megafunds
- Quote: "Dan Gray at Odin uses Dealroom data to show that emerging managers back more outlier profiles than megafunds."
Dealroom
- Description: Startup and venture data provider
- Why mentioned: Data source for the outlier-profile analysis
- Quote: "Dan Gray at Odin uses Dealroom data to show that emerging managers back more outlier profiles than megafunds."
- Description: Firm where Aaron Harris writes about fundraising timing
- Why mentioned: Source of the "Decisive Moment" framework for raising capital
- Quote: "Aaron Harris at Magid explains that raise timing follows what he calls the Decisive Moment, now that metrics no longer decide rounds."
BCG (Boston Consulting Group)
- Description: Management consultancy
- Why mentioned: Its 2026 Wealth Report quantifies AI's potential impact across wealth management workflows
- Quote: "BCG's chart estimates the share of work AI could take on at 45% to 55% in onboarding and KYC... and 40% to 55% in account servicing."
- Description: Publisher that shared the BCG wealth management analysis
- Why mentioned: Surfaced the AI-in-wealth-management findings
- Quote: "Mr Family Office shares BCG's 2026 Wealth Report map of how much work AI could take on across wealth management."
Granola
- Description: AI meeting-notes tool (newsletter sponsor)
- Why mentioned: Sponsored placement; differentiates by working without a meeting bot
- Quote: "There are no meeting bots. Nothing weird joins your call. Granola transcribes directly from your computer or phone audio."
- Description: Publisher of the newsletter and the 2026 DDVC Landscape Report
- Why mentioned: Maps VC firms across five stages of digitization
- Quote: "Our 2026 DDVC Landscape Report maps VC firms along five stages of digitization, from manual spreadsheets to fully autonomous investing, with Workflow and Fullstack builders at the DDVC core."
- Description: Venture firm associated with Tomasz Tunguz, named in the subtitle
- Why mentioned: Affiliation of the AI-inference market thesis
- Quote: "New insights from Creandum, Odin, Theory Ventures & more"
4. People Identified
Tomasz Tunguz
- Description: Venture investor (Theory Ventures) and software-market analyst
- Why mentioned: Argues AI inference will pass databases as software's most important market
- Quote: "Tomasz Tunguz argues that AI inference is about to pass databases as the most important market in software, and that every application becomes a reseller of AI."
Dan Gray
- Description: Investor/analyst at Odin
- Why mentioned: Data analysis showing emerging managers back more outlier profiles than megafunds
- Quote: "Gray finds that the average emerging manager backs more than twice as many outlier profiles per investment as a megafund."
Aaron Harris
- Description: Writer/investor at Magid
- Why mentioned: Developed the "Decisive Moment" model for timing a raise
- Quote: "Aaron Harris at Magid explains that raise timing follows what he calls the Decisive Moment."
Simon Schmincke
- Description: Investor at Creandum
- Why mentioned: Built a personal agentic AI system with strict approval boundaries
- Quote: "A fixed loop runs every ten minutes around the clock, about 52,560 times a year, on a Mac mini with 107 active jobs and 47 databases."
Andre Retterath
- Description: Author of Data Driven VC
- Why mentioned: Newsletter author; runs a live experiment automating the VC job with AI
- Quote: "Join 1,991+ investors in our free Slack group as we automate our VC job end-to-end with AI. Live experiment. Full transparency."
5. Operating Insights
Track AI cost per unit of use, not per seat
If you build on or buy AI, expect costs to scale with usage.
"For firms building on AI models or buying from AI companies, costs are likely to follow usage, so a heavily used tool can cost far more than its seat license suggests. The budget line starts to behave like cloud spend, which means tracking cost per unit of use rather than per user."
Build investor relationships months before you raise
Since the investor's "prepared mind" is the factor founders control least, shift effort earlier.
"Harris's practical lever is finding prospective investors before fundraising and shaping their thinking in low-pressure conversations months ahead of any pitch. Founders who keep tracking what investors currently think are better placed to choose their moment to raise."
Budget time for agent boundaries before capabilities
When deploying AI agents that act on your behalf, put safeguards outside the model.
"For VCs letting agents act on their behalf, the safeguards that matter sit outside the model: what each agent can read, what it can write, and what can leave the machine... starting with his advice to plan time for boundaries before capabilities."
6. Overlooked Insights
A people-centric memory layer is the foundation Schmincke would build first
The system's world model is a relationship graph with sourced facts, not a generic knowledge base.
"The world model holds what the system has worked out about people from Simon's own correspondence, such as their tastes, role changes, and what he promised them. It covers 20,589 people, about 1,300 of whom carry 5,730 facts, each stored with its source sentence. Simon would build it first if he started again."
Even a sophisticated agent stack has unverified repairs
The system can't confirm its own fixes worked, a gap that matters for anyone trusting autonomous operations.
"Repair verification is not built yet, so the morning report can only say a repair was 'tried'."