This issue of Discover with Jane is dedicated to what the venture industry thinks, does, and predicts. And as usual, I’ve pinpointed some trends you can ride!
VC opinions, political agenda, and stewardship
What’s going on:
Sequoia’s handover: Roelof Botha steps aside as steward; Alfred Lin & Pat Grady become co-stewards. Lin has been leading early stage, Grady on growth, and the structure with co-stewards has worked for Sequoia in the past. The move comes after a turbulent stretch, where Sequoia wrote down over 200 M$ after FTX crypto exchange blew up, and split operations with its Chinese and Indian subsidiaries. However, also returned 50 B$ to its LPs! Plus announced a new 950 M$ fund recently (I wrote about it in a past issue).
Bothas stepping down happened soon after Sequoia COO resignation, allegedly related to islamophobic posts of one of their other partners. Sequoia has a policy of “institutional neutrality” - partners can express their own opinions, separately from the firm’s.
Botha declined to comment extensively on the controversy, but said of Maguire: “I think he has made it clear what he stands for, and there’s a particular set of founders for whom it is very appealing that he’s been as firm in his opinion. Does it come with tradeoffs? Yes, it does.”
a16z’s DEI reset: Andreessen Horowitz paused TxO, its program for under-represented founders, and laid off staff tied to it. Reporting frames this as part of a broader DEI retrenchment and a consolidation around Speedrun as the primary intake channel. TechCrunch
Takeaway: In a politicized climate, VCs are bound to be careful, to maintain the reputation of the firms, and the institutional capital management approach, with its diligence and care.
But I’m thinking: doesn’t mainstream thinking hurt returns?
Market opportunity map — Elad Gil’s “won” niches vs “wide open”
What’s essentially “won”:
Foundational models: a short list (OpenAI, Anthropic, Google, maybe xAI/Meta/Mistral). TechCrunch
AI coding: Cursor leadership, and Devin (Cognition), with Lovable, and many other players, including model players like Claude Code and IDE players like Replit.TechCrunch
Medical transcription: Abridge front-runner; Ambience among notable players. TechCrunch
AI customer support: Decagon (Gil-backed) and Sierra plus incumbent suites layering in AI. TechCrunch
What’s still wide open (Gil’s list) and who’s building there:
AI security operations. Examples: Protect AI (MLSecOps suite), HiddenLayer (model attack defense), Noma Security (prompt-injection & pipeline risk), Corridor (SecOps with Alex Stamos joining). These illustrate the fragmentation—and the room for a platform to emerge. Axios+3TechCrunch+3TechCrunch+3
Accounting — next-gen ledgers and close automation are early. Examples: Quanta (Accel-led seed for AI accounting), Puzzle (AI GL & close), and Rillet (Reuters-covered Series B led by a16z/ICONIQ). Signal: adoption is real, but no de-facto standard yet. TechCrunch+2Venturebeat+2
Fintech tooling — decisioning, risk ops, and compliance agents. Examples: Taktile (automated decision workflows), Rulebase (finserv QA/compliance co-worker). Expect incumbent suites to buy or build here; greenfield remains. TechCrunch+1
Gil’s caution on “false signal”: enterprise trials can inflate early AI revenue—stickiness and post-pilot expansion separate durable winners from hype. TechCrunch
Policy shock absorber — Khosla’s provocation
At Disrupt, Vinod Khosla floated the U.S. taking a 10% stake in all public companies as a way to redistribute AI-era gains—riffing on the government’s Intel stake. Regardless of feasibility, it flags the scale of interventions major VCs now contemplate in an AGI trajectory. TechCrunch
And an opportunity for private businesses to profit from those grants: as companies like OpenAI and Meta are struggling to generate revenue, even with outsized investment and top talent hiring.
Techcrunch Disrupt Battlefield favored ‘moving parts’ of the real world, not AI
Winner: Glīd (pronounced Glide): hybrid-electric yard vehicles + orchestration software to move 20-ft containers straight to rail.
Runner-up: Nephrogen—AI-guided kidney-cell delivery for gene-editing medicines, claiming 100× efficiency vs FDA-approved vehicles.
Looks like a signal: hard tech solving real world problems with a high cost (and path to profitability) is back at center stage. No everything is AI..TechCrunch
🌐 Community offer
Scouts & analysts — join our invite-only deal network
Founders (global & immigrant) — want to be featured or intro’d to relevant VCs?
Submit here: https://forms.gle/ZQ2HR5BKgMTA5E4t7.
Let’s continue building!
And let our stories bring us into the land of opportunity!
Have a great week of ventures and adventures ahead!
Jane