Yes — agent execution will become a core cloud primitive No — it will remain a niche feature within broader platforms Maybe — it depends on whether agents need persistent environments Too early — the category is still defining itself
Yes — AI-native firms will outcompete traditional services models No — relationships and reputation will keep incumbents dominant Maybe — it depends on whether Solvera can scale quality with AI Too early — the company just launched in July 2026
Yes — home batteries will become the backbone of residential grid flexibility No — utility-scale storage will dominate flexibility markets Maybe — it depends on smart meter rollout and device adoption Too early — the VPP business model is still proving itself
Yes — neutral governance will win because no provider can police rivals No — platform-bundled controls will be good enough for most enterprises Maybe — it depends on whether security teams demand cross-provider visibility Too early — the market is still defining the category
Yes — founder-led speed and customer focus can beat scale No — scale and capital will always win in telecom Maybe — it depends on whether AloFor can build repeat demand Too early — the company just launched in July 2026
Yes — creators will become the primary distribution channel for consumer software No — traditional product-led growth will remain dominant Maybe — it depends on whether creator audiences convert to paying users Too early — the model has traction but no scaled success yet
Yes — Heartbeat AI will become the universal energy OS for homes No — hardware will remain the core business Maybe — software will grow but hardware will dominate revenue for years Too early — the software pivot is still proving itself
Yes — two-sided agent networks will become the standard for B2B payments No — one-sided AP/AR platforms will adapt and maintain dominance Maybe — it depends on whether large enterprises adopt both sides Too early — the network is proving traction but scale is untested
Yes — AI decision tools will become essential for network management No — budget constraints and inertia will slow adoption Maybe — it depends on whether the M25 results can be replicated Too early — the market is still proving the value proposition
Yes — airships will become the standard for routine infrastructure inspection No — helicopters will remain dominant for complex or urgent missions Maybe — airships will capture routine inspections, helicopters will handle the rest Too early — manufacturing scale-up is untested
Yes — AI-native certification will become the standard for medtech No — traditional Notified Bodies will adapt and maintain dominance Maybe — it depends on whether regulators accept AI-assisted reviews broadly Too early — Scarlet’s model is proven but scaling is untested
Yes — AI testing will displace most manual QA outsourcing No — human testers will remain essential for complex scenarios Maybe — it depends on whether Klarent can scale verification economics Too early — enterprise adoption cycles are still early
Yes — enterprises will prefer open-source for control and flexibility No — closed platforms will maintain dominance through compliance and support Maybe — it depends on the enterprise's technical capabilities Too early — the enterprise AI agent market is still forming
Yes — specialised hardware will accelerate neurotech from research to clinic No — device companies will continue building custom chips in-house Maybe — it depends on whether MintNeuro’s chips prove reliable at scale Too early — the neurotech market is still in early stages
Yes — telemetry quality will be as routine as linting or testing No — instrumentation will remain an afterthought for most teams Maybe — only large enterprises will adopt it systematically Too early — AI agents may solve this automatically
Yes — painless monitoring will dramatically expand CGM adoption No — traditional CGMs will maintain market dominance Maybe — it will win on cost and comfort, not accuracy Too early — pivotal trial and CE approval are still ahead
Yes — field-level data and AI can reliably predict crop risks early No — agriculture is too complex for reliable prediction Maybe — it works for some crops and regions, not all Too early — the platform launched in September 2026, results pending
Yes — alternative feedstocks are the key to scaling SAF No — other pathways will scale faster than feedstock innovations Maybe — it depends on whether CMU works at commercial scale Too early — pilot results will determine the real potential
Yes — minimally invasive approaches will become standard for ostomy creation No — surgical ostomies will remain the standard for most patients Maybe — PercStoma will serve inoperable patients, not replace surgery broadly Too early — clinical validation must come first
Yes — AI will catch most avoidable errors before trials begin No — human judgment will always be needed for complex tradeoffs Maybe — AI will reduce but not eliminate avoidable amendments Too early — the industry is still evaluating AI's impact on trial design
Yes — integration and data connectivity will win over standalone tools No — best-of-breed standalone tools will still attract advisers Maybe — it depends on the firm size and existing tech stack Too early — the market is still deciding on AI strategy
Yes — cost parity and emissions pressure will drive rapid adoption No — conventional cement will dominate for decades Maybe — recycled cement will capture a significant niche, not the mainstream Too early — supply chain and standards barriers are underestimated
Yes — infrastructure like Procuros will enable full automation within 5 years No — human oversight will remain essential for exceptions and compliance Maybe — automation will increase, but full autonomy is a decade away Too early — the data layer problem is harder than the agent problem
Yes — AI will handle most scheduling calls within 5 years No — humans will remain essential for complex patient needs Maybe — AI will augment, not replace, reception teams Too early — edge cases will slow adoption more than technology allows
Yes — agents will handle most routine servicing tasks within 5 years No — humans will remain essential for compliance and edge cases Maybe — agents will augment, not replace, servicing teams Too early — regulation will slow adoption more than technology allows
Yes — messaging agents will replace most apps within 5 years No — dedicated apps will coexist with agents indefinitely Maybe — only for simple tasks, not complex workflows Too early — the distribution layer is the real bet, not app replacement
Clinical validation — AI-designed therapies haven't proven efficacy in human trials Regulatory pathway — unclear how regulators will evaluate AI-designed drugs Manufacturing — scaling AI-designed therapies for commercial production Competition — larger, better-funded rivals may reach clinic first
AI and agentic systems HealthTech and life sciences Consumer and lifestyle platforms CleanTech and industrial
Pix (Brazil) UPI (India) SEPA (Europe) Mobile money (Africa)
Hardware compatibility — too many manufacturers, too few integrations 50% Consumer awareness — homeowners don't know heat pumps can earn money 20% Regulatory fragmentation — energy markets differ by country 10% Competition — better-funded VPP platforms scaling faster 20%
Live pitches 33% Emerging startups 29% Investor access 21% Community networking 17%