The AI Way

1. Macro Regime, Valuations & the AI Inflection

We enter the mid-2020s in a somewhat paradoxical environment: public markets are near all-time highs in many regions, monetary conditions remain relatively tight, and growth is moderating. At the same time, we are seeing a renewed capital infusion into frontier technologies - particularly AI infrastructure, models, and application layers, that may catalyse a new productivity super cycle. AI is increasingly viewed as a general-purpose technology (GPT) with structural implications for capital allocation, competitive moats, and return dispersion. 

But the AI wave isn’t uniform or risk-free. Valuation excesses have emerged, especially in high-profile AI names whose multiples seem to reflect future potential more than proven cash flows. The concept of Capability Realisation Rate (CRR), i.e. the gap between AI potential and what’s delivered, becomes critical in detecting mispricing. In short: AI matters now, because the structural shift is underway, but timing, discipline, and differentiation will decide winners and losers.

Takeaway: In this regime, passive, broad-based exposure is unlikely to deliver strong excess returns. Instead, you need active navigation—tilting toward innovation, rotating across sub-themes, and calibrating risk with macroeconomic signals.

2. Portfolio Lens - How AI Should Be Positioned Within a Broader Construct

The presence of AI as a thematic core should not displace the need for balance, diversification, and liquidity. Rather, AI should be treated as a convex engine - a tilted “bet within a bet” that lives inside a well-constructed base portfolio.

“Embedded Innovation — Integrating AI and DeepTech as a Structural Growth Driver”

• Public + Private Exposure: Listed thematic AI equities with private/VC innovation help capture both scalable platforms and early-stage asymmetry.
  
  
• Dynamic AI Allocation: Active rotation across the AI value chain - compute and chips, data infrastructure, model platforms, and applied verticals (e.g., health, energy, autonomy).
  
  
• Globalisation of Innovation: Extend AI exposure beyond traditional hubs; allocate to Asia/APAC and frontier ecosystems where leapfrog adoption may occur.
  
  
• Compounding Themes: Build for longevity by layering future frontiers - biotech, climate tech, and synthetic biology, always through an AI interoperability lens.
  
  Takeaway: The AI theme becomes not a bet on a single stock or sector, but an embedded, risk-adjusted engine that can deliver asymmetric upside while being tempered by broader diversification. 

3. Why 2025/2026 Vintage Could Be a Defining Moment for VC & AI

Venture capital has always been cyclical. The best vintages often come not at the peak of exuberance but in the reset years, when valuations correct, capital is more discerning, and top founders emerge in environments where others hesitate. I argue that 2025/2026 may be one of those inflection points, especially in AI and deep tech.

Here’s why:

1. Valuation reset and capital discipline
   After a frothy period of aggressive capital deployment into AI and tech (2020–2023), the correction in public markets and rising cost of capital have forced more rigorous underwriting. Higher bar means better entry points.
   
   
2. Surge of dry powder and investor rotation
   Many LPs and funds have held back capital during the more volatile period, now seeking compelling deployment options. AI remains the magnet. In Q1 2025 alone, a $40B AI deal lifted overall VC investment figures significantly. 
   
   
3. Emerging infrastructure maturity
   The AI stack is entering deeper phases: not just model training, but edge deployment, vertical AI layers, embedded systems, hybrid-cloud agents, and model marketplaces. These areas may yield more concentrated alpha than broad foundation models.
   
   
4. Improving exit environment
   As rate pressure eases and capital markets regain confidence, IPOs, M&A, and continuation vehicles will likely recover, unlocking value for earlier-stage funds. Continued growth in AI adoption and strategic acquirers motivates exit demand. 
   
   
5. Selective bifurcation and re-concentration
   The VC landscape is bifurcating: AI-native funds and AI-focused deals are commanding most attention and capital, while many non-AI sectors struggle under valuation compression. The winners from the 2025/2026 vintage will likely be the ones that navigate this concentration and pick niche leaders.
   
   Takeaway: In effect, a well-timed, disciplined investment in a top-tier 2025/2026 AI/tech fund can yield returns that compound strongly over the following decade. The “optionality” potential from that anchor vintage can drive large portions of total portfolio outperformance.

4. Balancing Opportunity and Risk - Key Guardrails

While the AI wave offers extraordinary upside, it is neither guaranteed nor linear. Any investment framework must account for the real risks. Here’s how to temper ambition with prudence

• Valuation discipline is paramount. Avoid frothy late-stage AI plays unless there is already robust traction or proof points. Use valuation models like CRR to identify misalignment.
• Diversify across the AI stack. Don’t put all capital into chips or foundational models. Spread across edge AI, vertical application layers, model tooling, data infrastructure, model ops, AI in domains (e.g. health, climate).
• Liquidity and pacing. Given the illiquidity of VC, carefully and consistently phase deployments, maintain optional dry powder, and avoid overcommitment in any single vintage.
• Manager selection and alignment. Pick GPs with domain expertise (in AI), technical networks, operational rigor, and aligned incentives.
• Hedge & tail risk strategies. Maintain convective hedges (e.g. long volatility, options, tail-risk instruments) and rebalancing triggers.
• Scenario playbooks. Define clear stop-loss, pivot, or hedge triggers in case the AI narrative slows, regulation intervenes, or macro shocks occur.

Takeaway: By embedding these guardrails, the portfolio can ride the upside of AI without being crippled by a collapse of exuberance.

5. This is the moment to invest in the “AI way”

In an era where many public and private asset classes sit near cyclic highs, the AI wave represents one of the few structurally asymmetric windows available to investors. The confluence of infrastructure maturation, capital discipline, capital rotation, and thematic momentum forms a compelling case for structuring a portion of capital toward AI and related deep tech.

Yet this is not a “spray and pray” regime. Discipline, diversification, pacing, and manager selection are crucial. The ideal portfolio will treat AI not as a standalone bet, but as a tilted engine nestled within a diversified base.

Critically, I believe the 2025/2026 vintage window has the potential to be one of the most fertile periods for VC / AI investing over the coming decade. That vintage may capture the reset valuations, the momentum of capital redeployment, and the emergence of foundational companies in next-layer AI applications.

Takeaway: in a world of compressed public-market upside and rising dispersion, venture capital - especially in AI-forward vintages - may offer one of the rare asymmetric levers left in your toolkit. Put differently: this is the moment to invest in the “AI way”, but only if done with rigor, balance, and selective conviction.