What's drawing LPs back to China
- 01Theme 1: China's AI & Robotics Edge Is Reversing Western LP Sentiment
- 02Theme 2: PE Is Moving Up the AI Stack
- 03Theme 3: Enterprise SaaS Is Surging Against the Prevailing Narrative
- 04Theme 4: AI Is Creating a Bifurcated Workforce
- 05Theme 5: European M&A Is Accelerating, Driven by AI Valuation Bifurcation
1. Key Themes
Theme 1: China's AI & Robotics Edge Is Reversing Western LP Sentiment
After four consecutive years of fundraising decline, China's private markets appear to be bottoming out — driven by structural advantages in energy costs, manufacturing scale, and a reopening IPO market in Hong Kong.
"I do observe very clearly that sentiment is improving from a Western LP perspective. … If you look across the world, Chinese private equity still churned about 14% net returns." — Brooke Zhou, LGT Capital Partners
"We've been seeing a lot of attention in large language models in the first half of 2026. But I believe in the second half of this year, we will see more activity in application robotics." — Lane Zhao, InnoVision Capital
Theme 2: PE Is Moving Up the AI Stack — From Data Centers to Compute
The Blackstone-Google JV represents a structural shift in how private capital engages with AI infrastructure. Rather than just owning real estate (data centers), institutional investors are now underwriting the computing power that runs inside them.
"The tie-up opens a new avenue amid the rush of private equity into AI infrastructure, moving it beyond data centers and into financing computing power itself."
"The JV also allows Blackstone to own the compute that sits inside data centers, not just the facilities themselves, of which it has acquired several over the past five years."
Theme 3: Enterprise SaaS Is Surging Against the Prevailing Narrative
Despite widespread conventional wisdom that AI would erode or cannibalize SaaS, VC funding into enterprise SaaS hit a record high in Q1 2026 — a 6x quarter-over-quarter jump.
"VC funding for the enterprise SaaS industry surged to a record $173 billion in Q1—a more than 6x increase quarter-over-quarter."
Theme 4: AI Is Creating a Bifurcated Workforce — Junior Roles Are Being Cut
CEO hiring plans have dramatically reversed in one year. The AI disruption is restructuring the workforce composition toward seniority, not just automating tasks. This has significant implications for talent strategy at startups and portfolio companies.
"A year ago, most CEOs planned to grow junior roles; now more than 40% plan to cut them. On the back of AI disruption, the composition of the workforce is shifting toward mid- and senior-level positions."
Theme 5: European M&A Is Accelerating, Driven by AI Valuation Bifurcation
Europe's M&A market posted a 42% year-over-year jump in deal value in Q1 2026. Simultaneously, a growing valuation gulf has emerged in European VC — with AI companies pulling far ahead of the rest.
"Q1 2026 sustained the strong M&A momentum that characterized the second half of 2025 in Europe, posting a 5.3% quarter-over-quarter increase in deal value and up an impressive 42% year-over-year."
"There's a growing valuation gulf in Europe's VC market. The culprit (you guessed it) is AI."
2. Contrarian Perspectives
China Is Already Exporting AI — in the Form of Cheap Tokens
The conventional framing is that China is locked out of leading AI chips (Nvidia export bans). The contrarian take: China's energy cost advantage means it can produce AI inference at dramatically lower cost per token — effectively exporting AI capability without needing the most advanced chips.
"China is already exporting low-cost electricity in the form of tokens. Ultimately, the LLM model relies on chips, energy, and which country can provide the lowest-cost token. China has double the electricity productivity than the US, but at a 50% lower cost per kilowatt." — Jing Hong, Gaocheng Capital
This reframes the US-China AI competition: the bottleneck may be energy economics, not just chip access.
AI Chips Are Too Hard to Underwrite — Yet Blackstone Is Doing It Anyway
The standard infrastructure investor playbook relies on long-duration, stable-cash-flow assets. AI chips are the opposite: they depreciate fast, bundle hardware with software and contracts, and have no established residual value framework. The contrarian move is leaning in anyway.
"AI chips are very difficult to underwrite. Google has rolled out four TPU generations in roughly three years, compressing the useful life of any given chip and making it hard to ascertain their residual value. A data center lease lasts 15 years, while a generation of TPU technology can feel antiquated in three."
The bet here is that demand certainty (from hyperscalers) compensates for asset depreciation risk — a novel and untested infrastructure thesis.
Enterprise SaaS Is Thriving Because of AI, Not Despite It
The dominant narrative has been that AI agents and LLMs would compress software spending and hurt SaaS multiples. The data says the opposite is happening — enterprise SaaS VC funding just hit an all-time high.
"VC funding for the enterprise SaaS industry surged to a record $173 billion in Q1—a more than 6x increase quarter-over-quarter, according to our latest report on the sector."
The implication: AI is expanding the total addressable market for enterprise software, not cannibalizing it — at least in the near term.
3. Companies Identified
| Company | Description | Why Mentioned | Key Quote |
|---|---|---|---|
| Blackstone | Global alternative asset manager | Formed JV with Google to offer compute-as-a-service; expanding from data center ownership to compute ownership | "The JV also allows Blackstone to own the compute that sits inside data centers, not just the facilities themselves." |
| Big Tech / cloud & AI | Supplying TPUs, hardware, and software for the Blackstone JV; using the deal to expand TPU market share beyond Google Cloud | "The deal is partly a way for Google to expand the market for its TPUs, AI chips that until now have lived almost entirely inside the Google Cloud software suite." | |
| Decart | AI startup | Raised $300M at $4B valuation, led by Radical Ventures | Notable deal size and valuation in AI |
| Armada | Modular data center builder | Raised $230M Series B at $2B valuation, led by Overmatch, BlackRock, 8090 Industries | Signals continued institutional appetite for AI infrastructure |
| Commure | AI for healthcare administration | Raised $70M led by General Catalyst at a $7B valuation | Highest valuation-to-round ratio in the deal list — notable for healthcare AI |
| Kikoff | Fintech / credit-building startup | Co-founder alleging CEO froze IPO plans and pressured discounted share sale | Case study in cap table governance failure |
| Linkerbot | China-based physical AI / robotics startup | Hired banks for a Hong Kong IPO | Signals China robotics IPO pipeline is active; validates H2 2026 robotics thesis |
| Sima.ai | Specialized chips for drones and cameras | In talks to raise $100M+ at $1.4B valuation | Represents edge AI / robotics chip investment wave |
| Radar | Retail intelligence platform | Raised $170M Series B at $1B valuation | AI application layer hitting unicorn scale in physical retail |
| Dust | Enterprise AI agent builder (Paris-based) | Raised $40M Series B backed by Sequoia, Snowflake Ventures | European AI application layer gaining tier-1 investor backing |
| Bunch | AI-native fund operations platform for private markets | Raised $35M Series B led by Portage | FinOps infrastructure for PE/VC going AI-native |
| Nourish | Telehealth for nutrition-related chronic conditions | Raised $100M Series C led by Menlo Ventures | Digital health at the intersection of GLP-1 trend and AI |
| Tenstorrent | AI chipmaker | Intel and Qualcomm reportedly interested in acquisition | Strategic M&A signal — legacy chip giants racing to acquire AI chip IP |
| Mistral | French LLM developer | Acquired Emmi AI (Austrian industrial simulation tech) | European AI consolidation underway; LLM players buying vertical application capabilities |
4. People Identified
| Person | Description | Why Mentioned | Key Quote |
|---|---|---|---|
| Brooke Zhou | Co-head of APAC business, LGT Capital Partners (Swiss asset manager) | Cited on improving Western LP sentiment toward China private equity and 14% net returns | "I do observe very clearly that sentiment is improving from a Western LP perspective." |
| Jing Hong | Founding partner, Gaocheng Capital (growth-stage investor) | Made the energy-economics argument for China's AI cost advantage | "China has double the electricity productivity than the US, but at a 50% lower cost per kilowatt." |
| Lane Zhao | CEO, InnoVision Capital (Beijing-based PE firm) | Called the H2 2026 shift from LLMs to application robotics | "I believe in the second half of this year, we will see more activity in application robotics." |
5. Operating Insights
Governance Risk Is a First-Order Startup Risk, Not a Legal Footnote
The Kikoff dispute — where a co-founder alleges a CEO froze IPO plans and pressured a discounted share sale — is a live reminder that cap table structure without governance guardrails creates leverage for abuse.
"Kikoff's co-founder is alleging in Delaware Chancery Court that the fintech startup's CEO froze IPO plans and pressured him into a discounted share sale—exposing the power struggles that can fester when cap table control goes unchecked."
Takeaway for founders and investors: Shareholder agreements, board composition, and drag-along/tag-along rights need to be structured before tension emerges — not after.
Workforce Strategy Must Now Plan for AI-Driven Junior Role Elimination
The rapid reversal in CEO hiring intentions — from growing junior headcount to cutting it — means companies should redesign onboarding, knowledge transfer, and career ladders now, before they face structural talent gaps at the mid-level.
"A year ago, most CEOs planned to grow junior roles; now more than 40% plan to cut them. On the back of AI disruption, the composition of the workforce is shifting toward mid- and senior-level positions."
Takeaway for operators: Invest in senior talent pipelines and mentorship infrastructure; junior-role pipelines are becoming unreliable feeders for the future org.
Anthropic Has Published a Public AI Startup Playbook
A directly actionable resource for builders was flagged without fanfare.
"If you're building or backing early-stage companies right now, Claude thinks you should read this. Anthropic just put out a playbook on how to build an AI startup."
Takeaway: For early-stage founders building on or around AI, this is a primary-source strategic document from one of the leading AI labs — worth reading for both product and go-to-market signals.
6. Overlooked Insights
Hong Kong Has Claimed the #1 IPO Market Position Globally in 2026
This is mentioned only in passing but is highly significant. It means the exit pathway for China-based private capital is functionally open again — which directly validates LP re-engagement with China PE/VC.
"Sentiment is further supported by China's robust exit market, particularly in Hong Kong, which claimed the world's top position for IPOs this year."
For investors evaluating China exposure, this is the structural exit signal that changes the liquidity calculus — not just the investment thesis.
"Cozy Games" Is Now a Billion-Dollar Market Driven by Burnout
Buried in the Side Letters section with no further commentary, this is a legitimate emerging consumer market signal.
"The 'cozy games' genre has grown into a billion-dollar market driven by burnout and escapism."
The underlying driver — widespread stress and screen-fatigue — is durable and growing. For investors in consumer entertainment, wellness, or digital engagement, this genre represents an underexplored consumer category with structural tailwinds.