2025 AI Investment

Perspective · Insights · Action

On-the-Ground Observations from Silicon Valley

PERSPECTIVE 视角

$200B capital flow & outlier cases signaling 2026

INSIGHTS 见闻

SV market entry case studies & strategic frameworks

ACTION 实战

Finding talent, funding & direction in Silicon Valley

CEIBS MBA  |  SVTR|硅谷科技评论  |  2026

Read full article →

PART 01

Perspective 视角

On-the-Ground Observations of 2025 AI Investment

Explore the flow of $200 billion in capital and the trends shaping its movement. Discuss noteworthy "outlier" cases that may signal shifts for 2026.

Where Did the $200B Go?

AI captured 50% of all global VC funding — not a bubble, a consolidation

$200B
Total AI VC Funding
50% of all global VC
302
AI Unicorns
15 new in 2025
$129.3B
Top 3 Mega-Deals
OpenAI + xAI + Anthropic

De-Nvidia Compute

Etched.ai $5B, custom ASICs

Energy Infra

Power > silicon at scale

Vertical AI Agents

Cursor $29B, Xaira $4B+

Capital follows infrastructure, not hype — the biggest checks go to compute, chips & energy.

Outlier Cases: Signals for 2026

Noteworthy anomalies that may reshape the landscape

DeepSeek

Cost Revolution

Open-source model rivaling GPT-4 at a fraction of the cost.
Signal: 90%+ drop in inference costs coming in 2026.

Cursor

$1B ARR → $29.3B

From dev tool to autonomous coding agent in 12 months.
Signal: Vertical AI agents replacing horizontal SaaS.

Physical Intelligence

$5.6B Pre-Product

Foundation models for robotics — zero mass-market product.
Signal: Embodied AI is the next frontier bet.

Circular Financing

Risk Pattern

Cloud providers fund startups who spend it back on cloud.
Signal: Inflated metrics — look for real unit economics.

PART 02

Insights 见闻

Practical Takeaways from the Silicon Valley Market

Analyze inspiring company cases, reviewing their early-stage strategies and typical challenges. Extract actionable tactical pathways and strategic frameworks for CEIBS entrepreneurs.

Market Entry: Case Studies

Crusoe Energy

Pivot Strategy

Bitcoin mining → AI data centers

  • Identified power > silicon at scale
  • Repurposed stranded energy for GPU clusters
  • Lesson: Infrastructure arbitrage opportunities

Etched.ai

De-Nvidia Play

$500M raised, $5B valuation

  • Custom ASICs for Transformer arch only
  • Bet specialization over general-purpose
  • Lesson: Find the monopoly gap, go narrow

Xaira Therapeutics

Vertical Deep Dive

$4B+ in generative biology

  • Foundation models for drug discovery
  • Top pharma + AI talent crossover
  • Lesson: Domain expertise × AI = moat

Strategic Frameworks for CEIBS Entrepreneurs

Two Playbooks: US vs. China

US: Universal Foundation Models

  • Software-first, model-centric approach
  • AGI-driven long-term R&D bets
  • Capital-intensive frontier labs

China: Hardware-First Mass Production

  • DeepSeek: cost-efficient open models
  • Unitree: 5,500 humanoids @ $16K each
  • Speed-to-market & price advantage

CEIBS advantage: Bridge both ecosystems

Tactical Pathways

Infrastructure Arbitrage

Find overlooked compute/energy gaps before they're priced in

Vertical Specialization

Pick one industry + go deep. Domain expertise = defensible moat

Cross-Border Leverage

US tech + China manufacturing + global distribution

Talent Network Effects

Google/Meta alumni, Stanford/Tsinghua grads = your hiring pool

Capital Timing

Follow infra cycles, not hype cycles. Build during corrections

PART 03

Action 实战

Finding Talent, Funding, and Direction in Silicon Valley

Drawing on SVTR's direct experience: core strategies for building top-tier teams, navigating venture capital networks, cultural adaptation, and avoiding common pitfalls.

找人 · 找钱 · 找方向

Finding Talent, Funding, and Direction — From SVTR's Experience

Finding Talent 找人

  • Google/Meta alumni dominate AI founding teams
  • Stanford/Tsinghua = top AI talent pipelines
  • Cultural fit matters as much as technical skill
  • Build cross-cultural teams for global reach
  • Referral networks > job postings in SV

Finding Funding 找钱

  • YC (283 deals) & a16z (237) most active in AI
  • Warm intros via network = 10× cold email
  • Understand circular financing — avoid inflated metrics
  • China capital + US innovation = powerful combo
  • Timing: infra bets now, app bets after correction

Finding Direction 找方向

  • Pick vertical AI > horizontal platform plays
  • De-Nvidia compute = near-term infra gap
  • Energy infrastructure = multi-decade thesis
  • SLM + edge AI = underexplored opportunity
  • Defense tech going mainstream — $50B+ TAM

Avoid Common Pitfalls 避坑指南

Inverse Pyramid

App-layer valuations lack profitability to support infrastructure costs beneath.

Validate unit economics before scaling. Focus on gross margin, not GMV.

Valuation Bubble

Embodied AI startups command $5B+ valuations with zero proven delivery.

Ask: Can this company ship a product in 18 months? If not, be cautious.

Circular Financing

Cloud providers fund startups who spend it back — inflated revenue loop.

Track net-new revenue vs. recycled cloud credits. Follow the cash.

Cultural Misread

Applying China playbook directly in SV (or vice versa) without adaptation.

Hire local operators. Build trust through ecosystem participation, not just capital.

Key Takeaways for CEIBS Entrepreneurs

$200B is consolidating, not scattering

Capital flows to infrastructure (compute, energy, chips). Position yourself in the value chain.

Outliers signal the next wave

DeepSeek, Cursor, embodied AI — watch cost disruption and vertical specialization.

Bridge US & China ecosystems

CEIBS alumni have unique cross-border advantage. Leverage both playbooks.

Build networks before you need them

SV runs on warm intros. YC, a16z, Stanford networks are entry points.

Execute with systematic methods

Pick vertical → validate economics → build team → tap network → scale globally.

1 / 12