Companies, capital, and market rhythm
Turing Award winner Yann LeCun founded AMI Labs after leaving Meta, raising $1.03B in seed funding to build world models that challenge the LLM paradigm. Europe's largest seed round ever, but also its biggest gamble.
DeepSeek completed a 50 billion yuan funding round, valuation jumping from $20B to $45B. Founder Liang Wenfeng contributed 40% personally, with the National IC Fund leading. V4 model plans a V4.1 upgrade in June, focused on coding capability improvements.
China's major AI labs have raised ~$9B cumulatively, with DeepSeek's new ~$7.35B round in progress. Open-source-first, aggressive pricing, and talent wars are reshaping China's AI landscape.
Akamai announces a $1.8B compute deal with Anthropic, shares surge 28%. Model companies are converting funding directly into locked-in global edge compute capacity.
DeepSeek launches $7 billion funding round, founder Liang Wenfeng personally contributes approximately $3 billion for 40%, while maintaining 90% ownership. In Chinese AI companies, this level of founder absolute control in funding structure is nearly unprecedented.
Alibaba and DeepSeek failed to reach agreement on funding terms, with ecosystem compatibility as the core disagreement. DeepSeek fundraising target up to 50 billion yuan, with no shortage of potential investors.
ByteDance raises AI infrastructure spending to 200 billion yuan while cutting 30% of internal AI projects. The money is not increasing - it is concentrating.
According to Bloomberg, SoftBank has scaled back its $10 billion margin loan plan backed by its OpenAI equity stake, as some creditors show hesitation. Concerns over an AI company valuation bubble are beginning to impact the financing side.
DeepSeek's first-round funding structure underwent a dramatic shift within 48 hours: founder Liang Wenfeng plans to personally invest up to 20 billion RMB (approximately a 40% stake), replacing the Big Fund as the largest single investor. The post-money valuation exceeds $51.5 billion, with the V4.1 model slated for release in June.
Baidu AI chip subsidiary Kunlunxin began STAR Market listing guidance on May 7, having already filed confidentially for a Hong Kong IPO in January. Jefferies values it at $16-23B. Baidu holds 58%. The capital window for China second-tier AI chips is opening.
According to Reuters, DeepSeek is conducting its first external funding round with a potential valuation of $50 billion, led by China's National Artificial Intelligence Fund. The model company that long rejected outside capital is changing strategy.
US National Science Foundation launched OMAI compute systems with NVIDIA Blackwell Ultra, turning $152M national investment into public infrastructure for open AI research. Academia compute anxiety temporarily eased.
Menlo Ventures partner updates AI Neolabs list to 63 companies, all valued over $1B with no scaled revenue, totaling ~$300B in valuation—up 13 from February.
Global AI funding in April 2026 totaled $37 billion, representing 66% of global VC investment that month. AI model companies raised $26.7 billion. Capital continues to concentrate heavily at the model layer.
Morgan Stanley's latest report raises the 2026 AI infrastructure spending forecast for the five major tech giants (Amazon, Google, Meta, Microsoft, Oracle) to $805 billion, with 2027 projected to reach $1.1 trillion. The AI infrastructure arms race enters the trillion-dollar era.
German AI company PriorLabs, just 18 months old, receives over €1 billion in strategic investment from SAP. Its core product TabPFN is a foundation model designed specifically for tabular data, outperforming traditional machine learning on small datasets without any parameter tuning.
According to The Information, Anthropic commits $200 billion to Google Cloud over five years for compute and chips, representing over 40% of Google Cloud disclosed RPO. Combined with $100B from AWS and $63B from Broadcom, Anthropic total compute bet exceeds $360 billion.
Israeli startup CopilotKit raises $27M. Its AG-UI open standard is already adopted by Google, Microsoft, Amazon, and Oracle. The agent interaction layer is moving from fragmented to standardized.
Alphabet issued six tranches of euro bonds in the European bond market, raising at least €3 billion. This is not a routine financing move — it signals an escalation in the AI arms race, as tech giants begin using debt markets to fuel compute expansion.
Reuters confirms with three sources that DeepSeek could reach a $50 billion valuation in its maiden external fundraising round, up from the previously reported $45 billion Big Fund-led deal. The company that rejected outside capital for three years is now shifting to a capital-driven strategy, marking a new phase in China AI valuation.
The US government AI Guidance public comment window closes on May 8, 2026. This is the last opportunity for industry participants to shape the AI regulatory framework, covering model safety, data privacy, and automated decision-making. Policy direction will directly impact compliance costs and competitive landscape for global AI enterprises.
Goldman Sachs publishes AI infrastructure industry report, arguing that Agent AI through the low token cost times high token consumption flywheel effect is turning AI infrastructure from money-burning to money-making. The profit inflection point is in 2026. This is Wall Streets first systematic argument of AI infrastructure profitability timeline.
Anthropic valuation surged from $120B in October 2025 to $1.2T in May 2026 — a 10x increase in 7 months. The driver was not traditional VC funding but on-chain pre-IPO trading tools — tokenizing private company shares via SPV structures for 24/7 secondary market trading and instant price discovery.
Anthropic valuation exceeds $900 billion with a new $50 billion funding round in progress. Meanwhile, Claude announces doubled usage for Pro, Max, and Team subscriptions, plus lowered API rate limits. GPT-5.5 pressure is forcing Anthropic to spend real money to retain users.
Q1 2026 global AI sector VC funding reached $226 billion, up 4.5x quarter-over-quarter, exceeding the full year 2025 total. 80% flows to AI, four companies take 65%. LLM players are the biggest winners, industry concentration reaches historic levels.
Infini-AI announces another round of funding exceeding 700M RMB, bringing cumulative total to nearly 2.2B. MaaS platform daily token calls surged over 20x compared to end of 2025, providing high-performance inference optimization for leading models like Kimi, GLM, MiniMax, and DeepSeek.
PixelBloom (AiPPT.com) announces completion of Series C funding, co-led by Guoke Investment and SenseTime Guoxiang Capital. The company shifts its strategy from an AI visual expression platform to an 'AI Office Solutions Agent,' launching Xiaofang (Fangan.com), a marketing agent with over 30 million global users.
PriorLabs raised just €9M in seed funding 15 months ago, now acquired by SAP with a €1B investment commitment. Its core product TabPFN is the world first tabular data foundation model, outperforming traditional ML on small datasets.
On May 6, 2026, Alphabet (Google) officially surpassed NVIDIA as the world's most valuable company. The AI investment narrative shifts from chipmakers selling picks to platforms controlling data and distribution, marking a new phase in the AI industry landscape.
Moonshot AI is closing a new $2 billion funding round, pushing its post-money valuation past $20 billion. Meituan DragonBall leads with a single investment exceeding $200 million, joined by China Mobile and CPE. Combined with three earlier rounds this year, Kimi has raised over $3.9 billion in less than six months.
A16Z launches a $2.2 billion new crypto fund, explicitly targeting the intersection of AI with crypto and traditional finance. Q1 2026 AI industry absorbed over $240 billion in investment, with AI VC funding up 4.5x quarter-over-quarter — the heat keeps rising.
According to Bloomberg and The Information, China National Integrated Circuit Industry Investment Fund is in talks to lead DeepSeek investment at nearly $45 billion valuation. This marks the fund first public investment in a domestic LLM company, signaling AI models are now on par with semiconductor national strategy.
SubQ closes $29M funding round to retrain a model for 12M token context; meanwhile evermind's MSA chooses to add an external memory layer to mainstream models. Two routes represent a fundamental split in the sparse attention field: vertical integration or horizontal embedding?
Anthropic released ready-to-use Claude Agent templates for financial services, covering pitch creation, valuation review, month-end close, and more. These templates can be used as plugins for Cowork and Claude Code, or deployed directly to production as Managed Agents.
UBS analysis points out that agentic AI workloads require 3-5x more CPU cores than traditional GPU training, with server CPU TAM expected to grow from $30B in 2025 to ~$170B by 2030, approximately 5x growth. Data center architecture is shifting from GPU-centric to complex orchestration.
The latest AI company valuation rankings for May 2026 show Anthropic surpassing OpenAI at $900B to take the top spot, xAI valued at $250B, DeepSeek at $45B entering the global top six, while capital concentration in the AI sector continues to shatter historical records.
The Forbes AI 50 2026 list shows 50 companies with combined funding of $305.6 billion, with OpenAI and Anthropic alone accounting for 80%. But the real signal worth watching is: vertical applications like Cursor ($29B valuation) and Gamma ($100M ARR) are proving commercial viability.
Greg Brockman confirmed in court testimony that OpenAI compute spending will reach $50 billion in 2026, a 1,600x increase from $30 million in 2017. This massive expenditure flows directly to NVIDIA, Microsoft, Amazon, and other infrastructure providers, while signaling that GPT-6 will operate at a scale far beyond existing models.
Q1 2026 semiconductor AI profit data is clear: NVDA leads at $42.3B quarterly, Broadcom at $7.3B, Micron at $13.8B, AMD at just $1.2B. The profit gap among the four major players has widened to 35x, and the winner landscape of the AI compute race is taking shape.
Neocloud reveals H100 GPU rental prices are higher than 3 years ago, with frontier labs locking up most 2026 GPU supply. Combined with hyperscaler $725B capex and Micron HBM supply sold out, 2026 AI compute presents a "money cannot buy" supply landscape.
Anthropic announces enterprise AI services joint venture, marking strategic shift from pure model provider to "end-to-end AI delivery." This positions them to compete directly with enterprises and consulting firms for AI implementation projects, not just at the API level.
OpenAI and 19 private equity firms including TPG, Brookfield, and SoftBank formed "The Deployment Company," raising over $4B and promising investors 17.5% annual returns. The goal: push AI software into 1,200 enterprises. Anthropic countered the same day with $1.5B.
The Forbes AI 50 2026 list shows $305.6B in combined funding among selected companies, with OpenAI and Anthropic taking 80%. But the real investment signal hides in the next tier: Cursor ($29B valuation), Harvey, Abridge, and Gamma ($100M ARR) are proving the commercial viability of vertical AI applications.
Anthropic annual recurring revenue surged from $9B at end of 2025 to $44B in May 2026, adding $35B ARR in 12 months, equivalent to $96M per day. Investors submitting allocations within 48 hours for a $50B round, valuation could exceed $900B, surpassing OpenAI.
Corgi launches AI Coverage, an insurance product specifically covering AI hallucination-caused decision errors, copyright infringement, data breaches, and autonomous AI Agent risks. This is the world's first dedicated insurance for AI application liability, marking AI risk management's transition from technical solutions to financial solutions.
Q2 2026 embodied AI/robotics funding exceeds $4 billion, with SkildAI setting a single-round record at $1.4B, followed by China's TARS at $513M and Apptronik at $520M. Capital is betting real money on AI's physical world outlet.
On May 4, 2026, Anthropic announced a partnership with Blackstone, Goldman Sachs, and Hellman & Friedman to launch an independent enterprise AI services company. Applied AI engineers will embed into mid-market firms to build custom Claude powered business solutions, marking Anthropic's transition from model provider to enterprise AI integrator.
Nvidia invested $420 million in the Bittensor/TAO network in Q1 2026, with 77% locked. Combined with Polychain Capital adding $200M exposure, decentralized AI infrastructure receives real-money endorsement from a chip giant for the first time.
Top-tier VC Coatue launched a new ventures entity to acquire land at scale for AI data centers, reportedly to lock in compute infrastructure for Anthropic. Compute investment is extending from chips to land and energy.
Moonshot AI completed three funding rounds in 90 days, with valuation surging from $4.3B to $18B. Kimi K2.6 is scheduled for June launch while K3 targets 2.5 trillion parameters. Capital markets are pricing in China AI next wave.
Y Combinator released its list of million-dollar AI startup directions for Summer 2026 batch. Analysis reveals vertical industry agents, AI safety infrastructure, and AI-native workflow tools as three high-signal tracks. AI funding exceeded 80% of global VC totals in Q1 2026.
NVIDIA announces resumption of RTX 3060 12GB production, with ASUS, MSI, Colorful resuming supply in June. In 2026, as MoE models drastically reduce VRAM requirements, this "budget GPU" becomes the cost-performance king of local AI inference.
Google Q1 2026 revenue $109.9B (+22%), cloud up 63% to $20B. But Pichai explicitly admits "compute constrained, cloud revenue would have been higher." Full-year capex raised to $190B. Backlog ~$460B, nearly doubled QoQ. AI demand far exceeds supply.
WTS again raises AI capital expenditure forecast for five hyperscaler tech companies: $805 billion in 2026, $1.1 trillion in 2027. 2026 spending alone roughly equals the combined capital expenditure of all non-tech companies in the S&P 500 in 2025.
Broadcom DSP chip prices have increased by 15–20%, pushing the average selling price (ASP) above $1,000 per unit; lead times have stretched to 50–70 weeks. Meanwhile, overall GPU lead times remain at 6–9 months. Multiple bottlenecks across the AI computing supply chain are intensifying, making it increasingly difficult for smaller companies to access high-end computing resources—and potentially further accelerating industry consolidation.
Gartner predicts that by 2026, 30% of enterprises will depend on AI agents capable of autonomously triggering transactions and completing tasks. Meanwhile, "AI Agent Orchestrator" emerges as a new profession — managing agent teams, fixing failures, and connecting agents to business outcomes. This marks AI transitioning from auxiliary tool to business execution entity.
Zhipu surged 690% after HK listing, MiniMax follows to HK exchange. Along with Manycore and a wave of AI companies going public, HK is becoming the primary capitalization venue for Chinese AI model companies. AI is rewriting the HK IPO landscape.
NVIDIA GTC 2026 unveils Vera CPU — the first processor designed for Agentic AI workloads, with 88 custom Olympus cores and 1.2 TB/s memory bandwidth. Meanwhile, Blackwell and Vera Rubin chip orders surpass $1 trillion through 2027. Agent workloads are reshaping AI server hardware design.
Stripe launched AI Agent wallet, Visa debuted Agent payment rail, Tether-backed Oobit provides AI Bot Visa corporate cards. In April 2026, three payment giants simultaneously built AI Agent payment infrastructure, marking the "Agent Economy" moving from concept to reality.
Microsoft $190B, Google $180-190B, Meta $115-135B, Amazon cloud +28%, Big Four 2026 AI CapEx combined exceeds $725B, doubling from last year. But market reaction is muted — growth stories need profit proof.
According to The Information, Tencent and Alibaba are in talks to invest in DeepSeek at a valuation exceeding $20 billion. In under two years, DeepSeek has matched Baidu market cap, validating the core logic that in the AI era, efficiency matters more than scale.
NVIDIA announced Groq 3 LPX + Vera Rubin racks shipping Q3 2026 with 35x inference throughput improvement. Meanwhile, hyperscaler annual Capex exceeds $600B, with most shifting from training to inference. AI chip market evolves from training monopoly to inference diversification.
Moonshot AI valuation jumped 4x in three months as Kimi K2.6 approaches Claude-level performance across benchmarks. The capital repricing reflects breakthrough technology, product maturation, and favorable domestic AI policy tailwinds.
The US Department of Defense reaches classified AI agreements with OpenAI, Google, and Nvidia — Anthropic, previously the primary model for classified information handling, is notably absent, reflecting a strategic shift from safety compliance to scale capability in military AI procurement.
Q1 2026 saw $300B in global AI venture funding. OpenAI ($122B), Anthropic ($30B), xAI ($20B), and Waymo ($16B) account for four of the five largest deals in history. AI has shifted from a technology narrative to a capital narrative.
Amazon's Q1 2026 earnings reveal $16.8B in pre-tax gains from Anthropic investment. Combined with Anthropic's $100B+ AWS spending commitment over 10 years, this marks a new 'investment-compute-binding' era in AI infrastructure.
Google, Microsoft, Amazon, Meta combined 2026 AI infrastructure capex hits $725B, up 77% from 2025. Q1 spent $130B already. AWS growing 28%, Google Cloud 63% but supply constrained.
OpenAI officially launches GPT-5.5, Codex Agent, and new Bedrock Managed Agents on AWS Bedrock April 29. This marks the end of OpenAI-Microsoft exclusivity, ushering in the multi-cloud Agentic era for enterprise AI.
San Francisco AI defense company Scout AI completes $100M Series A, signaling AI + national security as 2026 hottest capital direction.
AI chip maker Cerebras reportedly preparing up to $4B IPO at ~$40B valuation. As inventor of WSE wafer-scale engine, Cerebras seeks capital market validation between big tech custom chips and NVIDIA monopoly.
Lumenai Innovation Fund launches June 1, 2026, headquartered in Stamford, positioning as an institutional-grade hedge fund built agentic from day one — marking AI autonomous investment decisions entering live trading.
DeepSeek V4's delayed release is not a technical bottleneck but a deliberate strategic shift — deep integration with Huawei's Ascend ecosystem. CCTV-affiliated accounts reveal this marks Chinese AI leaders shifting from 'train on NVIDIA' to 'train on Ascend.'
During the 2026 Hong Kong Web3 Carnival, Binance co-founder He Yi revealed that YZi Labs, her family office with CZ, has invested in a Chinese AI model company. Moonshot AI (Kimi) is the most likely target based on multiple clues.
Meta reported Q1 2026 revenue of $56.3B while raising full-year CAPEX guidance to $125-145B. Zuckerberg stated directly that most of the increase comes from rising memory component costs. The bottleneck in the AI arms race is shifting from compute to memory.
Meta reports Q1 2026 revenue of $56.3B, while raising full-year CAPEX guidance to $125-145B. Zuckerberg directly attributes most of the increase to memory component pricing. The bottleneck in the AI arms race is shifting from compute to memory.
Ineffable Intelligence closes $1.1B seed round, breaking the seed round size record. AI investment is shifting from "validate then invest big" to "directly bet heavily" — mega seed rounds are becoming the new financing paradigm for frontier AI projects.
Apple admitted in Q2 2026 earnings call that Mac mini and Mac Studio supply shortages will last months. Tim Cook directly attributed this demand to agentic AI use cases like OpenClaw, exceeding Apple's own forecasts. This marks AI agents as a core driver of consumer hardware demand.
Cursor secured a $60B acquisition option from SpaceX, ARR hit $2.7B, but gross margin is -23% — paying $1.23 in API costs to Anthropic and OpenAI for every $1 of revenue. The economics behind the $50B valuation.
Former OpenAI researcher Leopold Aschenbrenner's Situational Awareness Fund grew from $225M to over $5.5B AUM in one year. Latest 13F filings show core holdings focused on AI infrastructure: energy, compute, optical interconnects, and memory. His investment thesis is being rapidly emulated by institutions.
AMD announced Advancing AI 2026 conference will be held on July 23 in San Francisco. With NVIDIA dominating AI training chip market and Huawei Ascend expanding rapidly in China, AMD next product roadmap will directly impact the global AI compute landscape.
During 2026 Hong Kong Web3 Carnival, Binance co-founder He Yi revealed YZi Labs invested in a Chinese AI LLM company, with clues pointing to Moonshot AI (Kimi). The first deep intersection of Web3 capital and Chinese LLM industry.
DeepSeek V4 natively compatible with Ascend 950 processors, combined with Cambricon Q1 revenue growth exceeding 150%, US export controls are transforming from "obstacles" to "forced monopoly accelerators" for Chinese AI chip makers. Domestic compute substitution has crossed the testing phase into full-scale deployment.
Q1 2026 earnings season shows collective acceleration in AI infrastructure spending: Meta raises CapEx to $125-145B, Google increases to $180-190B, AWS grows at fastest rate in 15 quarters, Azure up 40%. The AI arms race enters the ammunition phase.
DeepSeek valuation exceeds $40B, Zhipu GLM reaches $50B, MiniMax at $30B, and Moonshot AI Kimi prepares for HK IPO at ~$20B. The capital landscape of Chinese AI foundation models is reshaping, with valuations benchmarking against Silicon Valley.
Alphabet Q1 2026 revenue hits $109.9B, cloud business up 63% YoY to $20B, net profit surges 81% to $62.6B. Gemini Enterprise paid users up 40% QoQ — AI has officially become Google core revenue engine.
Based on Q1 earnings data, Google, Meta, Amazon, and Microsoft are expected to collectively spend over $600 billion on AI infrastructure in 2026. Google raised CapEx guidance to $180-190B, Meta to $125-145B.
Defense AI sector sees record capital inflow in 2026 Q1: Shield AI closes $2B Series G at $12.7B valuation; Anduril Industries seeks $60B valuation for new round. AI autonomous systems are reshaping global defense industry.
AI financial analysis platform Rogo closed a $160M Series D led by Kleiner Perkins, with valuation jumping from $750M in January 2025 to $2B. The company now serves 35,000+ financial professionals and 250+ institutional clients.
San Francisco defense startup Scout AI raised $100M in Series A funding, co-led by Align and Draper. The company is building AI models capable of commanding autonomous military assets like unmanned ground vehicles, marking accelerated commercialization of AI in defense.
According to the Financial Times, SoftBank plans to establish an independent AI robotics and data center company called Roze in the US, targeting a $100 billion valuation for a potential 2026 IPO. This represents the latest piece of Masayoshi Son AI strategy, integrating ARM, OpenAI investments, and self-built computing power.
SpaceX, OpenAI, and Anthropic plan to list sequentially in 2026 with combined target valuations of approximately $4 trillion. This is the mostconcentrated trillion-dollar IPO wave in history.
Anthropic completed a new $2 billion round led by Spark Capital, bringing its valuation to $20 billion.
OpenAI's pre-IPO valuation exceeds $1 trillion, but latest reports show revenue and user growth falling short of internal targets, sending a cooling signal to the overheated AI investment narrative.
Global AI funding was highly active in Q1 2024, with total financing exceeding $30 billion.
OpenAI, together with TPG, Bain, Advent and other PE firms, established DeployCo joint venture with $500M initial commitment and up to $1.5B total, targeting a $10B valuation to accelerate enterprise AI adoption.
China's NDRC, under the foreign investment security review mechanism, blocked Meta's approximately $2.5 billion acquisition of general AI agent company Manus, ordering the deal to be revoked. This is a landmark event in global AI M&A regulation.
Anthropic's valuation surged from ~$120B in October 2025 to $800B–$1T, with ARR doubling from $9B to $19B in under four months. The company plans to launch its IPO in Q4 2026.
Tesla CFO announced 2026 capex exceeding $25B for six new/expanded factories, AI infrastructure, and Austin semiconductor facility. Optimus robot mass production plan advances, marking traditional automakers AI transformation.
AI coding agent company Cognition AI is negotiating a new funding round that could value it at over $25 billion, doubling its previous valuation. The coding agent sector sees surging capital interest.
Q1 2026 saw $242B in AI company funding globally, representing ~80% of total VC. But OpenAI, Anthropic, xAI, and Waymo alone captured $188B, leaving 6,000 other startups to split the remainder.
Google will invest $10B in Anthropic at a $350B valuation, with up to $30B additional on performance targets. The deal covers 5GW of compute and 1M TPU chips, forming a deep compute-model binding.
Anthropic has hired Wilson Sonsini to prepare for an IPO as early as 2026, potentially becoming the largest AI company listing. Beating OpenAI to the public markets could reshape AI industry capital dynamics.
