$410M+ AI Funding Surge & Davos 2026 AI Predictions | January 23, 2026

Daily AI Blog

January 23, 2026

📋 Quick Takeaways

$410M+ in major AI funding including Inferact ($150M seed at $800M valuation) and Neurophos ($110M Series A for photonic chips)
Elon Musk at Davos 2026 predicts AI smarter than any human by end of 2026, warns electrical power is limiting US AI deployment
OpenAI introduces ads to ChatGPT, surprising Google DeepMind CEO who says Google is thinking through monetization “very carefully”
Apple reportedly training next-gen Siri on Google Cloud TPUs, signaling major shift in Big Tech compute partnerships
Anthropic CEO criticizes H200 chip sales to China, comparing decision to “selling nuclear weapons to North Korea”
Jensen Huang declares AI creates trade jobs rather than destroying employment, with plumber/electrician salaries nearly doubling

🚀 Major AI Funding & Infrastructure Investment

Inferact Raises $150M Seed at $800M Valuation for AI Inference Platform

Breakthrough Funding: The commercial startup behind open-source vLLM secured $150 million in seed funding on January 22, 2026, co-led by Andreessen Horowitz and Lightspeed Venture Partners.

Strategic Significance:

Valued at $800 million in seed stage, reflecting massive AI inference market opportunity
Focuses on making AI model inference faster and more cost-efficient for production deployments
Major cloud providers including Amazon Web Services already testing the platform
Additional investors: Sequoia Capital, Altimeter Capital, Redpoint Ventures, ZhenFund

Technology Impact: Inferact’s platform addresses the critical shift from AI model training to deployment, where companies need lower-latency, lower-cost inference solutions. The vLLM project has over 2,000 code contributors and is becoming industry standard for optimizing large language model performance.

Market Context: As AI companies transition from building models to running them in production at scale, inference optimization becomes increasingly valuable. The $800M valuation at seed stage underscores investor conviction in this infrastructure layer.

Source: Tech Startups | SiliconANGLE

Neurophos Secures $110M Series A for Revolutionary Photonic AI Chips

Photonics Breakthrough: Austin-based Neurophos raised $110 million in Series A funding led by Gates Frontier (Bill Gates’s fund), with participation from Microsoft’s M12, Carbon Direct Capital, Aramco Ventures, and Bosch Ventures on January 22, 2026.

Technology Innovation:

Develops optical processing units (OPUs) integrating over 1 million micron-scale optical processing elements
Claims up to 100x performance and energy efficiency gains versus current leading chips
Uses photons instead of electrons to bypass speed and energy limitations
Achieves over 300 trillion operations per second per watt in early tests

Commercial Timeline:

Pilot deployment with Norwegian data center operator Terakraft in 2027
First complete systems manufacturing by early 2028
Production scaling planned for late 2028

Industry Validation: Microsoft’s involvement through both M12 venture fund and strategic endorsement from Dr. Marc Tremblay (Corporate VP, Core AI Infrastructure) signals confidence in photonics as next-generation compute paradigm beyond traditional silicon GPUs.

Source: PR Newswire | SiliconANGLE | Yahoo Finance

Railway Raises $100M Series B for AI-Native Cloud Infrastructure

Developer Platform Expansion: Austin-based cloud platform Railway announced $100 million Series B on January 22, 2026, led by TQ Ventures with participation from FirstPath, Redpoint Ventures, and Unusual Ventures.

Platform Focus:

Developer-friendly infrastructure management for AI/ML workloads
Addresses data and networking bottlenecks hindering large AI model scaling
Total funding now reaches approximately $124 million
Targets hybrid-cloud and multi-cloud AI deployments for enterprises

Market Opportunity: Railway positions itself as challenger to AWS and other hyperscalers by offering AI-native infrastructure that simplifies deployment and scalability specifically optimized for machine learning workloads.

Source: Tech Startups | BEAMSTART

Additional AI Funding Rounds (January 22, 2026)

Superstate Raises $82.5M for Tokenized Securities London-based tokenization platform Superstate secured $82.5 million Series B co-led by Bain Capital Crypto and Distributed Global. The company manages over $1.1 billion in tokenized U.S. Treasury funds and is launching “Opening Bell” platform for tokenized stock issuance on Ethereum and Solana.

Cambio Raises $18M for Real Estate AI San Francisco startup Cambio raised $18 million Series A led by Maverick Ventures, with participation from Y Combinator and angel investors from OpenAI and Anthropic. The company uses large language models to analyze unstructured building data, serving the $20 trillion U.S. commercial real estate market.

Corxel Secures $287M for Biotech Development Biotech company Corxel raised $287 million Series D1 to advance oral GLP-1 obesity therapy development.

D-Orbit Locks In $53M for Space Infrastructure Space technology company D-Orbit secured $53 million Series D to expand in-space data processing capabilities.

Source: Tech Startups

🎙️ Davos 2026: AI Leadership Predictions & Warnings

Elon Musk: AI Smarter Than Humans by End of 2026

Historic Prediction: In his first-ever appearance at the World Economic Forum in Davos on January 22, 2026, Elon Musk predicted “we might have AI smarter than any human by the end of this year, I would say no later than next year.”

Key Statements:

AI intelligence surpassing smartest humans driven by exponential computing power and software advances
Robots will eventually outnumber people globally
Tesla could start selling Optimus humanoid robots next year
Future of abundance where AI and robots eliminate scarcity

Broader Vision: Musk painted future where “everyone will have a robot” and there will be “abundance for all,” solving poverty and providing high standard of living through AI-powered automation.

Source: Business Today | Euronews | CBS News

Musk Warns: Electrical Power Now Limiting AI Deployment in US

Critical Infrastructure Challenge: Musk told the World Economic Forum audience that electrical power is the fundamental limiting factor for AI advancement in the United States.

Key Warning:

“The limiting factor for AI deployment is fundamentally electrical power”
“We’re very soon—maybe even later this year—we’ll be producing more chips than we can turn on”
US hampered by outdated grid system from decades of underinvestment
China doesn’t face same constraint due to massive solar expansion

China Advantage: Musk noted China is deploying over 1,000 gigawatts of solar per year and has nearly 4x the operational solar electricity capacity of the US. He criticized high US solar tariffs making deployment “artificially expensive.”

Policy Implications: President Trump encouraged tech companies at Davos to build their own nuclear plants for AI data centers, claiming three-week approval timelines despite historical years-long processes.

Source: Fortune | WebProNews

Jensen Huang: AI Creates Trade Jobs, Doesn’t Destroy Employment

Optimistic Employment Outlook: NVIDIA CEO Jensen Huang told Davos on January 21-22, 2026, that AI will create significantly more manual trade jobs rather than eliminating employment.

Key Insights:

“We’re going to have plumbers and electricians… all of these jobs, we’re seeing quite a significant boom and salaries have gone up. Nearly double”
AI needs more energy, land power, and trade skill workers for infrastructure buildout
Europe has strong trade workforce positioning for AI infrastructure opportunity
Robotics represents “once in lifetime opportunity for European countries”

Industry Validation: Huang emphasized that instead of job displacement, AI data center construction and maintenance is driving unprecedented demand for skilled trades, with six-figure salaries becoming common.

Source: Euronews

💼 Major Corporate AI Strategy Shifts

OpenAI Begins Testing Ads in ChatGPT, Surprising DeepMind CEO

Monetization Pivot: OpenAI announced on January 22, 2026, it will begin testing advertisements in ChatGPT to generate revenue from its 800 million weekly active users who don’t have paid subscriptions.

DeepMind Response: Google DeepMind CEO Demis Hassabis told Axios at Davos he was “surprised” by OpenAI’s decision and that Google is thinking through AI monetization “very carefully.”

Strategic Context:

Hassabis emphasized DeepMind isn’t feeling pressure for “knee-jerk decisions” despite ads being core to Google’s business
ChatGPT is fundamentally different experience than Google Search, making ad integration more challenging
Users previously reacted negatively to OpenAI’s app suggestion feature, forcing its removal

Industry Implications: The move reflects OpenAI’s growing infrastructure and energy costs pressuring the company to diversify beyond subscriptions. Consumer backlash risk could impact ChatGPT’s user experience quality and competitive positioning.

Source: TechCrunch | Yahoo Finance | StartupNews.fyi

Apple Reportedly Using Google Cloud TPUs for Next-Gen Siri Training

Major Partnership Shift: Apple is reportedly using Google Cloud TPUs to train the “LLM Siri” system behind its next-generation conversational assistant, according to reports from January 22, 2026.

Strategic Significance:

Marks departure from Apple’s traditional reliance on in-house silicon and tightly controlled infrastructure
Signals pragmatic approach: when workload is massive and time-to-quality matters, Apple will rent best-in-class compute
Strengthens Google’s position as “arms dealer” for foundation-model compute infrastructure
Demonstrates AI pushing Big Tech toward “coopetition” where rivals become each other’s customers

Broader Impact: Apple workload on Google infrastructure highlights supply reality—the AI race is increasingly shaped by who can secure enough accelerators, networking, and power, not only who has best model architecture.

Competitive Dynamics: Apple and Google compete across platforms and services, yet AI compute access now dictates product velocity, potentially reordering Big Tech alliances.

Source: Tech Startups

🛡️ AI Geopolitics & Policy

Anthropic CEO Compares NVIDIA H200 Sales to China to “Nuclear Weapons”

Hawkish Stance: Anthropic CEO Dario Amodei strongly criticized the US allowing NVIDIA to sell H200 chips to China, telling Bloomberg at Davos on January 20, 2026, that the decision is comparable to “selling nuclear weapons to North Korea.”

Policy Context:

Trump administration announced in December 2025 it would allow H200 accelerator shipments to Chinese customers with 25% revenue cut to US government
Chinese authorities must still approve local buyers acquiring the GPUs
Amodei argues chip embargo is what’s holding back Chinese AI companies

Competitive Concerns:

Chinese AI firms releasing open-weight models that compete with Anthropic’s closed commercial offerings
Amodei admitted “almost never” losing contracts to Chinese models, but keyword “almost” acknowledges some competitive losses
Access to advanced chips would strengthen Chinese model developers’ enterprise adoption potential

Industry Split: Anthropic’s position contrasts sharply with chipmakers like NVIDIA and AMD, which warn that closing China access could harm US competitiveness by excluding half of world’s AI researchers.

Source: The Register | TheStreet

📊 Market Impact Analysis

The January 22-23, 2026 news cycle reveals several transformative trends reshaping the AI landscape:

Capital Concentration in Infrastructure: $410M+ flowing to just three AI infrastructure companies (Inferact, Neurophos, Railway) demonstrates investor conviction that the next phase of AI value creation lies in deployment efficiency, novel compute architectures, and developer tooling rather than foundation models alone.

Compute Paradigm Shift: Neurophos’s $110M raise for photonic chips and Apple’s reported use of Google TPUs signal the industry exploring alternatives to traditional silicon GPUs as Moore’s Law slows and energy constraints intensify.

Power Becomes Primary Constraint: Musk’s Davos warning that electrical power limits AI deployment more than chips or capital represents critical inflection point. The AI race now depends on energy infrastructure, grid modernization, and power generation capacity.

Monetization Pressure Mounts: OpenAI’s introduction of ChatGPT advertising reflects growing pressure on AI companies to generate revenue beyond venture funding and subscriptions, as infrastructure costs balloon.

Geopolitical AI Competition: Anthropic CEO’s nuclear weapons comparison for chip sales to China underscores how AI development has become central to great power competition, with export controls as primary policy lever.

Labor Market Transformation: Jensen Huang’s observation of doubled trade worker salaries challenges AI job displacement narrative, suggesting infrastructure buildout creates significant blue-collar opportunity even as white-collar roles face disruption.

🔮 Looking Ahead

Key Trends to Monitor:

AI Inference Market Consolidation: As companies shift from training to deployment, expect M&A activity and strategic partnerships around inference optimization platforms
Alternative Compute Architectures: Photonic chips, custom ASICs, and specialized accelerators may challenge NVIDIA’s GPU dominance if energy efficiency becomes paramount
Energy Infrastructure Race: Countries and companies that secure reliable, affordable power for AI data centers will gain significant competitive advantages
AI Monetization Evolution: Expect more AI companies to explore advertising, API pricing changes, and enterprise licensing models as path to profitability
Export Control Enforcement: Watch for escalating US-China AI competition manifesting through chip restrictions, talent mobility policies, and technology transfer controls

Stay Updated: Follow our comprehensive daily coverage of AI funding, technology breakthroughs, policy developments, and market analysis.

Last Updated: January 23, 2026, 8:30 PM EST