February 2026:
The Month That Redefined Capital

February 2026 defied every historical benchmark. The $195B+ in AI-tracked venture capital deployed in a single month approaches 60% of total global VC for all of 2024 (~$330B). But the headline number obscures a more interesting story beneath it.

Yes, OpenAI's $110B Series G and Anthropic's $30B Series G dominated. Together they represent $140B, or 72% of the month. But even excluding these two deals, February's remaining activity, roughly $55B across autonomous driving, robotics, semiconductors, cybersecurity, fintech, and energy, still matches January 2026's record month.

The real signal is not the top-line number. It's that the capital is no longer primarily flowing to model labs. February saw massive commitments to the physical infrastructure that AI requires: chips (Cerebras, Axelera), autonomous vehicles (Waymo, Wayve), humanoid robots (Apptronik, Bedrock), energy (SHINE, Heron Power), and sovereign compute (Neysa). The market is treating AI the way it treated the internet in 1999: as something that needs to be built, not just imagined. Whether it's pricing that correctly is the open question.

February's numbers look euphoric. The sentiment beneath them is more complicated.

The concentration problem. Two deals (OpenAI + Anthropic) represented 72% of all capital. Remove them and February looks like a strong but not unprecedented month. This extreme concentration raises the question of whether "venture capital" is even the right framework for $100B+ rounds that look more like infrastructure spending. When Amazon commits $50B to a single company, is that VC? Or is it a corporate procurement decision with equity attached?

The burn rate question. OpenAI forecasts $14B in losses for 2026 despite projecting $100B revenue by 2029. Anthropic reports $14B+ ARR but hasn't disclosed profitability. The model labs are spending faster than any startups in history, with the assumption that AI infrastructure requires the same scale of upfront investment as cloud computing or telecom networks. If the revenue materializes, these are generational bets. If not, the write-downs will be equally generational.

Physical AI breaks through. The defining shift of February wasn't the model lab mega-rounds (those were expected). It was the emergence of physical AI as a capital category rivaling software. Autonomous driving ($18B+ across Waymo and Wayve), robotics ($800M+ across Apptronik and Bedrock), energy ($400M+ across SHINE and Heron), and sovereign compute ($1.2B for Neysa) all represent bets on atoms, not bits. This is the most capital-intensive phase of AI investing yet.

Application layer velocity accelerates. Basis reached unicorn status in 28 months. Profound did it in 18. Code Metal jumped 5x in valuation in 3 months. Vega Security raised three rounds in one year. The application layer is now producing unicorns faster than the infrastructure layer did two years ago, suggesting that AI's value is shifting toward domain-specific solutions with real revenue.

The backlash is real but irrelevant (so far). Google warned that LLM wrappers may not survive. ~90% of firms reported no AI productivity gains. Bernie Sanders called for a data center moratorium. SBA restricted startup loans to citizens only. Yet capital kept flowing. The gap between mainstream skepticism and investor conviction widened further in February. Whether the money or the skeptics are right will define the next 12 months.

February set an impossible bar. March will reveal whether this is a new normal or a peak. Key signals to watch:

• OpenAI's execution. With $110B in fresh capital and $14B in projected losses, execution against the $100B revenue target by 2029 becomes the single most important data point in venture. Early enterprise adoption metrics and developer platform growth will be closely watched.
• Physical AI deployment timelines. Wayve, Apptronik, and Bedrock all raised at valuations that assume near-term commercial deployment, not indefinite R&D. Watch for partnership announcements and pilot programs with the automakers and manufacturers that invested.
• Energy infrastructure deals. Google's $1B Form Energy purchase and SHINE's $240M raise opened a new investment category. March may see more AI companies making energy procurement deals alongside traditional fundraising.
• IPO pipeline. Form Energy ($3.5B, targeting 2027 IPO), IQM Quantum ($1.8B SPAC), and Generate Biomedicines ($400M filing) suggest the public markets window may finally be opening. Whether these proceed will shape exit expectations across the industry.
• SBA policy impact. The March 1 restriction on SBA loans to citizens only will be felt first in early-stage rounds. Watch for shifts in angel, pre-seed, and seed activity, particularly among immigrant-founded companies.