The AI Arms Race Just Shifted Into Overdrive: $330B+ Deployed in Two Weeks

Key Takeaways

1. Anthropic closed a $30 billion Series G at a $380 billion valuation — the second-largest private funding round ever — while OpenAI is finalizing a $100B raise that could value it above $850 billion.
2. Meta signed a $100B+ chip deal with AMD for 6 gigawatts of AI compute capacity and expanded a multi-year, multi-billion-dollar infrastructure partnership with Nvidia spanning millions of GPUs.
3. Google unveiled new data centers in Texas and Minnesota as part of a $175–$185 billion CapEx plan for 2026, while Anthropic targets 10 GW of proprietary data center capacity — on par with Amazon Web Services.
4. Big Tech collectively plans to spend roughly $650 billion on AI infrastructure this year, nearly doubling the $410 billion spent in 2025.

Quick Recap

In an extraordinary two-week blitz that may define 2026 as the year AI went industrial, the world’s largest technology companies committed over $330 billion in capital raises, chip deals, and infrastructure partnerships. From record-breaking private funding rounds to unprecedented semiconductor agreements and new data center campuses, the AI buildout has shifted from strategic planning to full-scale deployment. The developments were tracked and compiled by financial analyst @SmallCapSnipa on X (formerly Twitter).

Inside the Mega-Deals: A Breakdown

Anthropic’s $30B Series G

Anthropic closed its $30 billion Series G on February 12, led by Singapore’s GIC and Coatue Management, with co-leads including D. E. Shaw Ventures, Founders Fund, ICONIQ, and MGX. The round — which includes earlier commitments from Microsoft and Nvidia — brings the Claude-maker’s total funding to nearly $64 billion since its 2021 founding. Anthropic’s annualized revenue has surged to $14 billion, up from roughly $10 billion in 2025, with approximately 80% of revenue coming from enterprise customers.

OpenAI’s $100B Raise

OpenAI is finalizing the largest private funding round in history: more than $100 billion at a post-money valuation exceeding $850 billion. Amazon leads with up to $50 billion, followed by SoftBank at $30 billion, Nvidia at $20 billion, and Microsoft. The company plans to use the capital for chips, data center capacity, and long-term compute access — even as it expects to lose $14 billion in 2026 and does not anticipate profitability until 2029.

Meta × AMD: $100B Chip Deal

Meta Platforms agreed to acquire 6 gigawatts of AI computing capacity from AMD in a deal exceeding $100 billion, announced February 24. AMD will deliver its upcoming MI450 flagship chips, with the first gigawatt shipping in H2 2026. In a unique twist, Meta will receive warrants for 160 million AMD shares (roughly 10% of the company), vesting as shipments and stock price targets are met.

Meta × Nvidia: Data Center Tech Partnership

One week earlier, Meta expanded a multiyear strategic partnership with Nvidia to deploy millions of Blackwell and next-generation Vera Rubin GPUs across hyperscale data centers. The deal also covers Nvidia’s standalone Grace CPUs — a first at this scale — and Spectrum-X Ethernet networking switches. Meta CEO Mark Zuckerberg framed the partnership around delivering “personal superintelligence to everyone in the world”.

Google’s Texas and Minnesota Data Centers

Google announced two new data center campuses — in Wilbarger County, Texas and Pine Island, Minnesota — as part of a broader plan to invest $40 billion in Texas alone through 2027. The Texas facility will feature “advanced air-cooling” to eliminate operational water use, a response to growing community concerns around data center resource consumption. Google’s 2026 CapEx guidance sits between $175 and $185 billion.

Anthropic’s 10 GW Ambition

Anthropic has an internal plan to secure up to 10 gigawatts of proprietary computing power, recruiting veteran data center talent including former Google executive Brett Rogers and Tim Hughes of Stack Infrastructure. Achieving the 10 GW target would place Anthropic on par with AWS and surpass OpenAI’s currently disclosed capacity agreements, though it would require hundreds of billions in investment. The company also committed to covering electricity price increases for consumers near its facilities.

The AI Infrastructure Arms Race: Why Now?

The timing of these announcements is no coincidence. Several converging forces are pushing AI companies into an all-out infrastructure sprint.

Compute demand is outstripping supply. Bridgewater Associates warns that the AI boom has entered a “more dangerous phase,” with exponentially rising investments and growing reliance on outside capital. Microsoft alone disclosed an $80 billion backlog of Azure AI orders that cannot be fulfilled due to power constraints.

Revenue growth is justifying the bets — for now. Anthropic’s annualized revenue hit $14 billion, up 40% year-over-year. OpenAI reached a $20 billion annualized run rate by end of 2025, with ChatGPT surpassing 800 million weekly users. Meta’s total revenue hit $59.89 billion in Q4 2025 alone, up 24% year-over-year.

IPOs are on the horizon. Both Anthropic and OpenAI are reportedly considering IPOs in 2026, adding urgency to demonstrate scale and infrastructure readiness. OpenAI is eyeing a potential Q4 2026 listing that could push its valuation past $1 trillion.

Geopolitics and energy are becoming competitive moats. Companies are no longer just competing on model quality — they are competing on access to power, land, chips, and permitting. Meta’s $600 billion U.S. infrastructure commitment through 2028 and Google’s 7,800 MW of contracted Texas grid capacity reflect a new era where physical resources define AI leadership.

Competitive Landscape & Comparison

AI Infrastructure Investment at a Glance

AI Model Comparison: Claude vs. GPT vs. Gemini

While Anthropic’s Claude leads in enterprise coding and long-form reasoning tasks, Google’s Gemini remains the most cost-effective option for high-volume API users at $1.25 per million input tokens. OpenAI’s GPT-4o occupies the middle ground, balancing broad multimodal capabilities with moderate pricing, though its 128K context window is the smallest among frontier models.

Sci-Tech Today’s Takeaway

I’ve been covering AI infrastructure for three years now, and I can say without hesitation: this two-week stretch feels like a genuine inflection point. What strikes me isn’t just the numbers — though $330 billion deployed in fourteen days is staggering — it’s the nature of the commitments. These aren’t research grants or speculative bets. These are hardware shipments, land deals, power contracts, and equity-linked procurement agreements. The AI race has moved from the lab to the construction site.

I think the Meta-AMD deal is the single most strategically significant move in this batch. Not because of its size, but because Meta is essentially becoming a 10% stakeholder in one of its own chip suppliers while simultaneously deepening ties with Nvidia and reportedly renting Google TPUs. That’s a company hedging every possible compute bet at once. In my experience, that kind of aggressive multi-sourcing only happens when a company genuinely believes supply will be the bottleneck for years to come.

As for the valuations — $380 billion for Anthropic, $850 billion for OpenAI — I’m cautiously bullish. The revenue growth is real. But Bridgewater’s Greg Jensen is right to flag the “dangerous phase” warning. Neither company is profitable. Both need product breakthroughs to justify these numbers ahead of potential IPOs. If there’s a sharp market correction, the capital pipeline could tighten fast.​

My bottom line: this is overwhelmingly bullish for AI adoption in 2026. The infrastructure being laid right now — data centers, chips, energy contracts — will unlock capabilities we haven’t seen yet. For investors, the supply chain (energy, semiconductors, networking) looks like the safest way to play this trend. For everyday users, expect AI tools to get dramatically faster, cheaper, and more capable by year’s end. The buildout is real, and it’s accelerating.