🕶️ TensorWave: $350M to ditch NVIDIA

The real scarcity isn't chips anymore; it's control over where they live, how they're powered, and who decides what gets built next. Two continents, two data center plays, same pattern: the people who own energy, land, and cooling infrastructure are suddenly as strategically important as the people who design the silicon.

What started as a supply crisis is turning into a geography game, and the founders who figure that out early will route around everyone else's waitlist.

Hiring great engineers shouldn’t feel like finding a needle in an AI-generated haystack. G2i connects you with pre-vetted contract and full-time developers—no scammers, no endless resume piles, just top talent with 5+ years of experience across every major tech stack.

We’ve already helped teams at Meta, Microsoft, and fast-growing startups like Shopmonkey and 1Password hire faster, smarter, and with full transparency through recorded technical interviews.

Whether you need one engineer or a whole team, we’ve got over 8,000 ready to go.

TensorWave just turned AMD capacity into a venture-scale asset.

The company raised $350 million at a $1.55 billion valuation to expand its AMD-powered AI infrastructure. The important part isn’t just the round size. It’s that investors are now backing a full alternative compute stack: bare metal, managed services, enterprise compliance, and AMD Instinct chips from MI300X up through MI455X.

That’s a real market signal. AMD is no longer the “nice-to-have” second source in AI infra. It’s becoming a strategic supplier for startups and enterprises that don’t want their roadmap set by one chip vendor or one hyperscaler’s pricing model.

For founders, the takeaway is simple: if you’re building in AI infra, plan for large capex, supply-chain risk, and more pressure on AMD allocation. If you’re selling to enterprise, certifications like ISO 27001, SOC2 Type II, and HIPAA are starting to look less like polish and more like table stakes.

The less obvious ripple: rounds like this can tighten AMD inventory for everyone else. If your roadmap depends on Instinct supply, your procurement strategy now matters almost as much as your product.

Meta isn’t just adding capacity in India. It’s helping turn India into an AI compute export market.

In its first India AI data center deal, Meta partnered with Reliance to lease capacity at a 168-megawatt facility in Jamnagar. The setup matters: renewable power, desalinated seawater cooling, and Meta covering the full energy and water bill. Reliance says it will be live in two years and expandable from there.

The signal is bigger than one site. This extends a relationship that already includes Meta’s $5.7 billion Jio investment and a $100 million enterprise AI JV. It also sits inside a much larger buildout: India’s data center capacity rose from roughly 375 megawatts in 2020 to 1.5 gigawatts in 2025, with projections above 8 gigawatts by 2030.

For founders, the takeaway is practical. India is becoming a real option for lower-cost, renewable-backed training and inference, especially for global workloads that don’t need to sit in the US. Add New Delhi’s tax incentives through 2047, and this starts looking less like local infrastructure and more like a new compute geography. The next bottleneck may not be GPUs. It may be access to power, partners, and priority allocation.

Scale your growth: Join names like HubSpot, Ahrefs, and Deel. Put your product in front of 85,000+ startup founders. Let's talk.