The CTO Show Brief: Issue 150

Author

Mehmet Gonullu
December 21, 2025

Welcome to Issue 150 of The CTO Show Brief.

The final weeks of 2025 delivered a decisive signal: the experimental phase of generative AI is ending and the deployment phase has begun. Capital is no longer chasing models—it's chasing the physical infrastructure to run them. This week's moves reveal a tech economy where the binding constraints are joules, silicon, and borders, not algorithms.

Whether you’re a long-time reader or a first-time reader—grab your coffee, and let’s dive into what’s shaping tech and venture this week!

Sub-Dollar Inference Unlocks the Agentic Economy

Google's Gemini 3 Flash reprices the inference layer: $0.50 per million input tokens, 1M-token context, and 3x speed gains over prior generations. This isn't a discount—it's an enabling technology. Multi-turn agentic loops that were cost-prohibitive in 2024 are now viable at consumer scale. The shift is from 'AI as feature' to 'AI as continuous loop.' Developers building always-on agents—coding assistants, autonomous workflows, robotic control systems—gain immediate leverage. The winners are infrastructure-light application builders who can now run inference-heavy products without burning runway.

Automakers Break from Nvidia—Vertical Integration Accelerates

Rivian unveiled its custom RAP1 chip and Large Driving Model, signaling a broader industry move to reclaim the compute stack. Built on TSMC's 5nm process with claimed 4x performance over Nvidia's Drive Orin, the silicon underpins a $49.99/month autonomy subscription priced to maximize attach rates. The strategic logic mirrors Apple and Tesla: control hardware, capture margin, reduce supply chain dependency. If Volkswagen and other OEMs follow Rivian's playbook, Nvidia's automotive segment faces structural headwinds even as its datacenter business booms. For founders, the signal is clear: vertical integration is defensible; pure software wrappers are not.

Energy Becomes Critical Infrastructure for AI

Radiant Nuclear raised $300M to mass-produce 1MW portable microreactors at Oak Ridge; NextEra formalized gigawatt-scale development partnerships with Google and Meta. The message: power is the hard cap on AI scaling. Grid interconnect queues run years long—bringing generation to the datacenter bypasses the bottleneck. Tech giants are becoming utility developers, not just power buyers. The capital thesis is simple: whoever solves the last mile of energy wins the next phase of compute. Multiple shots on goal—Radiant, Last Energy, nuclear SMRs—suggest investors are hedging across the fission stack.

Sovereign Capital Reshapes Tech Geography

Databricks closed $4B+ at a $134B valuation—$34B higher than three months ago—with MGX (Abu Dhabi) joining Insight and Fidelity. Days later, Databricks announced $300M into Saudi Arabia and a Riyadh HQ. This is circular capital: sovereign wealth funds invest in US infrastructure, and US tech reinvests in Gulf digital buildout. The pattern extends to nuclear (US-Saudi 123 Agreement talks), talent (UAE absorbing H-1B friction spillover), and data (Al Jazeera's Google Cloud deployment wrapping Gemini in 30 years of proprietary archives). For founders, the calculus has changed: Silicon Valley is no longer the only viable HQ. Capital and talent are becoming geopolitically distributed.

Talent Friction Creates Arbitrage Opportunities

US H-1B social media vetting pushed visa appointments to 2027, creating immediate operational drag for startups dependent on high-skill migration. The policy inadvertently stimulates rival hubs—Canada, UK, UAE—and remote-work infrastructure. Meanwhile, India's startup ecosystem raised $10.5B in 2025 (third globally), increasingly from domestic capital. The brain-drain risk for US tech is rising as friction increases. Operators should map talent pipelines now; investors should track secondary hub deal flow.

The question for 2026: Who owns the stack—chips, joules, data—and who merely rents it?

🎙️Episodes Recap:

In this episode of The CTO Show with Mehmet, I’m joined by Alexander Schlager , Founder and CEO of AIceberg, a company operating at the intersection of AI, cybersecurity, and explainability. We dive deep into why AI agents fundamentally change enterprise risk, how shadow AI is spreading across organizations, and why monitoring black-box models with other black boxes is a dangerous mistake.

In this episode of The CTO Show with Mehmet, I sit down with Khaled Nazif , COO of DSquares, one of the most influential yet quietly powerful enterprise loyalty platforms in the MENA region. Khaled shares his journey from Stanford and Silicon Valley back to the region, where he helped scale DSquares into a 150M+ end-user platform serving banks, telcos, governments, and large enterprises across 16 countries.

In this episode of The CTO Show with Mehmet, I’m joined by Ahikam Kaufman A, Co-Founder and CEO of Safebooks.ai, a seasoned finance executive turned entrepreneur with deep experience across startups, public companies, and large-scale acquisitions. We explore why finance has lagged behind other functions in digital transformation, how AI is fundamentally reshaping financial governance, and why the modern CFO is becoming a transformation leader, not just a financial steward.

 

📖 From Nowhere to Next

 

Every week I share startup lessons and stories through The CTO Show Brief. But if you want to go deeper, my book From Nowhere to Next brings together the experiences and insights that shaped my own journey.

 Thanks for reading — and for being part of this growing, global-minded network.

 — Mehmet