How do you know an AI system is good enough to use? Evaluation as the acceptance test, with a worked example and three questions.
AI Evaluation as Acceptance Test: How to Know It’s Good Enough Read More »
What happens when software becomes cheap enough to throw away: discardable software, a worked example, and what still has to last.
Discardable Software: When Code Is Cheap Enough to Throw Away Read More »
What an AI data centre actually produces: tokens. Why tokens per watt is the number that matters when power is the limit, with a worked example.
Tokens per Watt: What an AI Data Centre Actually Produces Read More »
Why a faster AI chip does not always mean faster AI: the memory wall, memory bandwidth, and a worked example with round numbers.
The Memory Wall: Why a Faster AI Chip Is Not Faster AI Read More »
A US jury ruled against Elon Musk in his lawsuit against OpenAI, clearing a major obstacle to what could become one of the largest IPOs in history. Strategic analysis of what this means for Microsoft, Anthropic, Google, and the AI industry.
OpenAI vs Musk: The Verdict That Clears the Path to IPO Read More »
Nvidia just did something that most investors missed entirely. Buried in its latest 10-K filing, the company replaced its four legacy business segments — Data Center, Gaming, Professional Visualization, and Automotive — with just two: Data Center and Edge Computing. This is not a press release refresh. This is a valuation instrument. And the playbook
Nvidia vs Google: Why Nvidia Just Rewrote Its Own Business — And What It Signals Read More »
A US jury ruled against Elon Musk in his lawsuit against OpenAI, clearing a major obstacle to what could become one of the largest IPOs in history. Strategic analysis of what this means for Microsoft, Anthropic, Google, and the AI industry.
OpenAI vs Musk: The Verdict That Clears the Path to IPO Read More »
Anthropic and the Gates Foundation committed $200M for AI in health and education. OpenAI pursues ubiquity through Malta, Plaid, and consumer scale. Two theories of trust, two paths to regulated-industry dominance.
Anthropic vs OpenAI: Two Models for Winning Trust in AI Read More »
Fractile, a UK-based AI hardware startup, just raised $220 million to build silicon designed exclusively for inference. Not training. Not general-purpose GPU compute. Pure inference acceleration. Business Engineer Beyond the Nvidia Tax — the compute layer economics Read → This is a direct challenge to Nvidia’s dominance and a signal that the AI infrastructure market
Fractile vs Nvidia: The $220M Bet That Inference Needs New Silicon Read More »
Two deals reveal competing enterprise AI distribution strategies: Anthropic expands through PwC consulting, OpenAI embeds through Fiserv banking infrastructure. A strategic analysis of which model wins.
Anthropic-PwC vs OpenAI-Fiserv: The Enterprise AI Distribution War Read More »
Publicis agreed to acquire LiveRamp for $2.2B, buying the identity-resolution layer that powers agentic marketing. WPP chose a different path. A strategic analysis of two diverging bets on AI marketing infrastructure.
Publicis vs WPP: The $2.2B LiveRamp Deal That Reshapes AI Marketing Read More »
OpenAI just made its boldest consumer move yet. By connecting ChatGPT — as explored in the intelligence factory race between AI labs — to 12,000+ financial institutions via Plaid, the company is no longer just building an AI assistant — it is building a financial operating system. US Pro users can now ask ChatGPT about
OpenAI vs Mint: How ChatGPT Is Coming for Your Bank Account Read More »
Google and Blackstone are creating a standalone TPU cloud company — a joint venture where Blackstone contributes $5 billion in equity for a majority stake, with the first 500MW of compute capacity coming online in 2027. This is not a routine cloud partnership. It is a structural move to distribute Google’s proprietary AI chips outside
Google vs Nvidia: The TPU Cloud Play That Changes AI Infrastructure Read More »
NextEra Energy is acquiring Dominion Energy in a $66.8B all-stock deal driven by AI data center electricity demand. This megamerger signals that power — not chips — is becoming AI’s most critical bottleneck.
NextEra vs Dominion: The $67B Merger That Proves AI Runs on Power, Not Just Chips Read More »
