Scale & Strategy
together with
This is Scale & Strategy, the daily newsletter that's like finding cash in an old jacket pocket, but every day.
Here’s what we’ve got for you today:
- Google's AI Reset: 3 Things Gemini Needs
- SoftBank Just Borrowed $10 Billion Against OpenAI
Google's AI Reset: 3 Things Gemini Needs
The company that helped spark the modern AI revolution just hit the reset button.
On Wednesday, Google DeepMind announced a major leadership shakeup as OpenAI, Anthropic, and a growing wave of Chinese labs continue pulling ahead in advanced models and agents.
Koray Kavukcuoglu, Google's chief AI architect, was promoted to SVP of Google DeepMind and will now report directly to CEO Sundar Pichai. Demis Hassabis is moving from CEO of DeepMind to Chairman of DeepMind and Chief Scientist of Alphabet, where he'll focus on AGI and scientific discovery. Meanwhile, longtime Google AI leader Jeff Dean is leaving the company altogether to launch a new startup, Discovery Loop.
The changes come at an uncomfortable moment for Google.
After briefly capturing the industry's attention with Gemini 3 last fall, the company has struggled to maintain momentum. Gemini 3.5 Pro arrived slowly, while OpenAI, Anthropic, and several Chinese labs have shipped major improvements at a relentless pace. At the same time, Google has largely missed the agentic coding wave that has become one of the defining stories of AI in 2026.
That's a difficult position for a company that invented the transformer architecture, employs some of the world's best researchers, and has virtually unlimited access to compute and infrastructure.
If Google is serious about regaining momentum, there are three areas it needs to address.
1. Build a Clear Answer to Claude Code
Nothing has defined the AI market this year more than agentic coding.
Anthropic's Claude Code became the product that transformed AI from something users chat with into something that actively gets work done. OpenAI's Codex has since caught up and, in many areas, surpassed it. Together, they've become the flagship products of the agent era.
Google, meanwhile, has Gemini CLI and Antigravity.
The problem isn't necessarily that either product is bad. The problem is that users still don't know which one matters. Having multiple overlapping products that solve similar problems creates confusion precisely when competitors are building iconic brands around a single clear experience.
Google doesn't just need a coding agent.
It needs a coding agent people can immediately identify as Google's answer to Claude Code.
2. Simplify the Brand Mess
Google's AI portfolio has become nearly impossible to follow.
Even people who spend all day covering the AI industry struggle to keep track of the company's growing collection of brands.
There's Gemini. Then Gemma. Then Genie. Antigravity. Google AI Studio. Firebase Studio. Opal. Flow. Lyria. Veo. Nano Banana. And that's before you get into all the product-specific integrations scattered across Google's ecosystem.
At some point, the problem stops being product development and starts becoming organizational sprawl.
Google appears to be shipping its org chart.
Instead of presenting a unified AI strategy, it often feels like individual teams are launching independent brands that happen to live under the same corporate umbrella.
Users don't want to understand Google's internal structure.
They want to know which product solves their problem.
3. Become More Like Copilot
This recommendation may be painful inside Google.
Microsoft spent years playing catch-up to Google. Now, in some important ways, Google should be paying attention to Microsoft's playbook.
Microsoft isn't winning because it has the best models.
It isn't winning because it has the best coding agents.
It's winning because Copilot is everywhere.
Whether you're using Word, Excel, Outlook, Teams, Windows, or GitHub, the same brand shows up. Users understand what Copilot is and where to find it.
Google has many of the same advantages.
It controls Android. It controls Workspace. It has billions of users touching its products every day.
Yet Google's AI experience often feels fragmented compared to Microsoft's increasingly unified approach.
The lesson isn't about copying Copilot's features.
It's about copying its simplicity.
Google Still Has a Path Back
Despite the recent struggles, writing Google off would be a mistake.
The company still has assets that no startup can replicate.
It has Android. It has Workspace. It has YouTube. It has Search. It has one of the deepest benches of AI researchers on the planet. And perhaps most importantly, it still has Google Research, the organization responsible for many of the breakthroughs that made today's AI boom possible in the first place.
The problem isn't talent.
The problem is urgency.
Over the past six months, Google has looked less like a company fighting for the future and more like a giant corporation trying to manage it.
Meanwhile, OpenAI, Anthropic, the Chinese labs, and a new generation of startups are operating like every release matters and every month counts.
That's the mindset Google needs to rediscover.
Because the next phase of the AI race won't be won by the company with the best history.
It will be won by the company that moves fastest from here.
Why it matters: Google still has the talent, compute, distribution, and research capabilities to remain a frontier AI leader. But leadership changes alone won't solve its problems. To regain momentum, the company needs clearer products, simpler branding, and a stronger sense of urgency. The biggest threat to Google isn't OpenAI or Anthropic. It's behaving like a mature tech giant while the rest of the industry acts like a startup.
Risk and regulations are ramping up, and customers expect proof of security just to do business. That’s why Vanta is a game changer.
Vanta automates your compliance and security processes in one AI-powered platform.
Whether you’re prepping for SOC 2, or running an enterprise GRC program, Vanta keeps you secure and your deals moving.
Companies like Ramp and Writer spend 82% less time on audits with Vanta. Faster compliance—and more time for growth.
Check out the on-demand demo to learn more. – trusted by top startups like Cursor, Linear, and Replit.
SoftBank Just Borrowed $10 Billion Against OpenAI
Masayoshi Son is doubling down.
SoftBank announced this week that it has borrowed $10 billion using its OpenAI stake as collateral, one of the largest debt-backed bets ever placed on a private AI company.
The move helps fund another $10 billion investment into OpenAI, the final installment of the $30 billion commitment SoftBank announced earlier this year. By October, SoftBank will have invested roughly $64 billion into OpenAI and own about 13% of the company.
That alone would be notable.
What's more interesting is what it says about where the AI race is heading.
The financing was backed by some of the biggest names in global finance, including Goldman Sachs, JPMorgan, Apollo, Mizuho Securities, and Sumitomo Mitsui. Using a private company stake to secure a loan of this size is unusual. Margin loans are common for publicly traded stocks, where lenders can easily monitor valuations and liquidate positions if needed. OpenAI isn't public, isn't profitable, and doesn't have an investment-grade credit rating.
In other words, this is not a normal financing arrangement.
If OpenAI's valuation falls, SoftBank may be required to contribute additional capital. The structure effectively ties more of SoftBank's balance sheet to the success of a single company.
And that's exactly what SoftBank wants.
While most major investors have spread their bets across OpenAI, Anthropic, Google, xAI, and a growing list of AI startups, SoftBank is doing the opposite.
"It's our choice to put all of our eggs in OpenAI," SoftBank executive Mark Agne said this week.
That's becoming increasingly clear.
The company isn't just investing in OpenAI. It's building infrastructure around OpenAI.
SoftBank plans to begin construction this year on the first phase of its massive Ohio AI campus, a project that could eventually reach 10 gigawatts and become the largest AI data center complex ever built. Reports indicate OpenAI is in talks to lease the entire facility, while Nvidia has reportedly discussed guaranteeing as much as $250 billion of OpenAI-related obligations tied to the project.
SoftBank has also signed a separate lease with OpenAI for a 1.2-gigawatt data center campus elsewhere in the U.S., despite the fact that OpenAI remains years away from profitability by its own projections.
The result is a growing web of financial relationships where the same handful of companies are simultaneously investing in one another, financing one another, leasing infrastructure to one another, and guaranteeing one another's obligations.
It's a level of interconnectedness the tech industry hasn't seen since the telecom buildout era.
The timing is also notable.
When SoftBank first backed OpenAI, the company was widely viewed as the undisputed leader in generative AI. Since then, the competitive landscape has become much more crowded. Anthropic has gained significant ground, growing revenue rapidly while raising substantially less capital. Google continues to invest heavily. Chinese labs are releasing increasingly capable open models.
Despite that, SoftBank isn't diversifying.
It's concentrating.
The bet isn't simply that OpenAI will win. The bet is that OpenAI will become the foundation upon which much of the AI ecosystem is built.
Whether that proves visionary or reckless will likely determine the outcome of one of the largest technology investments ever made.
Why it matters: AI is no longer just a technology story. It's becoming a capital markets story. SoftBank borrowing $10 billion against its OpenAI stake shows how much confidence investors have in the future of frontier AI, but it also highlights how concentrated those bets are becoming. If OpenAI succeeds, SoftBank could look brilliant. If it stumbles, the ripple effects will extend far beyond a single company.
Was this email forwarded to you?
That’s it for today and as always It would mean the world to us if you help us grow and share this newsletter with other operators.
Our mission is to help as many business operators as possible, and we would love for you to help us with that mission!
