- OpenTools' Newsletter
- Posts
- OpenAI Wants $100B More đź’°
OpenAI Wants $100B More đź’°
How 2025 ends explains how 2026 begins.
Reading time: 5 minutes
🗞️In this edition
The Google Move Most People Will Miss
OpenAI reportedly wants another $100B
Microsoft Copilot rolls out GPT-5.2 “Smart Plus”
In other AI news –
AI-native cloud is becoming a real category
Microsoft’s quiet agent breakthrough: long-term memory
The AI-native utility company you didn’t expect
4 must-try AI tools
Hey there,
As 2025 comes to a close, it’s a good moment to pause—not to celebrate, but to notice what quietly became normal.
This year, AI stopped feeling experimental. It moved into habits, workflows, and expectations. Not through loud announcements, but through decisions that reshaped how work actually gets done.
This final edition of the year looks at those decisions. Where power concentrated. Where trade-offs appeared. And what that tells us about the kind of judgment 2026 will demand.
Let’s take one last clear look before the year turns.
What's happening:
Google wrapped December with a dense set of AI upgrades across its ecosystem, anchored by Gemini 3 Flash described as “frontier intelligence built for speed,” positioned to be fast, more capable at reasoning, and lower cost.
They also shipped a Gemini Deep Research agent for developers, plus a new benchmark (DeepSearchQA) to evaluate how well research agents handle real web tasks.
Why this is important:
This is not “one more model release.” It’s a strategy: make AI feel instantaneous and everywhere.
Speed changes behavior. When AI responses happen in seconds and are embedded across products, you stop treating AI like a special tool. You treat it like electricity: you don’t “use” it, you assume it’s there.
The research-agent piece matters even more long-term. Once deep research becomes a portable capability developers can embed, you get a new baseline for what apps can do by default: investigate, synthesize, decide, and explain.
Our personal take on it at OpenTools:
Here’s the real shift we see: 2026 won’t be won by the team with the most impressive demo. It’ll be won by the team that changes what people do when nobody is watching.
Google’s play is habit formation. “Frontier intelligence built for speed” is a behavioral product decision as much as a technical one.
Because when AI feels frictionless, it moves upstream into your daily decisions: what you read, what you buy, what you build, how you work, how you learn.
The Deep Research agent and the new benchmark are also a subtle confidence move. Benchmarks aren’t exciting to most people, but they’re the scaffolding for trust. When research agents become measurable and comparable, the market starts to differentiate on reliability, not just cleverness.
Our lens for 2026: ask one question whenever you see an AI release
Does this create a new habit for millions, or does it just impress a few?
Google is optimizing for the former.
What's happening:
OpenAI is reportedly in talks to raise another $100 billion, pushing its implied valuation toward $800 billion.
But the real signal isn’t the size of the round.
Inside OpenAI, leadership is said to be navigating a growing tension: how much compute to allocate to training the next frontier model, versus how much to dedicate to keeping ChatGPT fast, reliable, and always available at global scale.
This isn’t a funding story.
It’s a systems-stress story.
Why this is important:
When you’re talking about nine-figure annual burn and 12-figure raises, strategy becomes survival: what to train, what to ship, what to delay, and what to serve at scale.
And there’s a second-order effect: once numbers reach this magnitude, AI companies stop being “tech startups.” They start behaving like nations: supply constraints, bargaining power, long-term obligations, and reputational risk become part of the product.
Our personal take on it at OpenTools:
Most readers interpret “$100B raise” as “AI is unstoppable.”
We read it as: AI is becoming expensive enough that only the most strategically disciplined players will thrive.
2026 will reward teams who can do three things simultaneously:
Build frontier capability
Ship products that people actually keep using
Control cost curves without breaking trust
If you’re a founder, operator, or investor, the actionable takeaway is simple: build offerings that do not collapse if compute gets tighter or pricier. Build systems that still deliver value under constraint.
In a world where even OpenAI is weighing “train vs. serve,” the winning products will be the ones that compound utility even when resources are limited.
What's happening:
Microsoft is rolling out GPT-5.2 inside Copilot as a new “Smart Plus” mode across web, Windows, and mobile.
This isn’t a replacement.
It’s a quiet elevation of the baseline.
Why this is important:
Not through launches people debate, but through upgrades people don’t opt into. They just wake up to them.
By embedding GPT-5.2 directly into daily workflows, Microsoft is turning advanced reasoning into an ambient capability.
When intelligence is always available, expectations shift.
“Good enough” work stops being good enough.
Our personal take on it at OpenTools:
Once AI becomes the default collaborator for writing, planning, analysis, and decision-making, skill gets redefined.
The advantage won’t come from knowing how to use AI.
Everyone will.
In 2026, the rare skill won’t be speed.
It will be judgment.
AI-native cloud is becoming a real category – InfoWorld lays out what “AI-native cloud” means, from GPU-first orchestration to vector databases as “memory,” and why 2026 may favor specialized “neoclouds.”
Microsoft’s quiet agent breakthrough: long-term memory – Microsoft Foundry Agent Service added a preview “memory” layer so agents can persist preferences and context across sessions, devices, and workflows.
The AI-native utility company you didn’t expect – Kraken, spun out of Octopus Energy, is valued at $8.65B after a $1B round and already serves 70M+ customer accounts using AI agents to automate utility operations.
👩🏼‍🚒Discover mind-blowing AI tools
Headshot Generator AI - Create polished, professional headshots in minutes with no studio, no photographer, and no hassle.
TurboSite - Build and launch high-performance websites in minutes, without design or engineering friction.
PDF.ai - Chat with your PDFs to instantly extract, summarize, and understand complex documents.
Chopcast - Turn your ideas into a consistent LinkedIn presence, without thinking about posting ever again.
2025 made AI feel inevitable.
2026 will decide who benefits from it.
The advantage won’t go to the loudest voice or the flashiest demo. It will go to the people who build durable habits, ship dependable systems, and make better decisions because they see the landscape clearly.
That’s the lens we bring to OpenTools.
If one idea in this edition sharpened how you’re thinking about 2026, reply and tell us which one. We read every message.
The OpenTools Team
PS: A quiet test for 2026
If AI disappeared from your workflow tomorrow, what would break first?
That answer tells you where you’re already dependent, and where you still have leverage.
Interested in featuring your services with us? Email us at [email protected] |