Good day, AI enthusiasts!
AI’s competitive edge is increasingly about what sits around the model. Google is tightening the connection between DeepMind and Gemini, Anthropic may spend billions on infrastructure expertise, and Canva is discovering just how expensive successful AI adoption can become.
In today’s digest:
A I__B U S I N E S S
Google’s AI reshuffle puts more power behind Gemini
News Digest: Google is tightening the connection between DeepMind research and its commercial Gemini push. Sergey Brin has reportedly been pressing teams to accelerate development, while Koray Kavukcuoglu has gained greater authority over major decisions — another sign that execution and distribution now matter as much as frontier research.
Key points:
Sergey Brin has reportedly been pushing teams to accelerate Gemini development.
Koray Kavukcuoglu now has greater authority over major DeepMind decisions.
Google is bringing frontier research and commercial product strategy closer together.
Why it matters: Google already has enormous distribution, compute and capital. Faster execution could make Gemini an even stronger competitor across everyday business AI.
Source: Reuters
A I__B U S I N E S S
Anthropic is reportedly in talks to buy Decart AI for around $6bn
Image source: New Digest AI
News Digest: Anthropic is in early discussions to acquire Nvidia-backed Decart AI, according to Reuters. Decart works on inference optimisation alongside real-time video and simulated-environment models. A deal could give Anthropic more control over the increasingly important infrastructure layer behind Claude — although talks may still go nowhere.
Key Points:
The discussions remain early and may not result in a completed acquisition.
Decart combines inference and infrastructure optimisation with generative video and simulation technology.
If completed, Decart’s team would reportedly join Anthropic’s inference and performance organisation.
Why It Matters:
As Claude usage grows, infrastructure efficiency becomes increasingly important. Acquiring Decart could help Anthropic serve more users at lower cost while adding expertise in video, simulation and model performance beyond its core language-model business.
Source: Antropic
A I__+__F I N A N C E
OpenAI-backed Thrive Holdings raises $2bn to bring AI into traditional businesses
Image source: Verdict.co.uk
News Digest: Thrive Holdings has raised $2bn at a $12bn valuation to buy traditional service businesses and rebuild their workflows around AI. Rather than selling another AI subscription, its model is to own the companies where productivity gains happen — and capture more of that value directly.
Key Points:
Thrive raised $2 billion from investors including SoftBank, D1 Capital Partners and Altimeter Capital.
Its strategy combines business acquisition with direct AI implementation rather than simply selling software.
The company plans to expand into regulatory services linked to areas such as data centres, manufacturing, power, water and transport.
Why It Matters:
Thrive is betting that the biggest AI gains will come from redesigning how businesses operate — and then owning the companies that benefit. It represents a shift from selling AI tools to capturing the productivity gains directly.
Source: TechCrunch
A I__+__B U S I N E S S
Canva slows its AI rollout as inference costs bite
Image source: medium.com
News Digest: Canva has slowed parts of its AI rollout after heavier-than-expected usage pushed up inference costs. It says the cost per AI task has already fallen by nearly 90%, but demand keeps climbing. It’s a useful reminder that adding successful generative AI to SaaS can improve the product while making its economics harder.
Key Points:
Canva reduced its expected revenue-growth rate to around 20%.
Users of its newer AI experience are reportedly creating around three times as many designs.
Canva says the cost per AI task has fallen by nearly 90%, but overall usage continues to rise rapidly.
Why It Matters:
Traditional SaaS benefits from very low marginal costs. Generative AI changes that equation because every AI action consumes compute. Canva shows how successful AI adoption can simultaneously strengthen a product and put pressure on its margins.
Source: Fortune
U S E__T H I S__T O D A Y
Build a reusable company context file for AI
Find the customers most likely to buy again
Export your last 3–6 months of customer or order data and ask AI to spot repeat-purchase patterns, high-value customers and people who may be due for a follow-up. Use the results to build a simple re-engagement list.
Try this prompt:
Analyse this customer/order data and identify the best opportunities for repeat sales. Rank customers by likely value and explain why each group is worth targeting. Look for purchase frequency, average order value, time since last purchase and any useful patterns. Then give me three customer segments to contact first, the offer or message most likely to work for each, and any risks or assumptions in your analysis. Keep the recommendations realistic for a small business with limited time and budget.
Best for: Founders, marketers and sales teams.
O T H E R__N E W S
Everything Else in AI
Nvidia is working with major financial groups on structures designed to mobilise more than $500bn for AI infrastructure, potentially turning GPUs and compute capacity into something closer to a conventional infrastructure asset class.
AI-driven job destruction still isn’t showing up as a broad employment collapse. The more immediate shift appears to be higher productivity expectations, slower hiring for automatable work and growing demand for practical AI skills.
Google’s Pixel Watch 5 is bringing more Gemini capability onto the wrist, including some offline tasks, proactive suggestions and new AI-assisted health and fitness features.
Foxconn now generates 51% of quarterly revenue from cloud and networking, overtaking the consumer-electronics division associated with iPhone manufacturing as AI-server demand reshapes the company’s business.
A I__T O O L S
🛠️ Top AI Tools & App Releases
A few recent products and updates worth a look:
💻 Prelint — checks AI-written pull requests against product specs, business rules and previous decisions.
📣 Denovo — AI business builder spanning websites, branding, campaigns, outreach and customer acquisition.
🔎 Prefactor — monitors production AI agents for quality, drift, risk, cost and risky actions.
That's it for today!
See you soon,
News Digest AI team