How AI is quietly creating a new digital economy.
@IrenaCronin and I write this newsletter every week.
AI is creating far more than smarter models. It's giving rise to an entirely new digital economy built on cloud services, infrastructure, AI agents, subscriptions, and
ECONOMY
-
AI Is Shaping a Digital Economy Around Agents and Cloud Infrastructure
By
–
-
Fish Audio Combines Conversational Voice AI Launch with New Funding
By
–
ANYONE CAN CLONE A VOICE. ALMOST NOBODY CAN HOLD A CONVERSATION.
— Nico (@nicos_ai) 28 juillet 2026
fish audio just turned 1, raised $52M and shipped S2.1 Pro
→ 2x faster than Cartesia
→ 1/6th the cost of ElevenLabs
→ 83+ languages
→ word level control over emotion, pacing and delivery
→ on prem with zero… https://t.co/oWAoR2yxGtANYONE CAN CLONE A VOICE. ALMOST NOBODY CAN HOLD A CONVERSATION. fish audio just turned 1, raised $52M and shipped S2.1 Pro → 2x faster than Cartesia
→ 1/6th the cost of ElevenLabs
→ 83+ languages
→ word level control over emotion, pacing and delivery
→ on prem with zero -
Mythos Conducts Cryptography Research with Limited Human Guidance
By
–
Mythos Preview did most of this work autonomously, with occasional human guidance. Each of the two results cost roughly $100,000 in API usage. We disclosed the findings in advance to the algorithms’ authors, as well as to US government and industry partners.
-
RBC Models Higher Costs for Replacing Commercial Software with AI Builds
By
–
Kudos to Rishi Jaluria of RBC Capital for his 37-page report modeling what it costs to replace commercial software with DIY AI stuff. A replacement, AI-built version of Microsoft Office, for example, is hypothetically 11 times as expensive as the license cost, $343,000, a big
-
Cost per Task Better Captures AI Inference Economics Than Token Prices
By
–
incidentally, $ per input/output tokens died as a relevant cost measure sometime last year if you haven’t updated your x axes to $/task per @ArtificialAnlys then idk if you can be taken seriously anymore these days
-
Agentic AI Case Study Cuts Monthly Cost from $1.2M to $100K
By
–
Ben just showed what real agentic AI actually costs.
— Robert Scoble (@Scobleizer) 27 juillet 2026
$1.2 million a month when every action is an agent doing the work.
Then he fixed it. Dropped the bill to about $100,000.
That’s the moment this whole thing gets interesting.
When AI stops just talking and starts executing… https://t.co/wId8nP5KC8Ben just showed what real agentic AI actually costs. $1.2 million a month when every action is an agent doing the work. Then he fixed it. Dropped the bill to about $100,000. That’s the moment this whole thing gets interesting. When AI stops just talking and starts executing
-
AI Infrastructure Stocks Face Pressure as July Trading Weakens
By
–
Trust me when I tell you that the Technology Letter 20 group of stocks to consider is the best barometer of the AI trade. It includes the picks and shovels companies such as Vertiv and Amphenol. The group has been under severe pressure, down 7% for the month of July. It hit a
-
Databricks Unity AI Gateway Aims to Govern Enterprise Token Spending
By
–
AI spend is climbing faster than the value it creates. Join Databricks Co-founder and CTO @matei_zaharia and Product Director @kuhlenhuth to learn how Unity AI Gateway helps organizations move from token maxing to value maxing. See how it already governs more than 1 quadrillion
-
Routing Work Away from Expensive Frontier Models Changes AI Economics
By
–
Really solid breakdown, Ben. The part that stuck with me is how quickly the economics flipped once you stopped treating expensive frontier models as the default for every single action. Most people are still underestimating how fast that shift is coming. Looking forward to seeing
-
Open-Weight Models and Rented GPUs Cut Polsia’s Monthly AI Spending
By
–
Renting GPUs and moving to open weight models took a $1.2M monthly AI bill down to roughly $ 100K.
— Chubby♨️ (@kimmonismus) 27 juillet 2026
That is the core of Ben Cera's new episode. Four months of frontier model spend while Polsia went from 500 users to 5,000 paying customers, then about two months of configuration… https://t.co/WANBJPKGSQRenting GPUs and moving to open weight models took a $1.2M monthly AI bill down to roughly $ 100K. That is the core of Ben Cera's new episode. Four months of frontier model spend while Polsia went from 500 users to 5,000 paying customers, then about two months of configuration