Large law firms are paying to train AI on their own deal and trial files
Kirkland & Ellis agreed this year to pay Palantir $500 million to build technology for its private equity fundraising advice, Business Insider reported. With Harvey saying it has users at 80% of the largest 100 firms, one consultant says the edge has to come from what the models know.
Reality
- Evidence35
- Adoption50
- Hype gap+30
- Incentives70
- Confidence50
Thomson Reuters kept its copyright win over ROSS Intelligence on appeal, the first US appellate ruling on fair use in AI training. ROSS had been refused a Westlaw licence as a rival, so the ruling lets owners of training data set a price or decline to sell.
Perspective Coverage
3 publishers
- Builder
- Builder 19%
- Operator
- Operator 38%
- Investor
- Investor 43%
Reality
- Evidence55
- Adoption30
- Hype gap+25
- Incentives55
- Confidence60
Thomson Reuters moves the Microsoft 365 bill onto Virtual Office CS firms
Thomson Reuters retired Hosted Exchange for Virtual Office CS on Aug. 31, 2026, so firms running Office there now need their own Microsoft 365 licenses. Whether they then move to a managed cloud host depends more on the software they run outside the CS suite than on the new email bill.
Reality
- Evidence35
- Adoption
- Not scored
- Hype gap+40
- Incentives85
- Confidence35
Kentucky's court administrators have confirmed appellate data was taken from the vendor's file systems. The count of affected people, the other states, and the dates of the intrusion all remain undisclosed.
Perspective Coverage
6 publishers
- Builder
- Builder 21%
- Operator
- Operator 58%
- Investor
- Investor 21%
Reality
- Evidence68
- Adoption
- Not scored
- Hype gap+15
- Incentives55
- Confidence62
Astra for Law is GPT-6 Astra with a legal search index and analysis instructions wrapped around it. OpenAI will also sell that index through an API to Harvey and Legora, at a price it has not published.
Reality
- Evidence50
- Adoption30
- Hype gap+20
- Incentives70
- Confidence60
Snowflake's $9.77 billion backlog covers nearly two years of its current run-rate
Product revenue grew 30% to $1.23 billion in the quarter to January 31 while remaining performance obligations grew 42%, and the 12-point gap is the closest thing budget owners have to a forward reading on committed data spend.
Publishers:snowflake.com
Reality
- Evidence57
- Adoption64
- Hype gap+22
- Incentives86
- Confidence56
buildConfirmed3 publishers The legal toolkit ships as its own entry in the ChatGPT model picker and its own API string. Firms weighing it against retrieval stacks they already built still have no price to compare.
Perspective Coverage
3 publishers
- Builder
- Builder 38%
- Operator
- Operator 30%
- Investor
- Investor 32%
Reality
- Evidence55
- Adoption35
- Hype gap+25
- Incentives70
- Confidence62
DeepSeek's first CFO arrives with CITIC Securities engaged and a 2027 listing as the goal. The near-term cash is a 50 billion yuan private round, and the two published accounts of June's valuation differ by nearly seven times.
Reality
- Evidence42
- Adoption33
- Hype gap+38
- Incentives66
- Confidence48
Supio's agent chases a medical-records request for weeks before escalating to the legal team
Supio told SiliconAngle its long-horizon agents own a records request end to end, alerting staff only when it stalls. The company says roughly two-thirds of the work in a case is communication with outside parties.
Reality
- Evidence24
- Adoption14
- Hype gap+44
- Incentives76
- Confidence57
Chinese models clear 75% of enterprise engineering tasks at a fifth of the US cost
American agencies say six Chinese labs bought bulk subscriptions to US models and trained on the outputs since 2024. Enterprise buyers are meanwhile paying a fifth as much for models that clear most of their engineering work.
Reality
- Evidence45
- Adoption55
- Hype gap+25
- Incentives70
- Confidence45
C-Track breach reached at least twelve judiciaries, including Ontario
Court users are getting the letters and the courts are answering the phones, but the supplier that held the files set both the detection date and the disclosure date, which is where the exposure sits.
Reality
- Evidence57
- Adoption74
- Hype gap−12
- Incentives79
- Confidence62
Ord's generation-time argument makes runaway AI unlikely, not just slower
The gate on recursive self-improvement is how long one loop takes, and experiments, training runs and new chips all take real time. That makes capability curves fast but bounded, which changes what you sequence first.
Reality
- Evidence34
- Adoption56
- Hype gap+18
- Incentives58
- Confidence42
Thomson Reuters pins the AI value gap on process redesign over more tooling
A survey of 1,800 professionals finds 91% saying their firm still falls short on AI. The remedy its chief operations officer names is fewer tools with owners attached, plus a date on which a failed pilot dies.
Reality
- Evidence40
- Adoption48
- Hype gap+15
- Incentives68
- Confidence45
Thomson Reuters spent $40 million to own the layer above the open weights
Thomson Reuters puts development of its first proprietary model at roughly $40 million, with under $450,000 of that in the final training run, and says the total still excludes the startup it bought and the cost of keeping the model running.
Reality
- Evidence45
- Adoption20
- Hype gap+20
- Incentives68
- Confidence52
Ramp counts 6.1% of AI-spending businesses paying for platforms that serve Chinese weights
The share buying model-serving platforms went from 4.5% in January to 6.1% in July, and two named buyers have now moved production work onto Chinese open weights while incumbent revenue has not yet moved.
Reality
- Evidence54
- Adoption38
- Hype gap+21
- Incentives66
- Confidence55
buildConfirmed2 publishers Gemini Enterprise for Legal ships as skills, connectors, partner agents and a control plane. One of the systems it plugs into belongs to Thomson Reuters.
Reality
- Evidence52
- Adoption18
- Hype gap+32
- Incentives76
- Confidence58
Legal tech's two leaders are buying the field, and nobody has disclosed a price
More than $7B has gone into legal AI in two years. Harvey is reported to be chasing $15.5B while it and Legora absorb at least eight startups, none at a stated price.
Reality
- Evidence52
- Adoption58
- Hype gap+34
- Incentives62
- Confidence55
Google's legal AI play is a connector list, and Harvey is on it
Gemini Enterprise for Legal plugs into iManage, NetDocuments, RelativityOne and two rival assistants. What Google is actually selling is the governance layer, not the drafting.
Reality
- Evidence34
- Adoption26
- Hype gap+16
- Incentives71
- Confidence41
Thomson Reuters trades Claude for a Qwen derivative it cannot let customers audit
The company says its in-house Thomson-1 matches Claude Opus 4.8 on internal evals, and that the Alibaba base is de-biased. Nobody outside can check either claim.
Reality
- Evidence30
- Adoption30
- Hype gap+38
- Incentives72
- Confidence42
buildOne report1 publisher OpenAI's top model at $4/$20 is a three-month answer to a permanent build decision
The frontier price cut concedes that buyers now price intelligence per task. The discount expires in a quarter; Thomson Reuters' own domain model does not.
Reality
- Evidence42
- Adoption48
- Hype gap+30
- Incentives62
- Confidence44