AI Weekly’s directory lists 28 named AI deployments across research and development, from drug discovery and laboratory biology to semiconductor design and clinical-trial screening. Last updated September 28, 2026, it classifies 17 as “in production or with results,” 17 as having “a reported outcome,” and four as “halted or reversed.” These are directory labels—not independently audited, comparable measures of scientific or commercial success.
What the 28-deployment count represents
The figure of 28 is the number of deployments listed in AI Weekly’s directory, not a count of all AI projects in R&D. The directory groups entries by industry and includes a mixture of operating systems, pilots, reported results, and efforts it says were halted or reversed. Its status totals should be read as the directory’s classifications: it does not provide a common definition of “with results” or “reported outcome,” nor does the supplied summary establish whether those two groups overlap.
As an Amazon Associate I earn from qualifying purchases.
The directory says it was last updated on September 28, 2026. Its counts and descriptions belong to that roundup and do not establish the current status of every project.
What’s actually slowing this PC down?
Pick the symptom - the matching free tool is one click away.
Industries and research tasks represented
The directory’s sector totals add up to 28. The tasks below describe the range of work represented in its summaries; they do not assign a particular task to a particular organization unless noted by the source.
#1 Best Overall
| Directory sector | Entries | Examples of work covered |
|---|---|---|
| Software & Tech | 11 | Internal research agents and model development |
| Pharma & Biotech | 9 | Molecule and drug discovery, including laboratory biology |
| Science & Research | 5 | Scientific hypothesis generation and research workflows |
| Manufacturing | 1 | Semiconductor simulation and design |
| Transportation | 1 | Autonomous-vehicle training data |
| Healthcare | 1 | Clinical-trial screening |
Among the organizations named by the directory are NaiveAI, OpenAI, Anthropic, Hugging Face, Enveda, Novo Nordisk, Anew Labs, Isomorphic Labs, Gamgee, Eli Lilly, Amgen, Moderna, Allen Institute, Thermo Fisher, Google, Fermi Explorer Mission, the U.S. Department of Energy National Laboratories, Intel, Uber, and Cleveland Clinic. These names are included as reported by the directory; the supplied summaries do not independently verify every deployment or map each name to a specific task and outcome.
There is also a count discrepancy in the directory summary as provided: it reports nine Pharma & Biotech entries but names eleven organizations in that group. The available description does not explain the difference, so the organization list should not be treated as a one-name-per-entry count.
How to read the status labels
“In production or with results,” “with a reported outcome,” “pilot,” and “halted or reversed” are not equivalent stages or evidence grades. The directory gives aggregate totals for some labels, but the summary available here does not define a uniform threshold for production, specify the status of every named case, or provide consistent outcome measures. A reported result could therefore mean something very different across research agents, drug discovery, and vehicle data.
Quick wins for a faster PC:
Clear out junk files and repair common Windows errorsFree Scan →Scan for outdated or missing drivers - takes under a minuteDriver Scan →Repair Windows errors before they cause bigger problemsFix Now →- Production indicates a directory classification, not proof that a system has delivered durable value or is still operating today.
- Pilot describes a limited deployment stage, not necessarily an outcome. The directory’s summary does not give a full case-by-case pilot breakdown.
- Reported outcome means an outcome is described in the roundup; it does not by itself show independent validation or comparable measurement.
- Halted or reversed is important evidence about deployment risk and status, but the aggregate count alone does not explain why an effort stopped or what impact it had.
What an AI-assisted R&D workflow can look like
Novartis describes using digital technologies, many powered by AI, across its R&D work. Its examples include identifying promising biological targets, selecting molecules that may have fewer side effects, and pairing generative AI with knowledge graphs to summarize prior studies and real-world evidence for clinical-trial design. These are examples of the company’s stated strategy, not independent performance findings for the 28 deployments.
The distinction matters: helping researchers search, summarize, prioritize, or design a study is not the same as demonstrating that an AI system discovered a successful drug or improved a clinical outcome. Novartis frames the potential benefit as faster decision-making in a resource-intensive process where success is uncertain. That is a rationale for using AI, rather than evidence that a particular deployment has achieved a measured improvement.
Independent reader supportYour contribution helps us test, update, and keep practical guides available for everyone.What the roundup can—and cannot—establish
AI Weekly’s list is useful as a map of reported activity across very different R&D settings. It shows that deployment claims span software research, biotech, scientific investigation, manufacturing, transportation, and healthcare. But the roundup’s headline counts cannot tell readers which systems produced reproducible scientific advances, saved time or money, improved safety, or changed real-world outcomes.
The available directory summary does not independently confirm every case against its original documentation. A peer-reviewed review of AI across drug development offers broader context and company case studies, but it does not verify the current status or details of every directory entry. For a fair comparison, each deployment needs to be checked against its own primary evidence: what task the system performed, whether it was an announcement, pilot, or sustained use, what humans reviewed, who reported the outcome, and how that outcome was measured.
Accordingly, the 28 cases should not be ranked on a single success scale. A research agent that accelerates literature review, a molecule-screening workflow, a semiconductor simulation system, and a clinical-trial screening tool solve different problems and require different evidence. The directory is a starting point for understanding where organizations say they are applying AI—not a common audit of whether those applications work.
Quick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
