Generative AI

Bible
The ultimate guide
to genAI disruption

Pablo Xavier via BuzzFeed News

Unleash the
generative AI
revolution
Dive into a treasure trove of generative AI
companies, investors, and market insights.
Elevate your decisions with exclusive software
buyer interviews to understand product pricing,
CSAT, and post-deployment experience.

CB Insights – Your Gateway to Generative AI

Sign up for access

2
Generative AI is a theme CB Insights covered before the
hype. We've consistently been ahead of the curve

2019 2020 2020 2021 2022

1 2 3 4 5

Generative Commercial Synthetic data, Transformers, Code generation,

adversarial deepfakes, AI self-supervised multilingual deepfake
networks voice spoofing learning models detection,
multimodal AI
CB Insights CB Insights CB Insights CB Insights CB Insights
2019 Trends Report 2020 Trends Report 2020 Trends Report 2021 Trends Report 2022 Trends Report

3
 And we're staying ahead today – so you can too

2023
Understand 25+ generative AI markets Stay on top of the landscape with research & transcripts
Large language model developers
Customer support operations
Text generation
Protein & drug design
Voice synthesis & cloning
+ more

Discover 400+ genAI vendors

Tracked in our analyst-curated
Expert Collection

CB Insights helps the world’s leading companies understand everything they need to know about 4
disruptive technologies — find out more about why our customers love us here.
Contents
Generative AI Bible

The generative AI boom 6 So where is generative AI headed? 63

Gradually 9
1. Race to dominate genAI infrastructure 64
Then suddenly 19
2. Cross-industry applications face pressure 88
And now suddenly is accelerating 37 from large players

Hundreds of startups pile into genAI 38 3. Opportunity in vertical genAI 93

Healthcare & life sciences
Funding soars as investors flock 41
Financial services & insurance
Big tech is all in and ready to fight 50
Retail

Promising companies to watch 119

5
The
generative AI
boom

6
 ERNEST HEMINGWAY, THE SUN ALSO RISES

“How did you go bankrupt?”

7
 ERNEST HEMINGWAY, THE SUN ALSO RISES

“Two ways…
Gradually and then suddenly”

8
2014

Gradually
Generative AI has been in the works for years

9
 How did we get here? A recent timeline of select
events in the development of generative AI
2014 2016 2017 2018 2019

1 2 3 4 5

Generative adversarial WaveNet and audio New neural network Google AI releases BERT, OpenAI releases GPT-2,
networks (GANs) generation introduced architecture called the a leap in the ability of gaining attention for
introduced by Ian by DeepMind “Transformer” introduced machines to understand text generation
Goodfellow by Google researchers context in language capabilities

2020 2020 2021 2022 2022

6 7 8 9 10

OpenAI releases GPT-3, “Deepfakes” become OpenAI releases Text-to-image models OpenAI launches GPT-
accelerating interest in widely known text-to-image model from Google, Midjourney, 3.5-based chatbot
language models DALL-E Stability AI, and OpenAI ChatGPT, unleashing
proliferate genAI boom

*Generative AI is artificial intelligence that can generate new content (text, code, images, audio, etc.). 10
 GANs tap into the idea of “AI versus AI” — advancing
image generation dramatically
1
2014 2016 2018 2020 2022

Images from 2018 paper where DeepMind researchers trained GANs on a large-
scale dataset to create “BigGANs”

Image source: Large Scale GAN Training for High Fidelity Natural Image Synthesis 11
*”AI versus AI”: A breakthrough where two neural networks try to outsmart each other, creating and refining
synthetic outputs.
 WaveNet produces synthetic audio, showcasing the
potential of generative models beyond images
2
2014 2016 2018 2020 2022

Image source: Google DeepMind 12

 The Transformer architecture can better understand and
generate human language, paving the way for further R&D
6 authors of the seminal research paper have gone on to raise $1.7B across 5 AI companies*

3
2014 2016 2018 2020 2022

175 +661%

Source: CB Insights — 6 authors of a seminal research paper by Google 13

*As of 10/12/2023
 Google AI releases BERT, a leap in the ability of machines
to understand context in language
4
2014 2016 2018 2020 2022

The AI language model predicts a word based on not only the preceding words, but
also the succeeding ones (bidirectional understanding of context).

BERT is deeply bidirectional, OpenAI GPT is unidirectional, and ELMo is shallowly bidirectional.

Image source: Google 14

 OpenAI incorporates the Transformer architecture into its
language models
5
2014 2016 2018 2020 2022

GPT-1 – June 2018 GPT-2 – February 2019

Image source: OpenAI 15

 OpenAI’s GPT-3 is a major leap forward, showcasing ability
to generate code, jokes, and more
6
2014 2016 2018 2020 2022

November 2020
July 2020

Image source: NYT, MIT Technology Review 16

 Deepfakes go mainstream, highlighting power and pitfalls
of video generation
7
2014 2016 2018 2020 2022

Image source: YouTube, TikTok via ABC 17

*Deepfakes refer to synthetic media where a person in an existing image or video is replaced with someone
else’s likeness using neural networks.
 GPT-3 is the foundation for DALL-E, which can generate
images from text descriptions
8
2014 2016 2018 2020 2022

Image source: OpenAI 18

2022

Then suddenly
GenAI goes from experiment to everywhere

19
 Models get bigger…

Image source: The Economist 20

 …and better, beating human performance benchmarks

Image source: Science 21

 Text-to-image generators take the internet by storm

Image source: Imagen (Google), Midjourney, DALL-E 1 vs. DALL-E 2 (OpenAI), Stable Diffusion 22
 Generative AI startups raise major funding to fuel growth
Deals worth $100M+ to generative AI startups in 2022

Source: CB Insights – Advanced Search - Deals 23

 AI coding assistant GitHub Copilot
becomes widely available

Image source: GitHub 24

 ChatGPT goes viral, getting to 1M users in 5 days and
100M in 2 months — unleashing genAI boom
Time to 1M users for select platforms/apps from launch

ChatGPT 5 days

Instagram 2.5 months*

Spotify 5 months

Dropbox 7 months

Facebook 10 months

Twitter 2 years
0 5 10 15 20 25 30

Source: Media mentions 25

*App downloads
 Almost overnight, exec interest in generative AI
skyrockets and companies feel pressured to react
Earnings call mentions of “generative AI” (as of 9/30/2023)
2,500

2,000
2,081

1,546
1,500

1,000

446
500

1 0 1 0 0 0 0 28
0
Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q2 Q3
2021 2022 2023

Source: CB Insights — Advanced Search - Earnings transcripts 26

 Microsoft invests $10B into ChatGPT-maker OpenAI in
Q1’23, propelling genAI funding to new heights
Disclosed equity funding & deals to generative AI companies (as of 9/30/2023)
Funding funding Deals
Funding
$17.4B
250
$16.
16000
0B

$14.
14000
0B Deals
168 170 200
$12.
12000
0B
154
$10.
10000
0B 150

$8.0B
8000 103
$6.0B
6000
64
$5.3B $10.0B 100

$4.0B
4000 $3.2B
50
$1.7B
$2.0B
2000
$0.7B
$1.0B $2.0B
$0.0B
0 0

2019 2020 2021 2022 2023 YTD

Source: CB Insights — What are customers saying about generative AI startups? 27

 Microsoft debuts AI-powered Bing, running on a new
OpenAI large language model (LLM)

Image source: Microsoft *A large language model is a deep learning algorithm that analyzes and produces text 28
by learning from extensive language data.
 Meta introduces Llama, an open-source language model

Image source: Meta 29

 Google sounds alarm bells and releases Bard chatbot

On the AI side, it is a really exciting time. I think we've been

investing for a while, and it's clear that the market is
ready…Obviously, we need to make sure we're iterating in
public, these models will keep getting better, so the field
is fast changing. The serving costs will need to be
improved.

So I view it as very, very early days, but we are committed

to…actually bringing direct LLM experiences in Search,
making APIs available for developers and enterprises and
learn from there and iterate like we've always done.
Alphabet CEO Sundar Pichai
Q1’23 earnings call

Source: CB Insights – Advanced Search - Earnings transcripts; Google 30

*Reflects quarter call occurred
 GPT-4 becomes OpenAI’s most powerful model yet,
crushing human exams

Image source: OpenAI 31

 Reddit, after providing years of free training data for AI
systems, plans to charge for access to its content

Image source: Reddit 32

 Chegg blames ChatGPT for declining revenue, sees its
share price tank, and, in response, pivots to build LLMs

Chegg attributes
declining growth
to ChatGPT

Source: CB Insights – Chegg company profile, Chegg 33

 Nvidia enters the $1T market cap club as demand
for GPUs used in genAI sends its revenue soaring
US companies to reach $1T+ market cap (as of 10/30/2023)

$2.7T
$3.1T
$2.5T
$2.7T
$1.6T
$2.0T
$1.4T
$1.9T
$1.0T
$1.2T
$0.8T
$1.1T
Current market cap
$0.6T
All-time high
$1.2T
$0.0T $0.5T $1.0T $1.5T $2.0T $2.5T $3.0T $3.5T

*GPUs = graphics processing units, which are used to run intensive AI applications. 34
 Stack Overflow sees declining traffic and lays off
employees amid AI coding boom

Image source: SimilarWeb, The Verge 35

 Even Apple scrambles as it works on its own LLM called
Ajax and is on course to spend $1B a year on genAI push

If you take a step back, we view AI and machine learning as

core fundamental technologies that are integral to virtually
every product that we build…And of course, we've been doing
research across a wide range of AI technologies, including
generative AI, for years. We're going to continue investing
and innovating and responsibly advancing our products with
these technologies with the goal of enriching people's
lives…And as you know, we tend to announce things as they
come to market, and that's our MO, and I'd like to stick to that.
Apple CEO Tim Cook
Q3’23 earnings call

Source: Bloomberg; CB Insights – Advanced Search - Earnings transcripts 36

*Reflects quarter call occurred
And now suddenly
is accelerating
Ambitious & flush with cash, young
companies and big tech players are all
rushing into this next platform shift

37
Hundreds of startups
pile into genAI

38
 Commercial genAI
applications are
proliferating
Generative AI Market Map

Explore the full map

Source: CB Insights — Generative AI Market Map 39

GENAI LANDSCAPE LAYERS

300+ vendors have

emerged across:

• Cross-industry generative applications

(visual media, text generation, code generation,
etc.)
• Industry-specific generative applications
(healthcare, finance, etc.)

• Generative AI infrastructure
(foundational models, vector databases, etc.)

Source: CB Insights — Generative AI Market Map 40

Funding soars as
investors flock

41
 As investors look to ride the generative AI wave,
funding soars in 2023
Disclosed equity funding & deals (as of 9/30/2023)
Deals
$20. 0B 168 170 180

$18. 0B
154
160

$16. 0B
Funding 140
$14. 0B
103 $17.4B 120
$12. 0B
100
$10. 0B

$8.0B
64 80

60
$6.0B $5.3B
$4.0B $3.2B 40

$1.7B 20
$2.0B
$0.7B
$0.0B 0

2019 2020 2021 2022 2023 YTD

Source: CB Insights — The state of generative AI in 7 charts; Deals Story 42

 Generative AI infrastructure attracts the bulk of funding,
due to the capital-intensive nature of developing LLMs
Disclosed equity funding & deals to generative AI categories, from Q4’22 to Q3’23
Funding Deals

$11.6B
Generative AI
infrastructure
30 deals

$5.5B
Cross-industry
generative applications
129 deals

$0.8B
Industry-specific
generative applications
48 deals

Source: CB Insights — The state of generative AI in 7 charts 43

 A record number of $100M+ mega-rounds drive funding
surge, with money primarily going to infrastructure layer
Disclosed $100M+ genAI equity deals (as of 9/30/2023)

1 25

0. 9
Mega-rounds
0. 8
20 20

0. 7

0. 6 15

0. 5

9
0. 4 8 10

0. 3

0. 2
4 5

0. 1 1
0 0

2019 2020 2021 2022 2023 YTD

Source: CB Insights — Advanced Search - Deals 44

 Alongside big deals, generative AI is minting unicorns left
and right
New unicorns ($1B+ valuation), Q1’23 — Q3’23
Company Valuation Country

1 $4.4B United States

2 $2.2B Canada

3 $1.5B United States

4 $1.4B Israel

5 $1.2B United States

6 $1.0B United States

6 $1.0B United States

6 $1.0B United Kingdom

6 $1.0B United States

6 $1.0B United States

6 $1.0B China

Source: CB Insights 45
 Out of the 16 new AI unicorns in 2023 so far, 11 are
genAI companies
New AI unicorns ($1B+ valuation)
30 AI unicorns 2023 genAI unicorns

25

20
24
21
15

16
10
14 14 14
7
4 5 4
5
3 3 3 2 6 5
0
3 3
Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q2 Q3
2020 2021 2022 2023

Source: CB Insights — State of AI Q3'23 Report 46

 Biggest M&A deal of 2023 lands in infrastructure, but the
rest of M&A reflects push at the industry/application level
Generative AI M&A exits, Q1’23 — Q3’23

Company Round Valuation Acquirer GenAI Focus Area Country

1 MosaicML $1.3B Databricks Infrastructure United States

2 InstaDeep $125M BioNTech Healthcare & life sciences United Kingdom

3 Casetext $650M Thomson Reuters Legal United States

4 Light Year AI $234M Meituan Infrastructure China

5 Valence $47M Recursion Healthcare & life sciences Canada

Neeva N/A Snowflake Enterprise tech United States

Thankful N/A Gladly Enterprise tech United States

Fig N/A Amazon Web Services Enterprise tech United States

Codiga N/A Datadog Enterprise tech United States

Source: CB Insights — Advanced Search - Deals 47

 The US is poised to own the genAI boom: 2x more deals in
the US than the rest of the world combined
Disclosed generative AI equity deals to startups by company HQ, Q4’22 — Q3’23

United States 143

Rest of world 64

0 20 40 60 80 100 120 140 160

Source: CB Insights — Advanced Search - Deals 48

 But it's still early days for genAI startups — 71% are
early-stage or haven't raised any funding
Percent of companies by latest disclosed round (as of 9/30/2023)

Seed, pre-seed, angel, &

convertible note
16% Series A

39% Series B & C

19%
Series D+
17 +66%

12% Not raised outside funding

1% 13%

Other

Source: CB Insights — The state of generative AI in 7 charts 49

*Other includes non-equity funding rounds and equity rounds not tied to specific stage of investment.
Big tech is all in and
ready to fight

50
 Generative AI is a new battleground for big tech, with
overlapping alliances and commitments into the billions
Generative AI companies with two or more big tech investors (as of 9/30/2023)
Big tech investors

Indicates where big tech invested

Hugging Face
Adept
AI21 Labs
Anthropic
Company

Inflection AI
Inworld AI
OpenAI **

Runway
Synthesia 1,646
Typeface +370%

Source: CB Insights *Includes investments from M12 and Google Ventures 51

**AWS
 Big tech backs every top deal in 2023 so far, with other
CVCs & corporates joining the action
Top generative AI equity deals, Q1’23 — Q3’23
Round Round
Company Round Amount Big Tech Investors Other Select Investors Country
Date Valuation
Corporate Minority - III
1 OpenAI $10.0B 2023-01-23
N/A Microsoft United States

Series B
2 Inflection AI $1.3B 2023-06-29
$4.0B Microsoft, Nvidia Gates Frontier United States

Corporate Minority - V
3 Anthropic $1.25B 2023-09-25
N/A Amazon United States

Series C Menlo Ventures, SK telecom ventures, Salesforce

4 Anthropic $450M 2023-05-23
$4.1B Google
Ventures, Zoom Ventures
United States

Corporate Minority
5 Anthropic $400M 2023-02-03
$4.1B Google United States

Series B General Catalyst, Spark Capital, Atlassian Ventures,

6 Adept $350M 2023-03-14
$1.0B Microsoft, Nvidia
Workday Ventures, Greylock Partners
United States

Generate Series C 175Fidelity

Abu Dhabi Investment Authority, Amgen, +661%
7 $273M 2023-09-14
N/A NVentures
Investments, Flagship Pioneering, MAPS Capital
United States
Biomedicines
Series B Inovia Capital, Index Ventures, Oracle, Salesforce
8 Cohere $270M 1,646
2023-05-02
$2.2B Nvidia
Ventures, SentinelOne
Canada

Series+370%
D Amazon, Google Ventures, Salesforce Ventures, AMD, IBM Ventures, Intel
9 Hugging Face $235M 2023-08-23
$4.5B
NVentures Capital, Qualcomm Ventures
France

Series B
10 Imbue $200M 2023-09-07
$1.0B Nvidia Astera Institute United States

Source: CB Insights — Advanced Search - Deals 52

*Big tech includes Amazon, Apple, Microsoft, Meta, Google, and Nvidia
 Nvidia increases its investment activity dramatically...
Equity deals backed by Nvidia (as of 10/25/2023)

1 Deals 18

0. 9 17 16

0. 8
14
0. 7
12
0. 6
9 10
0. 5
7 8
0. 4

6
0. 3
3 4
0. 2
2
0. 1 2

0 0

2019 2020 2021 2022 2023 YTD

Source: CB Insights — Nvidia company profile – investments 53

 …becoming the most active investor in generative AI
Top generative AI investors by company count, Q1’23 — Q3’23

Investor Company Count Investor Group Country

1 Nvidia 9 Corp United States

2 SV Angel 7 Angel United States

3 Salesforce Ventures 6 CVC United States

3 Index Ventures 6 VC United States

3 Andreessen Horowitz 6 VC United States

6 GV (Google Ventures) 5 CVC United States

7 Microsoft 4 Corp United States

7 Sequoia Capital 4 VC United States

7 Lightspeed Venture Partners 4 VC United States

Source: CB Insights — Advanced Search - Deals 54

 The dominant chipmaker is cashing in on the genAI
computing boom, backing startups using its chips…
Nvidia-backed equity deals to generative AI companies (as of 9/30/2023)

Source: Inflection AI; CB Insights — Advanced Search - Deals 55

 …as demand for Nvidia GPUs far exceeds supply

If you're trying to do the training There is [a] significant bottleneck in terms

of Nvidia’s GPU chips…in short, the
of the models, then having the demand far exceeds the supply in the
absolute latest, greatest GPU, market. And that is not the situation for us
the H100 from Nvidia right now, as one company, it is a general situation for
there's a lot of constraint in the industry…We had originally expected
the revenue from AI to be shown in our
getting those chips. financials in the third quarter and that is
likely to be delayed due to the shortage of
supply of GPU servers through the fourth
quarter or even the first quarter of 2024.

CEO Matthew Prince, CEO Tao Zou

Q3’23 earnings call Q3’23 earnings call

Source: CB Insights — Advanced Search – Earnings transcripts 56

*Reflects quarter call occurred
 Microsoft steps up its genAI investment activity beyond
its $13B invested in OpenAI…
Microsoft-backed equity deals to generative AI companies (as of 9/30/2023)

1,646
+370%

Source: CB Insights — Advanced Search - Deals *Includes deals backed by Microsoft’s venture arm M12 57
 ...betting that generative AI could tip the competitive
scales in its favor for decades to come

Microsoft
Investment Thesis
Map – Generative AI

Source: CB Insights — Analyzing Microsoft’s generative AI strategy: How Microsoft is expanding past OpenAI 58
to transform the way we work
 Microsoft’s investments in genAI help reverse slowing
Azure growth and contribute to 3 percentage point bump
Azure and other cloud services revenue growth (year-over-year) by quarter
70%

60% 62%
59% 59%

50%
50% 50% 51% 50%
47% 48%
46% 46%
40%
40%

30%
35% 29%
31%
27% 26%
20%

10%

0%
Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 Q1
FY 2020 FY 2021 FY 2022 FY 2023 FY 2024

Source: Microsoft 59
 Alongside its Bard chatbot, Google puts billions toward
internal AI research & a range of AI startups
Google-backed equity deals to generative AI startups (as of 10/27/2023)

1,646
+370%

Source: CB Insights — Advanced Search - Deals *Includes investments by Google, Google Ventures, and Gradient 60
Ventures
 Amazon and Google commit billions to LLM developer
Anthropic in battle with OpenAI-backer Microsoft

175 +661%

1,646
+370%

Source: CB Insights — Anthropic company profile - funding 61

 Share of genAI
accelerator cohort

Amazon launches
$100M generative AI
accelerator, looking
to feed its cloud
computing business
AWS Generative AI
Accelerator Investment
Thesis Map

Source: CB Insights — Where the AWS Generative AI accelerator is placing its bets across 7 industries 62
So where is
genAI headed?
Generative AI will soon be
impossible to ignore as disruptive
applications spread

1. Race to dominate genAI infrastructure

2. Cross-industry applications face pressure from
large players
3. Opportunity in vertical genAI

63
1. Race to dominate genAI infrastructure

No clear winner yet in

foundational models

64
 Highest-valued genAI unicorns compete primarily at the
infrastructure layer
Most highly valued private generative AI companies (as of 09/30/2023)

Company 1 $29.0B
Company 2 $7.3B
Company 3 $4.5B
Company 4 $4.1B
Company 5 $4.0B
Company 6 $2.1B
Company 7 $1.8B
Company 8 $1.5B
Company 9 $1.5B
Company 10 $1.4B
$0M $5,000M $10,000M $15,000M $20,000M $25,000M $30,000M

Source: CB Insights — The state of generative AI in 7 charts 65

 4 LLM developers – Anthropic, Cohere, AI21 Labs, and
Adept – join the unicorn club in 2023
Leading LLM developers by valuation (as of 09/30/2023)

Source: CB Insights — Generative AI — large language model developers market report 66

 While OpenAI has clear lead, vendors are competing on
multiple fronts to become the go-to model developer

Key KPIs for evaluation

Safety & compliance
Accuracy & quality
Customization
Pricing & deployment
1,646
Token limits
+370%

Source: CB Insights — Generative AI — large language model developers market report - scorecard 67
 Executive interest in AI has surged

Enterprise demand for AI and accelerated computing is strong.

We are seeing momentum in verticals such as automotive,
financial services, healthcare, and telecom, where AI and
accelerated computing are quickly becoming integral to
customers' innovation roadmaps and
competitive positioning.
CFO Colette Kress
Q2’23 earnings call

Source: CB Insights — Advanced Search - Earnings transcripts 68

*Reflects quarter call occurred
 Enterprises are spending millions with LLM developers…
Annual spend ranges displayed

$100K - $500K

$20K - $2M

$25K - $300K

$50K - $100K

$15K - $800K

$40K - $5M

Source: CB Insights — Software buyer interview transcripts and Analyst Briefing data 69
 …but reducing costs & time to train are key priorities

Mosaic ML offers what's called One of the things that we're really
programmatic optimization, which is not trying to do is reduce the cost for a lot
so much on the hardware side of things, of our large language models and
but rather on the algorithmic side. Can training…What we liked about Mosaic
you find ways of optimizing the time it was it is a lot less expensive in terms
takes to get to a certain performance of the training models…Right now, I
bar? I think that's really what drove us to would project, just based on our usage,
evaluate MosaicML. In fact, it was pretty I think the initial spend was $15,000 per
much the main tool out there right now annum. I would expect next year to
that offers this. I don't think there's any probably be $200,000, $250,000. We're
other tool that really has this a very large organization and there's
programmatic optimization layer. been a lot of interest in MosaicML.

Senior Manager, Data Science, Vice President, Innovation,

$1B+ valuation technology company Fortune Global 500 company

Source: CB Insights — MosaicML software buyer interview transcripts 70

 Safety & compliance will be in focus for enterprise
customers

Source: CB Insights — Cohere software buyer interview transcript; Anthropic software buyer interview 71
transcript
 Concern around AI
risks puts responsible
AI solutions in the
spotlight
These tools help
enterprises build and
deploy AI in an ethical 175 +661%

and legal manner

Source: CB Insights — The responsible AI market map 72

 As LLM developers burn through cash, focus will shift to
customer adoption — and revenue
Latest disclosed or whisper revenue (as of 10/18/2023)

$1,300M

$200M $50M $40M $20M $10M

Company A Company B Company C Company D Company E Company F

Source: CB Insights — Figures represent the latest disclosed revenue (based on company discussion or media sources).
73
All revenues are full-year 2023 ARR projections, expect for MosaicML (reported ARR at time of acquisition in June 2023).
 Companies need to grow into big valuations and will come
under pressure to build real business models
Revenue multiples (as of 10/18/2023)

113x
100x

65x

28x
22x 21x
Company A Company B Company C Company D Company E Company F

Source: CB Insights — Figures represent the latest disclosed valuation divided by revenue (based on company discussion or
74
media sources). All revenues are full-year 2023 ARR projections, except for MosaicML (reported ARR at time of acquisition
in June 2023).
 There’s no winner yet in foundational models

Strengths, I would say, the ethical We were considering obviously OpenAI, The first two things that came to my
considerations of privacy and bias, Cohere, Anthropic, and deepset…To be mind is, of course, OpenAI is still the
fairness…their model outperformed the honest, the reason that we actually market leader. That gives us the
other models, including GPT-3 and chose AI21 is because the interface is sense of comfort and we are confident
ChatGPT…In terms of weaknesses, the super easy to use for non-tech on this market-leading product…On the
specificity of the model output and the people…I actually tested Cohere pretty flip side of the open-source platforms
interestingness of the model output… I extensively…I think that their models that I just mentioned, the Hugging
think that other weakness also was in are really strong and also for some of Face and the Llama 2, we don't have
terms of speed and efficiency, like the creative work, I thought they were much faith and information about how
latency, and once you ask a question, doing slightly even better than even they are going to deal with our data.
how long does it take to fully respond. OpenAI when it comes to creative stuff, That's the key to the enterprise world.
like ad copy and marketing related If we are not certain, then I would
tasks. rather pass my roles to OpenAI.

Senior Manager of Data Science, Head of AI, Cloud, Data & AI Lead,
Model-as-a-service platform $100M+ funded technology startup Fortune 500 company

Source: CB Insights — Anthropic, AI21 Labs, OpenAI software buyer transcripts

75
1. Race to dominate genAI infrastructure

Open-source AI
movement gains steam

76
 Need for AI model transparency and rapid innovation
is fueling the open-source AI movement
News mentions of open-source AI and related terms (as of 9/30/2023)

175 +661%

1,646
+370%

Source: CB Insights — Advanced Search - News mentions 77

*The open-source approach to AI development is focused on making source code available for public use and allowing a
community of developers to contribute to improving software.
 Meta is leveling the playing field with its open-source
LLM, Llama 2

Notably, we recently announced a Azure AI is ushering in new AI is the flavor of the day. And
collaboration with Meta on Llama born-in-the-cloud, AI-first thanks to ChatGPT’s great
2-based AI implementations on workloads with the best selection launch, everyone has discovered
flagship smartphones and PCs of frontier and open models, this potential. We can expect
that will enable developers to including Meta's recent new changes by the day in the
create new and exciting genAI announcements supporting tech world. Just take yesterday,
applications using the AI Llama on Azure and Windows, Meta’s launch of Llama 2, which
capabilities of Snapdragon as well as OpenAI. will be available on Azure free of
platforms beginning in 2024. charge, including for
commercial 175 purposes.
+661%

Qualcomm CEO Cristiano Amon Microsoft CEO Satya Nadella Publicis CEO Arthur Sadoun
Q3’23 earnings call Q3’23 earnings call Q3’23 earnings call

Source: CB Insights — Advanced Search - Earnings transcripts 78

*Reflects quarter call occurred
 The private market is split into open vs. closed
Disclosed equity funding to LLM developers (as of 10/27/2023)

Closed-source LLMs Open-source LLMs

*Some developers may offer open-source versions of their models *Excludes open-source developers that have not raised equity funding
but keep their core models proprietary

Source: CB Insights — Generative AI – Large language model developers market report 79

 OpenAI customers highlight potential cost-savings,
customizability benefits of open-source models – though
OpenAI has the edge on performance & support

OpenAI’s developer API and the developer Meta's invention… it may be cheaper than
experience is definitely the best. It's really the OpenAI [model] because it's open-
managed, super clean APIs, well- source. Then we believe the performance
documented, and has the most of the Hugging Face and also the Llama 2
integrations. There was a big push toward is also comparable to the OpenAI [model].
using open-source models and then fine- Maybe just a little bit weaker than that, but
tuning them with your own data. That's maybe the overall…ROI is quite a good
still a big thing. We're actually evaluating deal.
that for some of the public data set stuff
because it's a lot cheaper, especially if you
add in more huge amounts of data versus
a giant OpenAI model.
Partner, Early-stage VC firm VP, Machine Learning, Fortune 500 company

Source: CB Insights — OpenAI software buyer interview transcripts 80

 Databricks pays $1.3B for MosaicML, which makes AI
development tools and has its own open-source model,
in June 2023

$20M in FY’22
revenue (65x
multiple)

Source: CB Insights — Databricks acquired Mosaic ML for $1.3B. How do the valuations of other generative AI 81
companies compare?
 Growing number of
vendors are developing
open-source tools to help
enterprises build and
deploy AI projects
Open-source AI
development market map 175 +661%

Source: CB Insights — The open-source AI development market map 82

1. Race to dominate genAI infrastructure

LLM infrastructure market

grows rapidly

83
 Tech vendors
supporting LLM
operations are
gaining traction
with enterprises
LLMOps market map
175 +661%

Source: CB Insights — The large language model operations (LLMOps) market map 84
*LLMOps refers to the end-to-end workflow that organizations employ to build, fine-tune, and deploy LLMs into production.
 Execs are buying infrastructure tools for better training
data, observing performance of models, and more

We had, basically, messy data and we I led a small data science team that
needed a better way of providing, of created a lot of models in production. As
doing better training data for generative we scaled, our observability of our
content... We have a lot of machine machine learning models in production
learning models that are fueling us here. was limited, and we felt blind to issues
In both cases, it was the desire to have, or ways to improve the model once in
basically, a higher level of data hygiene production. We tried to build a solution
in our training data. in-house, which showed the difficulty of
the challenge.

Chief Product Officer, IT company Senior Manager, $10M+ funded data analytics
platform

Source: CB Insights — Snorkel AI software buyer interview transcript, Fiddler AI software buyer interview 85
transcript
 New vendors are emerging for LLM fine-tuning &
customization
Leading LLM application development vendors by disclosed equity funding (as of 10/30/2023)

Source: CB Insights – LLM application development market report 86

 Vector database startups, which make data more accessible
for AI systems, raise record funding amid LLM boom
Disclosed equity funding and deals (as of 9/30/2023)

$200M 7 Funding
$180M
$176M 7

$160M 6

$140M
5
$120M
Deals
4
$100M
$109M 5
$80M 3

$60M
2
2
$40M 1
$44M 1
$20M
0 $10M
$0M 0

2019 2020 2021 2022 2023 YTD

Source: CB Insights – Vector database market report 87

*Vector databases provide enterprises with an easy way to store, search, and index unstructured data.
2. Cross-industry applications face
pressure from large players

88
 Execs are demanding their tech vendors keep up with
genAI advances and opportunities

From a technical perspective, I think SlashNext is working on a generative AI-based

this wave of ChatGPT and OpenAI large solution, which means it will generate its own
kind of phishing and malware and it will train
language models is going to open up a
the software to automatically be aware of any
lot of opportunities for Sourcegraph new kind of threats arriving in the market. So
because they already have a lot of the even if tomorrow some human is creating a
code and can say, "We'll take a new kind of malware or any other software is
customized model, throw in your code, creating some new kind of phishing or
and give you super good suggestions for ransomware, because SlashNext is based on
AI, it is already aware of these kinds of
your developers." So I think that's an changes and it will be able to detect them
area that is super interesting. And again, before any other software can do so. That is a
they have to compete with GitHub, differentiator from the technology perspective.
which already has Copilot.

VP, Technology at Publicly traded e-commerce Senior Design Engineer at Fortune 500
company company

Source: CB Insights — Sourcegraph software buyer interview transcript; SlashNext software buyer interview 89
transcript
 Generative interfaces, like Anthropic’s AI assistant Claude,
lead in funding among cross-industry tools
Distribution of generative AI funding, Q3’22 — Q2’23

175 +661%

1,646
+370%

*Based on an analysis of 210+ generative AI companies

Source: CB Insights — The state of generative AI in 7 charts building cross-industry solutions; excludes deals to 90
industry-specific companies and model developers such as
OpenAI.
 Growing competition is a threat to vendors in some cross-
industry markets, like text generation & editing

Mutiny and Jasper announce layoffs

So in this newly emerging world of generative AI it's

hard to keep up with all the changes that are going on.
It's probably not fair to ask this, but I'll say it: Jasper
needs to stay up to date faster to make me a definite
yes to renew. We need to look at their pricing model to
make sure, as I'm beginning to use it more and more at
a higher and higher scale, that it keeps working and the
price remains right for me. I'm seeing other lower cost
options; the price of calls to GPT-3, for example, has
gone down to really minimal numbers. It might be
harder to say yes to a renewal when we're due next
year, so I'll have to really see that our people have
picked this up and are finding great value.

C-level executive at $10M+ funded research platform

Source: CB Insights — Mutiny company profile - headcount; Jasper software buyer interview transcript 91
 Watch for vendors to scramble to build defensible
moats in specialized areas

Source: CB Insights — Generative AI — legal case search & summarization; Virtual medical scribes & 92
summarization tools
 3. Opportunity in vertical genAI
How generative AI is going to be used to…
Drive growth Improve customer experience Reduce costs & risk
Healthcare & • Copilots for doctors automate • AI companions address well-being & • GenAI drug discovery & design reduces
life sciences tedious tasks & improve EHR mental health time-to-market
documentation
• Synthetic patient data protects • Biomedical NLP supports clinical decision-
• De-noise radiology scans patient privacy making
Industry

Financial • GenAI assistants analyze & • GenAI chatbots simplify day-to-day • Synthetic training data improves financial
services & synthesize financial data at scale financial tasks models & ensures compliance
insurance • Automated underwriting decisions • Personalized interactions in • Pattern identification in unstructured
insurance sales process claims filings to minimize losses

Retail • LLM-powered search improves • Smarter, more relevant search • GenAI automates product catalogs
conversion
• Personalized avatars • Synthetic humans save on model costs

93
3. Opportunity in vertical genAI

Healthcare &
life sciences

94
HEALTHCARE & LIFE SCIENCES

Health systems and

pharma players are
using genAI to scale
everything from drug
design to EHR
documentation

Source: CB Insights — 7 applications of generative AI in healthcare 95

HEALTHCARE & LIFE SCIENCES

AI expertise is a necessity in sectors like pharma to

reduce time-to-market
Select generative AI drug discovery & design exits in 2023

Acquired by Recursion Acquired by BioNTech Filed for Hong Kong

Pharma in May 2023 in January 2023 IPO in June 2023

CB Insights — Understanding generative AI’s potential in healthcare - webinar 96

HEALTHCARE & LIFE SCIENCES

GenAI copilots for

doctors automate
tedious tasks like
note-taking

Source: CB Insights — Virtual scribes & summarization tools market report – ESP, Generative AI copilots for 97
doctors have raised more than $240M
HEALTHCARE & LIFE SCIENCES

Up-and-comer Corti raises $60M Series B in September

2023, taking on Microsoft’s Nuance

Source: CB Insights — Corti Analyst Briefing; Virtual scribes & summarization tools market report 98
HEALTHCARE & LIFE SCIENCES

Applications to enhance well-being and mental health emerge,

including AI-generated music, VR landscapes, and companions
Top-funded companies developing AI companions (as of 10/30/2023)

175 +661%

1,646
+370%

Source: CB Insights – AI companions market report 99

HEALTHCARE & LIFE SCIENCES

EHR workflows are ripe for LLM disruption, from document

search to summarization to suggested diagnoses

I can potentially ask ChatGPT, hey, We have tens of thousands, if not

does this person have out of network hundreds of thousands, of patients on
coverage and is this person eligible for our devices. We have device populations
spine surgery or something like that? in the millions... For service and
Then, we are having to look at multiple operations, there's a high demand in
documents and you don't know where to terms of the support that they can give
look, essentially, and you're essentially to a patient if they're in a trial or if they're
just giving the combination of all these just going about their day-to-day life on
documents as an input to ChatGPT. therapy, on one of these devices…So,
why we were investigating these
chatbots was to lower their cognitive
burden so that 10:1, 5:1 ratio could be
equalized.

VP, Machine Learning, Fortune 500 company Sr. Research Engineer, Fortune 500 company

Source: CB Insights — OpenAI software buyer interview transcript, Cohere software buyer interview transcript 100
HEALTHCARE & LIFE SCIENCES

Bundling genAI tools with existing cloud subscriptions is

giving big tech companies an advantage with market reach

Advantage is with, for example, John Snow Labs, it's a very sort
of clinically trained model... It's not just trained on wiki pages or
like general text. In that sense, I think it's much better…in terms
of entity recognition and things like that.

But the limitations, I would think, are these models are not
getting trained on the volume of data anywhere as close to
what ChatGPT is trained on… It [OpenAI deployment] was pretty
minimal overhead… the goal… is to essentially enable the use
of ML tools that are available from the Azure subscription at
the Enterprise level…

VP, Machine Learning, Fortune 500 company

Source: CB Insights — OpenAI software buyer interview transcript 101

HEALTHCARE & LIFE SCIENCES

Vendors will compete to license real-world datasets to

build genAI models tailored for the healthcare industry

LLM trained on 10 years of

GenAI for imaging trained on LLM trained on 2M patients’
health records and 400K
X-rays and radiology reports clinical notes
patients’ clinical notes

Source: Company research & announcements 102

HEALTHCARE & LIFE SCIENCES

Privacy and data handling practices will be major issues

as the market matures

175 +661%

1,646
+370%

Source: CB Insights — Cohere software buyer interview transcript 103

HEALTHCARE & LIFE SCIENCES

Expect healthcare-focused genAI vendors to lean into

data security as a point of differentiation

1,646
+370%

Source: CB Insights — Generative AI Expert Collection 104

3. Opportunity in vertical genAI

Financial services
& insurance

105
FINANCIAL SERVICES & INSURANCE

In financial services,
generative AI will
automate tasks and
transform how
organizations
use financial data

Source: CB Insights — 3 applications of generative AI in financial services 106

FINANCIAL SERVICES & INSURANCE

Insurers are using

genAI to drive
more personalized
customer experiences
and internal
automation efforts

Source: CB Insights — 3 applications of generative AI in insurance 107

FINANCIAL SERVICES & INSURANCE

Finserv incumbents are experimenting with generative AI;

prime use cases in document summarization & extraction

175 +661%

1,646
+370%

Source: CB Insights — business relationships and news mentions; company websites and press releases 108
FINANCIAL SERVICES & INSURANCE

Generative AI will automate & enhance underwriting

We are on the large language models and the

potential benefit that that will ultimately bring
beyond algorithmic, particularly in underwriting and
claims and the ability to work — either replace work
that is done or make it more accurate, or work
alongside underwriters.
Chubb CEO Evan Greenberg
Q2’23 earnings call

Source: CB Insights — Advanced Search - Earnings transcripts 109

*Reflects quarter call occurred
FINANCIAL SERVICES & INSURANCE

Incumbents with access to vast financial data & resources

will develop LLMs purpose-built for financial services

Bloomberg is developing JP Morgan Chase is Intuit partnered with OpenAI

BloombergGPT – an LLM developing IndexGPT, an AI to develop GenOS – a genAI
built for financial tasks like chatbot for investment operating system that will
sentiment analysis and news advice and selection. power customer experiences
classification. across its product suite.

Source: CB Insights news mentions — Bloomberg, JP Morgan Chase, Intuit 110

FINANCIAL SERVICES & INSURANCE

B2B fintechs developing genAI tools across compliance,

document processing, and chatbots land big-name customers

Cognaize's solution had better accuracy,

evolution, AI and machine learning
capabilities, and user interface than the
other vendors we considered… Cognaize
is integral to our workflow as they
provide a tool that our team uses to
extract relevant data from public
documents and map it to our templates.

Senior Vice President, Publicly traded financial

management company

Source: CB Insights — Kasisto Analyst Briefing; Cognaize software buyer interview transcript 111
3. Opportunity in vertical genAI

Retail

112
RETAIL

GenAI use cases in

retail will cut costs
and create more
engaging content

Source: CB Insights — 6 applications of generative AI in retail 113

RETAIL

E-commerce search is getting smarter, with retailers

developing search powered by ChatGPT

Recommendations, product attributes, Help choosing items based on budget,

dietary considerations food constraints, menu ideas

Source: Company announcements 114

RETAIL

Platforms will move quickly to revamp text on product

pages & automate product catalogs with genAI

We have a custom contract with AX

Semantics that contains the usage of the
platform itself, as well as extended
Shopify Magic generates, revises, Scans customer reviews and customer support. Our contract allows us
and expands product descriptions generates a summary to create automated texts in nine
languages for different channels, such as
the product description on the web shop
or the catalog, search engine optimized
texts for our category sites, and also
content for social media.

Manager, Manufacturing company

Source: Company announcements; CB Insights — AX Semantics software buyer interview transcript 115
RETAIL

Specialized search vendors are moving fast to catch up,

with a focus on relevance to drive conversion
Leading e-commerce search vendors by equity funding (as of 9/30/2023)

We ended up rebuilding our entire

catalog infrastructure using
Algolia…Visual merchandising and all
the rules that come with it was a big
important feature that we really wanted.
…Until I met Algolia for the first time, I
have never seen a search engine of that
scale that was not Google or Bing or
Yahoo operate that quickly.

Head of Digital Product, E-commerce company

Source: CB Insights — Generative AI – e-commerce search market report; Algolia software buyer interview 116
transcript
RETAIL

Leading brands are working with genAI vendors to create

personalized avatars and diverse models at scale

Source: CB Insights — Generative AI – synthetic humans & fashion design market report 117
RETAIL

Retail-specific genAI solutions improve digital shopping

operations, engagement, and conversion

175 +661%

1,646
+370%

Source: CB Insights — The generative AI in retail market map 118

Promising
companies
to watch

119
 We identified the 50
most promising
genAI startups
Generative AI 50

Explore the full list

Source: CB Insights — Generative AI 50 120

 CB Insights customers can find and track every generative AI
company mentioned in this report using our analyst-curated
Expert Collection

Explore the Collection

CB Insights helps the world’s leading companies understand everything they need to know about disruptive 121
technologies — find out more about why our customers love us here.